The exciting future of customer experience innovation

00:00

good evening everybody before I begin

00:02

the proceeding is not behalf of all

00:03

those present I'd like to acknowledge

00:05

the gadigal people of the eora nation

00:07

I'd also like to pay respect to elders

00:09

both past and present acknowledging them

00:11

as traditional custodians of knowledge

00:13

for this land and that's something we

00:15

say particularly at the University of

00:16

Technology Sydney which is where I'm

00:18

from I'm delighted to see so many people

00:20

here tonight I'd like to give her very

00:21

special welcome to the minister

00:23

the Honorable Andrew constants Minister

00:24

for transport infrastructure howard

00:26

collins CEO of sydney trains tim Rearden

00:29

secretary for transport us at Wales as

00:31

many as well as many of our industry and

00:33

government partners in the UTS community

00:35

for those of you I haven't met I'm Glen

00:38

why I'm the deputy vice-chancellor of

00:39

innovation and enterprise at UTS it's a

00:42

glorious title for actually the last

00:45

three years I've actually been looking

00:46

after our research portfolio at UTS and

00:49

and actually as of last Tuesday I've

00:50

moved into this new portfolio and as

00:52

part of that role I look after our

00:54

innovation and entrepreneurship we do a

00:56

lot of work with students and industry

00:57

but the enterprise part of my work

00:59

effectively takes all of the work we do

01:01

in our research and helps connect it to

01:04

our community to our stakeholders to

01:06

industry and so on we are at UTS we are

01:14

a University of Technology I know that's

01:16

part of our name and we're very

01:18

committed to using technology to improve

01:20

lives whether it's from things like low

01:23

cost easy to operate systems to remove

01:26

arsenic from groundwater in Vietnam

01:28

which is an interesting project we're

01:29

involved in at the moment - the

01:31

technology that we'll be talking about

01:32

here tonight the responsive passenger

01:34

information systems this is all

01:36

characterized were characterised by our

01:38

commitment to collaborating with

01:40

partners to develop new technologies for

01:42

very practical and yet innovative

01:44

applications I won't say too much more

01:46

on there are responsive passenger

01:48

information systems you'll hear more

01:50

from our experts on that but this is a

01:52

really important project will help

01:53

improve customer comfort and overall

01:55

experience at busy stations like when

01:58

your d'Or Town Hall on the Sydney train

02:00

network which would be welcomed by many

02:01

including me I'm a person who loves

02:03

public transport I catch the ferry and

02:05

the train to work each day and I walk

02:07

through the lovely tunnel underneath

02:08

central so the idea of responsive

02:11

passenger information systems is

02:13

very appealing what's made this possible

02:16

that's been the commitment in close

02:17

collaboration between our researchers

02:19

and the industry partners involved

02:20

Sydney trains the rail manufacturing

02:22

Cooperative Research Centre and

02:24

transport for New South Wales I'd like

02:26

to congratulate everyone involved I

02:28

remember a meeting we had with the

02:30

minister

02:31

probably 12 or 18 months ago where we

02:33

kind of threw around some ideas and it's

02:36

been very very exciting to see since

02:38

that point a number of us coming

02:41

together and working on this on this

02:43

technology

02:44

I'd like to as they congratulate

02:46

everybody

02:47

clearly there's unprecedented passenger

02:50

growth I just learned that there are 1.3

02:53

million passenger journeys on Sydney

02:56

rail everyday I think 388 million year

02:58

that's an incredibly large number

02:59

clearly there are people like me who

03:01

have increasing expectations of the

03:04

responsiveness of our of our public

03:07

transport system and clearly there's

03:09

lots of infrastructure constraints as we

03:12

grow Sydney and and demand more

03:15

Transport and rail capacity I believe

03:18

that only by working together can we

03:20

deliver innovative solutions that are

03:22

going to be of benefit to to all of us

03:25

here in in Sydney and then to take those

03:27

solutions and translate them into

03:29

capability working with smaller medium

03:31

enterprises and other organisations

03:33

industry to potentially take them out to

03:35

the world so you've probably heard

03:37

enough from me

03:37

I'd like now to introduce one of the

03:39

people behind this great work at UTS and

03:42

she'll be our host for this evening dr.

03:44

Michels eyebots from the UTS Transport

03:46

Research Centre Michelle is a transport

03:48

planner specializing the analysis of

03:50

sustainable urban passenger transport

03:52

systems in her research at UTS she works

03:56

with our Institute for sustainable

03:57

futures as well as within our Faculty of

03:59

engineering and information technology

04:00

where she lectures and transport

04:02

engineering we're very fortunate to have

04:04

her at UTS she's been absolutely

04:06

integral to this research project and to

04:08

building collaborative research

04:10

partnerships with all of the parties

04:11

involved please join me in welcoming dr.

04:14

Michels eyeballs

04:16

[Applause]

04:20

Thank You Glenn so tonight's very

04:23

exciting for us it's the culmination of

04:25

several months of research that's taken

04:28

place over the last year it's been a

04:31

very big team of researchers at UTSA

04:34

from a number of different faculties and

04:36

many of whom are here tonight what we

04:40

want to do tonight is we want to talk

04:42

about the nature of the the problem that

04:45

we've been charged with trying to solve

04:48

and we also want to talk about the

04:50

technology so as you can see we're all

04:53

set up for that and done hopefully

04:55

that'll be quite exciting so what we're

04:58

going to try to do we had a dress

05:00

rehearsal earlier that went sort of okay

05:02

so hopefully now that we're here and

05:05

doing the real thing and we've got all

05:07

of you here to to watch and participate

05:09

in this in this demonstration hopefully

05:13

we'll it'll go really smoothly but

05:15

before we tell you that story what I'd

05:18

like to do is just give you a little bit

05:21

of background about about the project so

05:25

basically we are looking at the

05:28

development of a new class of passenger

05:31

information system and it uses robotic

05:34

technology so it's only been made

05:36

possible with the new sensor

05:38

technologies and an actuator

05:41

technologies that are around at the

05:44

moment that weren't available 10 years

05:46

ago and responsive passenger information

05:50

systems they differ very much to the

05:51

earlier forms of passenger information

05:53

system technology which we call static

05:56

and dynamic so statics are basically

05:58

just the signs that we see on railway

06:00

stations and throughout the railway

06:02

environment and dynamic signs are those

06:05

signs that change so we often see the

06:08

the the dynamic passenger information

06:11

systems on platforms where you can see

06:14

things changing and the times are being

06:15

updated and what-have-you so what we're

06:17

proposing here goes goes far beyond that

06:21

and I guess it's also a form of

06:24

technology that we hope will demonstrate

06:26

tonight that utilizes our very

06:28

new and different approach a very new

06:30

philosophy to technology and we're not

06:33

just seeing it happen in transport we're

06:35

seeing it in a number of different areas

06:37

you'll often hear the term human in the

06:40

loop or human centered design and what

06:43

that basically means is that designers

06:46

engineers technologists are now looking

06:48

at ways in which to look at the the

06:51

natural instincts the the predilections

06:53

the needs of users of a system and

06:56

actually integrating that into the

06:59

structure and function of the system in

07:02

order to make the system work even

07:04

better than what it than what it

07:06

otherwise would and this is a very big

07:08

step away from old approaches to

07:10

technology that involve what what are

07:12

often called command and control

07:14

approaches where you've got actors or

07:17

operators directing everybody but not

07:20

necessarily or really considering what

07:23

it is that users users are after and

07:25

what what they might want so this is all

07:28

very compatible with the customer

07:30

service philosophy that we've seen

07:32

blossom here in Sydney recently networks

07:35

like the Sydney trains Network are very

07:37

different now to what they they were and

07:39

how I remember them only just a few

07:42

years ago so it's been wonderful to see

07:45

that change and we're hoping as UTS

07:49

researchers that we'll be able to

07:50

contribute to that ongoing change and

07:53

improvement so okay back to our story

07:57

we're drunk going to try and tell it as

07:59

a story in order to make it more

08:01

interesting so I remember an engineer

08:03

once saying to me you know Michelle

08:05

people respond to stories they don't

08:07

respond to bullet points so here we go

08:10

let's let's hold on tight and see what

08:13

happens all right so every new

08:15

technology needs a problem so what I'd

08:18

like to do in order to get us on our way

08:20

with this story is to ask Tony ade to

08:23

come up and join us

08:24

so Tony is the executive director of the

08:28

future network delivery Directorate and

08:30

he's responsible for planning our future

08:33

network the stuff exactly so I didn't

08:39

know where the clicker

08:41

okay all right so to get us going so

08:45

Glenn's mentioned it and it's been

08:48

through all the papers lots of people

08:50

are talking about these incredible

08:51

growth rates that we're seeing on the

08:53

Sydney trains Network can you give us

08:55

some background on that what's been

08:57

happening just right it's run through it

08:58

well certainly there's no doubt about it

09:00

there's been unprecedented growth on our

09:03

network in fact it's so busy out there

09:05

it's it's really becoming something we

09:08

need to address very quickly if I draw

09:10

your attention to that graph up there

09:11

and you can see from 1990 there's been a

09:13

steady growth right through till about

09:15

today but when we get further on into

09:18

the future and we predict the future

09:19

it's absolutely growing in a very rapid

09:22

pace and it means we have to do

09:24

something differently to address that in

09:26

the immediate future right now here and

09:28

now with more trains more services and

09:30

obviously we have to think carefully

09:32

but the innovative in smart ways of

09:34

tackling the future is it just the

09:36

morning and afternoon hope that we're

09:38

seeing this growth in or is there

09:39

something no not at all like you know if

09:42

he if you guys are on the trains all the

09:43

time you can now see that it's not

09:45

typically just a name and PM peak our

09:48

trains are fall all the time not just

09:50

any time and if you take traditionally

09:53

an aim and P and peak we don't we're not

09:55

a name picked railway anymore we've in

09:57

the off-peak between Peaks and on

09:58

weekends it's quite extraordinary and if

10:00

you have a look at that slide but it's

10:02

got 2021 on weekdays by 2021 we're gonna

10:06

have a 20% growth on weekdays but take a

10:09

look at the weekend a hundred and twenty

10:12

percent and for those who don't know we

10:14

carry about a half a million people on a

10:16

weekend just imagine 1.3 million people

10:20

travelling on our network on a weekend

10:22

and that's a really good story yeah so

10:24

what what what is causing this cuz it's

10:26

it does seem to have happened quite

10:28

suddenly well it hasn't really been

10:31

suddenly don't take this station here

10:33

Town Hall it's been happening for a

10:36

while okay but take Town Hall and have a

10:38

look at that station in the a.m. P it's

10:40

full of people and we'll talk about that

10:42

a bit later on but when you look at that

10:45

and you look at things that we've been

10:47

doing as Sydney trains over the years

10:49

you know we've introduced Opel we're

10:51

much cleaner we're lots

10:53

customers in our staff of law interact

10:55

in staff very visible look at their

10:58

stations and there are corridors and you

11:00

see development everywhere so people are

11:02

attractive to the set to the network and

11:04

I think the biggest success for us more

11:07

so than just our brand is the fact that

11:08

we've built something here that's quite

11:10

unique we have been able to demonstrate

11:13

to the community and our customers that

11:15

we care and we move them they trust us

11:18

to get about their day and I think

11:19

that's a very big deal for us yeah ok so

11:22

to the project now so in the very early

11:25

days when we started you gave us the

11:28

direction you said look I want you to

11:30

focus on Town Hall station what could be

11:32

done there with the technology and also

11:34

at Parramatta station so just give us a

11:37

bit more background on why you why you

11:38

asked us to focus on that well there's

11:40

three key reasons why we did that let's

11:42

start with Town Hall Town Hall is a busy

11:44

station very unique 250 people interact

11:49

with that station every day that's a

11:51

large number in any shape or form the

11:53

staff at Town Hall have to move that

11:56

type of number on and off trains and

11:58

they do a fantastic job every day doing

12:00

it but just imagine that train that's

12:02

just on that picture they're arriving on

12:04

that train is about 1,200 to 1,500

12:06

people about a thousand of those people

12:09

are going to get off onto that platform

12:10

we're on that platform there's seven

12:13

hundred plus people to get on and how do

12:15

you do that and manage that with the

12:17

train every three minutes coming in and

12:19

out with the same scenario and hold that

12:21

train for no more than 60 seconds so you

12:23

keep the railway running it's quite

12:25

remarkable

12:26

so Town Hall is one of the reasons

12:28

because if there's anything we can do to

12:29

make the innovation and the operation a

12:32

lot more seamlessly over there and

12:34

predict something like that wasn't

12:35

really much so help us then we talk

12:38

about Parramatta in Parramatta is the

12:39

other location risks and phenomenally

12:42

starting to grow to the point where it's

12:44

a city in its own right very soon

12:46

there'll be 20 trains per hour running

12:47

through that place curved platforms same

12:50

scenarios and what we found ourselves in

12:52

a position is that we may have a new

12:54

Town Hall on the hands so we have to

12:56

bring that into play so I guess for - is

12:59

that by bringing incorporating - very

13:01

busy places on our network the west part

13:03

of Sydney and the

13:04

every day anything that we come up with

13:06

here we can get immediate benefit and

13:08

that's the Russian help me I live in

13:09

those stations so I'm hearing you on

13:12

that but I have to say having worked on

13:14

the project now Town Hall is I I mean I

13:17

was quite overwhelmed by what it is that

13:19

Sydney trains does every day in order to

13:21

keep that operation going

13:23

I've got all the respect for the staff

13:24

it's taken a long time for those guys to

13:26

get it to a position where they move in

13:29

seventy plus thousand people in two

13:30

peaks getting them through there

13:31

seamlessly very safely and in the third

13:34

and final thing is that map that you see

13:36

up there we've got quite a few trains in

13:39

the morning peak making their way into

13:40

the city carrying thousands of people

13:43

and when you think about it for one hour

13:45

there's a hundred and twenty trains

13:46

merging from 15 lines curving their way

13:49

to six lines to make their way through

13:51

and every one of those trains will make

13:54

its way to Town Hall some shape or form

13:58

250,000 people touch that place every

14:00

day and we really need to respect that

14:01

as they work through this okay so with

14:04

them I mean I'm also very mindful that

14:06

we've got a new timetable

14:08

just about to be introduced so with that

14:11

new timetable how does that impact on

14:13

Town Hall ah be facts beautifully let's

14:17

start let's start let's start very

14:19

simply it's trains are gonna go through

14:21

there very frequently and you could

14:23

argue that if you've got a train

14:24

traveling through there on time very

14:26

frequently and more rather than coming

14:29

through with more choice of where to go

14:30

you will move the people a lot easier

14:32

our customers a lot easier but but to do

14:34

that takes extraordinary effort by the

14:36

station staff to train crew and all the

14:38

operators to make that work so you still

14:40

need to do more but when we talk about

14:42

the timetable we talk about the biggest

14:44

permanent uplift of a timetable ever

14:47

brought on this network and and look

14:49

I've been here a long time and this is

14:50

the biggest thing that we've ever done

14:51

in terms of preparing for a timetable

14:54

1,500 extra trains per week that is

14:57

quite extraordinary have a look at the

14:59

weekend to address that growth 750 extra

15:03

weekend trains will now appear that

15:05

doubles all the services everywhere

15:08

including the airport line where it's

15:09

much needed and between the AM and PM

15:11

peak we're pushing that to 250 extra

15:14

trains in that period

15:15

and that's a challenge for everybody but

15:17

it really starts to address the the

15:19

crowding issues that we have on the

15:20

network and it starts to produce us as a

15:23

very effective rolling so so you are

15:25

going to try and push more train paths

15:27

through Town Hall during the peak

15:29

periods with this new time travel not

15:32

quite right okay

15:33

I probably have to take a step back

15:35

which have put more trains through there

15:37

we need more singling systems and we

15:39

need new and innovative technologies to

15:41

help do that we are at track capacity

15:44

and we need to do more if we're looking

15:45

at that projected number into the future

15:47

and we are investing in that space and

15:50

in fact when we move to new signalling

15:52

and the technologies that we're talking

15:53

about here today and the demonstration

15:55

will will help I think people understand

15:57

the effort going into managing Town Hall

15:59

but we also don't see relief for quite

16:01

some time

16:02

the new set that the minister and the

16:06

government of the day has invested

16:07

heavily in Sydney Metro by 2019 we have

16:11

the first cut but it won't be until 2024

16:14

when that train that Metro makes its way

16:16

under the harbour and through Central

16:18

until we get some form of relief for

16:20

that area when that happens and we

16:22

introduce new signalling system we will

16:24

be able to run more but at until that

16:26

time until 2024 we have to think of a

16:28

lot smarter ways of doing it so we've

16:30

still got this problem at the at the

16:32

core at the center at the network centre

16:34

that we have to deal with okay so more

16:40

growth going back to what you were

16:41

originally saying so that the the

16:43

further growth it does need more train

16:45

pads so we've still got we've got some

16:48

time to wait for that there is new

16:50

technologies on the horizon but we've

16:52

still got to just we've just got to do

16:54

something in between and that's what

16:56

we're here today I mean this is a good

16:58

example of what we can do to support the

17:00

staff give them the tools let them know

17:02

some behind the scenes activity that

17:04

will help nudge the customers to the

17:07

areas they need to go so that we can

17:08

actually load those trying to get them

17:10

out much quicker and you never know if

17:12

we get that right in sequence a try it

17:14

might actually stimulate another train

17:16

path I guess we won't know until we test

17:18

it okay so it's so it's sort of about

17:19

stabilizing the system as much as we

17:22

care as

17:22

trying to expand it that's great okay so

17:24

train paths are really valuable here

17:26

they are they are all right so hold that

17:28

thought

17:29

and if you'd like to just wait here or

17:32

there what well the growing there could

17:34

be a growing crowd here ok no I'll wait

17:36

over here outside there's the Train the

17:38

body right here all right

17:40

do you want to give me the clicker oh

17:44

they took the clicker off me earlier

17:46

already so what I'd like to do now is

17:49

I'd like to invite professor John Rose

17:51

to join us so professor professor John

17:54

Rose is the director of the the Center

17:58

for business intelligence and data

17:59

analytics at UTSA and John is an

18:02

economist so quick question how much is

18:06

a train path worth so the simple answer

18:09

to that is about 50 to 60 million

18:10

dollars annually mm-hmm okay and how did

18:15

you get that figure so a bunch of

18:18

calculations of course being an

18:20

economist a bunch of assumptions so if

18:23

you assume that each traditional train

18:25

can carry say 1200 people and that

18:30

people will also be you'll have

18:31

additional trains on the shoulders of

18:33

the peak you can generate about 2,000

18:36

jobs from an additional additional train

18:39

path then based on that most of the jobs

18:43

that were likely to be created in the

18:45

CBD from that additional train path

18:47

they're going to be higher value type

18:49

jobs so you could occur you can say that

18:52

that's basically worth 37 million

18:53

dollars add on additional tourism and

18:57

movement of lower skilled jobs and lower

18:59

value jobs to suburban centers and you

19:02

get about 50 million dollars okay

19:04

so that's quite a lot I'm impressed by

19:06

that so even if we were conservative and

19:09

we took it down to say 40 million that's

19:11

still quite a lot so just a trained par

19:13

from our existing net worth trying to

19:15

squeeze that out of it it's worth it's

19:17

worth quite a bit so and a lot of that

19:19

is because of the the particular role of

19:23

the CBD isn't it so could you tell us a

19:25

little bit about that yes so stephenie's

19:27

are really important to two cities like

19:29

Sydney because that's where most of the

19:32

high value jobs are actually created and

19:34

actually

19:35

exist so the jobs that are basically

19:37

designed to service international trade

19:40

global trade that's not to take away

19:44

from suburban centers but the more

19:46

suburban centers are basically serviced

19:49

the local residential areas and the

19:53

workforce that is needed for the CBD

19:55

centers so the CBD is the really big

19:57

income generator for the the city as a

19:59

whole that's why we see these bigger

20:01

numbers here and it shows how

20:03

structurally it's important so the UM

20:06

Tony mentioned Parramatta CBD as well

20:09

and there's also been a lot of

20:11

discussion not just recently but for a

20:13

long time about paramater is the second

20:15

CBD so it does that somehow conflict

20:19

with the idea of trying to get more

20:20

trains through the Sydney CBD or is

20:23

what's going on there do you think so I

20:25

don't think it actually conflicts with

20:27

what I'm saying I think that we will get

20:30

to the point where somewhere like

20:31

Parramatta could be a second CBD I think

20:34

at the moment though the problem is

20:36

capacity so basically the public network

20:41

public transport network is designed to

20:43

get people into the city from all across

20:45

the Sydney region so once we can

20:48

actually build a network and get people

20:50

to be able to travel to Parramatta from

20:53

all over Sydney I think you'll see some

20:55

shifts from the city towards that but at

20:58

the moment we just have these

20:59

constraints that I think preventing that

21:01

from actually happening so it's really

21:03

the the important thing for firms and

21:05

these global businesses is that they're

21:06

able to access the entire metropolitan

21:08

workforce and not just a local area

21:10

which is why maybe Parramatta has

21:13

struggled to take on that same that same

21:15

role and I think we'll get there in the

21:16

end but I just don't think we're there

21:18

right this moment okay all right so from

21:22

a people perspective that's really

21:26

interesting

21:28

so what I think we might do now is ask

21:33

Susanna LeBron to join us thank you so

21:36

Susanna is the executive director of the

21:39

customer service director at Sydney

21:41

trains and and Suzanne is responsible

21:43

for representing the customer inside

21:46

Sydney trains and lifting the value

21:48

proposition so Susanna Taney was talking

21:52

a lot about the growth and what's been

21:54

responsible for that and he also

21:55

mentioned very briefly that the the

21:57

quality of the services or what people

21:59

are experiencing is just different now

22:01

and it's made it more attractive so what

22:03

sorts of things have been happening and

22:05

I've just remembered to give you the

22:06

clicker that's right so what is really

22:10

important to remember is that in 2013 we

22:14

took on a really ambitious program of

22:17

uplifting our customer service across

22:19

Sydney trains but transport at large and

22:22

what I'm really pleased to say is that

22:25

we have sat at 90% now for a strong 18

22:29

months but we did start that journey at

22:31

78% what I want to show you next is are

22:35

the nine key drivers that we focus on

22:38

which our customers naturally when you

22:40

look at them they're quite simple but

22:42

these are the areas that our customers

22:44

consistently remind us that we need to

22:46

focus on I'm just going to talk about a

22:48

couple straight-up machine so first of

22:52

all ticketing we have had one of the

22:55

biggest changes in the ticketing product

22:57

that you would have seen you know across

22:59

the whole of I suppose Australia and it

23:01

has been a resounding success

23:03

so the Opel option the Opel product 13

23:07

million people per week use that to

23:09

travel so 30 million travel on on Opel

23:13

half of those are on rail so naturally

23:17

we need to pay attention to looking

23:19

after those customers that are

23:20

travelling on us with the Opel product

23:22

the other thing that's really important

23:24

for our customers is timeliness so what

23:26

does that mean on-time performance our

23:28

customers really need to know and need

23:31

to know that we've got a reliable

23:32

service again pleased to say that in the

23:35

last financial year ninety three point

23:37

four percent on-time performance which I

23:39

think any of you that travel regularly

23:41

know that you can pretty

23:42

much turn up and nearly go in most of

23:44

our stations and great with the more

23:46

trains more services that's going to be

23:48

even more enhanced one of the other

23:51

things that I think it's important to

23:52

call out because I think if you go to a

23:53

barbecue now people often say oh the

23:56

trains are so clean the station's are so

23:58

clean that's the reality huge investment

24:01

has occurred not only in our station

24:03

environment but with our trains as well

24:06

and the actual benefit of cleanliness

24:08

also means that our customers feel safe

24:11

and secure so through that investment we

24:14

also have customers that are now really

24:16

pleased to be at stations at all times

24:19

of the day even through the night

24:20

because of the way it's been maintained

24:23

and cleaned okay so can I just interrupt

24:25

you there a little bit so on all of

24:29

these areas not just the the ones that

24:31

you've covered there so the customer

24:34

satisfaction survey result is what's

24:36

been driving and guiding a lot of what

24:38

you're doing

24:39

so you're scoring on ninety ninety

24:41

percent are you are you looking to just

24:43

stay there or what tell us what well I'm

24:46

not going to do I'm not gonna do crazy

24:47

straight and say yet we're at ninety

24:49

percent and we'll just leave it there

24:51

the reality is to maintain such a high

24:53

score and any major organization that

24:56

has customer services its key

24:57

deliverable needs to consistently look

25:00

at how they need to improve so to

25:02

maintain we need to keep improving one

25:05

of the other two is that's really

25:06

important and actually leads into why

25:08

this UTS partnership has really started

25:11

to bring around success for us is around

25:14

customer service so if you get customer

25:17

service right which we clearly are doing

25:19

really well with that the model of out

25:21

and about has only been really supported

25:24

through technology it's enabled our

25:26

people at stations to be responsive and

25:29

dynamic and have real-time information

25:31

through the technology that we now have

25:33

the other thing that leads into that is

25:35

information information is critical and

25:39

if we have information that is proactive

25:41

and available not just for us but for

25:44

our customers is actually the game

25:46

changer to match what Tony is absolutely

25:49

articulated as a growing patronage for

25:52

us so

25:53

you need to you need to find new ways of

25:56

doing things in order to just to stay

25:58

let's click on it's just the clicker all

26:01

right right we didn't really cause that

26:03

but but this is this is the clincher

26:06

page okay this is the one so we've had

26:09

some great workshops with UTSA and the

26:13

reality is that not only have we had a

26:15

fantastic partnership and we are

26:18

benefiting and you know transport at

26:20

large is benefiting but our customers

26:23

are now going to benefit from the

26:24

partnership that we've now got through

26:26

UTSA the key on this diagram is to

26:29

actually look at the fact that we've got

26:30

arrows going in both ways we know that

26:33

with the growth and with the success of

26:35

the service that we now offer we

26:38

actually need our customers to help us

26:40

out because as you can see with Town

26:42

Hall if we just rely on us as operators

26:46

to gather information and work with that

26:49

information we may limit our success on

26:52

how we look after these customers in

26:53

their journey but if we actually get our

26:56

customers and we've had some fantastic

26:58

our apps already today but if we

27:01

actually had our customers to be part of

27:03

this and this is where you heard the

27:04

words bye-bye Tony's saying nudging and

27:07

behind the scenes technology it means

27:09

that we really will fulfill all the

27:12

information that we need to actually

27:13

look after our customers and have them

27:15

part of the operational journey so if

27:17

you've got new ways to think about all

27:20

of this information and engage with your

27:22

customers then you're able to to get

27:25

that relationship going in more of a

27:27

two-way thing cuz that this is this is

27:29

what's exciting was exciting for us as

27:31

researchers is this human-in-the-loop I

27:33

think so people have got these wonderful

27:36

instincts and desires like they want to

27:39

be more comfortable they want to have a

27:40

better experience so it's it's just

27:43

trying to find the ways to harness that

27:45

um you're probably are thinking okay

27:48

seriously what is it that we're gonna be

27:50

shown tonight but this is laying the

27:52

foundation for this somehow we need to

27:55

get information that is proactive and

27:58

real beyond what we have currently which

28:01

is our people on stations position yes

28:04

they've got the bright orange on but

28:06

visually they

28:07

can only see so much to be able to get

28:10

technology and the ability to understand

28:12

what our customers are doing on a

28:13

regular basis means that we can be

28:15

proactive in addressing what we know is

28:18

about to come around the corner which is

28:20

consistently more requirements of our

28:23

public transport system so if we've got

28:25

those that information going to people

28:27

and they're getting more stuff on their

28:29

mobile phones and their staff not the

28:33

staff people the Sydney trains people

28:35

they're also getting more information

28:36

about what's going on then we're able to

28:40

get that more symbiotic relationship

28:42

going so a lot of you that travel on our

28:45

on our transport network at large would

28:48

be familiar with a multitude of apps and

28:51

information that you can access we are

28:54

consistently growing that technology to

28:56

meet the demands working very closely

28:59

with the department headed up by Toni

29:01

Braxton Smith in transport around all

29:04

the apps and the ability to understand

29:06

intermode connections of our customers

29:08

so for us we can't just stop here we

29:11

need to understand behind the scenes

29:13

what else can we feed into these apps

29:14

into this technology to help not only

29:16

our people in the operator environment

29:18

but our customers understand what is it

29:21

that they need to do to actually get

29:23

their experience better ok so just to

29:26

finish off what's your vision how do you

29:28

want to use all this tech what do you

29:30

want people to fasten aspirationally and

29:32

I'd like to think that I can drop that

29:34

word and we'll eventually make this

29:35

happen I would love for someone to get

29:38

to their destination whether it's

29:40

sitting at an office or sitting at the

29:42

movie cinema and actually just think for

29:45

a minute how on earth did I get here I

29:48

caught a bus and then I caught a train

29:50

but I glided through those experiences

29:53

because I had information at hand that

29:56

enabled me to make choices and wise

29:58

choices but I also had people in the

30:01

operational environment nudging me and

30:04

getting me around certain obstacles that

30:07

when I got to my destination it was so

30:09

seamless they're really the recollection

30:11

is that they just glided through so

30:13

that's that's what I'd like to think

30:15

we're done yeah all right thanks for

30:17

that now there's some other people that

30:19

have had vision

30:20

in the past there have been very

30:22

critical to this story oh thank you so

30:27

I'd now like to ask the minister the

30:29

Honorable Andrew Constance to to join us

30:32

up here so Andrew of course is the

30:35

Minister for transport and

30:36

infrastructure and he's actually played

30:39

quite an important role in this project

30:41

you might not think so but you have

30:42

because I think I think if you hadn't

30:46

told us to go and do it it wouldn't have

30:48

happened so really that's that's what

30:50

did occurred but look there's a question

30:53

that I've been dying to ask you for a

30:54

long time I've never had the opportunity

30:56

to do it because I was really impressed

30:58

and taken as a transport planner with

31:00

the whole customer service thing that

31:02

glad is particular and introduced and I

31:04

thought frankly it was going to be very

31:06

difficult for you to top that but don't

31:08

television I know it's pretty good no

31:13

it's pretty good but the whole

31:16

technology thing in the technology focus

31:18

that you've really given the portfolio

31:20

has been I think incredibly significant

31:23

and I just want how did you know how did

31:27

you know that it was going to pan out to

31:29

be as significant as what it has been

31:32

well I think there's a couple things

31:34

first of all and transports a technology

31:38

business and it will grow so rapidly and

31:42

so quickly so as we progress through a

31:44

world where you know people basically as

31:47

Suzanne I said you know just to be it to

31:50

be completely seamless so the impacts

31:52

that artificial intelligence will have

31:54

the impacts in which certainly we're

31:57

already seeing in terms of utilization

31:59

of the data behind Opel ultimately

32:03

people want a personalized service

32:06

despite it being a public transport

32:08

system that we're talking about and and

32:11

ultimately they want to be able to use

32:13

technology in a way where it's pretty

32:15

much the speed of thought so

32:16

we we're going to see I think

32:18

increasingly and we've seen people

32:20

obviously take-up yes we we open the

32:22

data to allow those apps to be developed

32:23

this is much bigger than this and I

32:26

think we've got a tired old Network and

32:28

we're trying to utilize technology here

32:30

to make it work more effective

32:33

and and ultimately as we see the advent

32:36

of for instance autonomous vehicles

32:38

we've also got a little ways in which we

32:40

can apply some of those principles to

32:42

the way in which people move in and

32:43

around this tiedoll network so I think

32:46

that's shot of Town Hall where it's

32:48

crowded between the stairs and and that

32:52

magical yellow line again we can't widen

32:56

the platform so the question then

32:59

becomes well how do we position people

33:00

to get on to a train more effectively I

33:03

mean ultimately because of the way that

33:06

behaviors have been embedded into us you

33:10

know we want to speed up the way in

33:12

which people get on and off the trains

33:14

as Tony alluded to but we also of course

33:16

want to get the customer satisfaction

33:18

high I say you know it wouldn't matter

33:22

if if your your mind cities will I need

33:24

to jump on carriage number two because I

33:26

know that if I do that I can get up the

33:28

escalators at Martin place more quickly

33:30

to get to Parliament House why you'd

33:32

want to do that the first place I don't

33:33

know a bit but just as a principle

33:35

because I do it but but the point is in

33:40

many ways we're setting out ways and

33:42

this is our gonna take it to a different

33:43

sphere of thinking okay so when when

33:48

when we first came to see you because I

33:49

came to see you when we were launching

33:51

the UTS Transport Research Center and I

33:54

thought well we should actually tell you

33:56

what we're doing so you know what we're

33:58

launching and I have to say I was a bit

34:00

surprised at you know the the

34:02

supportiveness that you had for the

34:03

technology that that we were showing you

34:05

so okay can I ask just quickly what what

34:08

was it that that picture interest or or

34:10

what when we were showing you I think it

34:12

was some of the sensor technology work

34:15

that the guys at the Center for

34:16

autonomous systems have done with the 3d

34:18

sensing that we're going to see so what

34:21

was there anything that sort of struck

34:23

you about that I knew a technology sort

34:25

of person yourself like do you did some

34:28

consulting work but I was doing

34:31

consulting away from Microsoft Asia

34:33

alright so yeah I mean look I think that

34:36

the and it was in the sort of government

34:38

relations corporate space but I think

34:42

that the thing about all of these

34:43

is that we can do so much better and the

34:48

technologies are there but it's a

34:50

question now of how we take it to a

34:52

different level and some of the

34:53

technologies will advanced very quickly

34:55

you know people don't want timetables I

34:59

mean if you were running a service every

35:01

three minutes and the peak there's no

35:02

need for a timetable and so with as of

35:07

this weekend we'll be making some

35:08

announcements about those timetables but

35:10

the the key point is I mean around 70%

35:13

of the train network is going to have a

35:15

train arrive at a station anywhere from

35:17

an average 3 to 15 minutes max so where

35:22

we're moving away so it becomes a turn

35:24

up and go system the Metro certainly is

35:27

but you know it's a system which will we

35:29

could almost deliver a train every two

35:30

minutes similar to what's being

35:32

delivered in parts of Asia now so we we

35:35

want to change the behaviors but if

35:36

we're able to use the technologies to be

35:38

able to do it it does mean some of the

35:40

political sensitivities that you know

35:42

have dogged particularly politicians in

35:45

the past about crowded services and you

35:48

know trains it alight and buses it'll a

35:50

you know we we move beyond so that's

35:54

where the technology is taking us and

35:55

it's it's a good political outcome as

35:57

far as I'm concerned I remember I

36:01

remember talking to quite specifically

36:03

about interchanging and from a

36:05

professional perspective we we've often

36:07

thought that people on Sydney's train

36:09

network don't really interchange very

36:11

much and and I guess it's that's related

36:15

to the the unprecedented growth levels

36:17

that we're seeing as well so what what's

36:20

your take on the on the growth levels

36:22

like it what what do you think that that

36:24

means for the future yeah I mean

36:26

everyone's talking about over

36:27

development at around the city at the

36:28

moment people certainly don't want

36:32

growth if they can get around I think

36:34

that's fair to say but the the

36:37

interesting thing for us is we inserted

36:39

a a transfer discount into the Opel

36:42

system last year 12 months ago and I

36:46

think at that time it's about you know

36:47

30 percent of commuters would

36:50

interchange at least once a month so

36:52

they either get from a bus to a train or

36:54

vice versa

36:56

as of yesterday we're now at 50% on a

37:00

regular basis cut communities are

37:02

starting to interchange so there I had

37:05

to I had this problem where our revenue

37:08

growth was going that was going down a

37:11

patronage growth going the other way

37:13

that's unsustainable and what we found

37:16

was that the eight trips and then the

37:18

rest of the week free was causing that

37:22

to occur I mean we need money to run the

37:24

system so we had to make a decision as

37:25

to well how do we deal with this

37:27

interestingly by putting the $2 transfer

37:29

discount in it has started to change the

37:33

transport planning and we had some

37:35

ridiculous anomalies in the network

37:37

where for instance of you know each

37:39

cliff to the 10 to the city we wouldn't

37:41

have people get to educate for one bus

37:43

and then jump on the train they wanted

37:45

to stay or interchange to another bus so

37:47

that they didn't have to pay twice so

37:49

just doing some of those sensible things

37:51

so we had a quite literally a bus

37:53

running the exact same route as a train

37:54

route on an underutilized line so just

37:58

changing some of those things is really

37:59

important the other thing is is that

38:01

what we are now experiencing is because

38:04

of Opel that the data is far more

38:06

accurate people aren't standing there

38:07

with clipboards from work out passenger

38:09

movements across the network so the

38:11

interchange points are also going to

38:13

start to change we've just separated the

38:15

T 1 and the T 2 line out so that again

38:18

we can deliver trains more effectively

38:21

in from Parramatta and to the people for

38:23

the Greater Western Sydney into into the

38:25

town come the year's end and we so

38:27

that's where the 1,500 additional

38:31

service uplift is coming from by just

38:32

reconfiguring some of those lines and

38:35

those interchange points so

38:36

traditionally there's been about 20 odd

38:38

interchange points across the network

38:40

that will grow the customer experience

38:42

knows interchanges very much as one

38:44

about how people can get through many

38:46

changes more quickly if we can improve

38:47

the customer experience better through

38:49

retail and other avenues so that they

38:52

become destination points central is a

38:55

classic example where there's enormous

38:56

opportunity there

38:59

and we are going to go to market

39:00

eventually in terms of what we can do to

39:02

redevelop an uplift central in the say

39:05

the same way we have seen some of the

39:08

the train stations around the world some

39:09

of those ground stations around the

39:11

world so it's going to be a very

39:12

interesting time as we we move through

39:15

through the technological change but

39:16

also the way in which we do run the

39:18

network okay so how about we have a look

39:21

at some of that technology now so I hope

39:24

you might still might stay here and help

39:26

us out and I'm also hoping that um Tony

39:31

and others come on come on up guys and

39:33

so we often talk about sensing in

39:35

perception and that's like the eyes of

39:37

the system that's where we're taking

39:38

data in the perception algorithms turn

39:41

it into useful information and that then

39:44

goes to another part of the system which

39:46

is what we call the cognition system

39:48

which is like the brain it thinks about

39:50

what's going on given the data it's got

39:52

and then it decides or make some

39:54

decisions about what can be done to

39:57

improve the system and that then is sent

40:00

out to what we call actuation pieces so

40:02

the actuation pieces are the bits that

40:04

actually do something they change

40:07

people's or they change people's

40:11

behavior they influence people's

40:12

behavior potentially by giving them more

40:15

options and in our in the case um in

40:17

this case what we're wanting to do is to

40:19

do that in a way that actually leverages

40:24

the the natural sort of desires and

40:27

instincts of people something else that

40:30

developed during the course of the

40:32

project was that we talked about micro

40:34

systems or the micro system environment

40:36

and the macro system environment and

40:37

you'll hear that a bit in the discussion

40:40

later on but basically what that means

40:41

is that the micro system is like the

40:44

platform in the concourse area or inside

40:46

the sydney trains network and the macro

40:50

system is outside and it's the the

40:52

broader urban environment in which

40:54

people are coming from and going to and

40:56

the reason why that significant is

40:58

because the sorts of information that

41:00

people need changes depending on whether

41:02

they're in the micro system or the macro

41:05

system so in the micro system they've

41:07

made the decision to travel what they

41:09

need is information about the the

41:11

details of what's going on in order to

41:13

make their trip more comfortable and the

41:15

macro system people can change there too

41:19

Asians quite quite considerably they

41:21

might change there at the time of

41:22

journey they might even decide not to

41:25

travel at all so there's lots of

41:27

opportunities for ways to influence the

41:30

system last one here a lot of the tech

41:34

that you're going to see tonight is a

41:37

bit rough around the edges and some of

41:40

it is a bit more sophisticated and it's

41:42

been a bit more polished so when when

41:45

people are developing new technologies

41:48

they go through a set of development

41:50

stages so there's the early discovery

41:51

period where you get planners like me

41:53

who are very broad in conceptual do a

41:55

lot of workshops think about the problem

41:57

and frame it then we hand over usually

42:00

to engineers who start working it up

42:02

into technologies and this is where

42:04

there is more cogent see around the

42:06

ideas and we then start seeing proof of

42:08

concept which is still a bit rough

42:09

around the edges

42:10

pilot prototyping in operations

42:12

evaluations so universities usually do a

42:15

lot of research in that area

42:17

and then we hand it over to a very

42:19

different kind of engineer that then

42:22

polishes it and turns it into something

42:24

that is going to be more like a finished

42:26

project product that can be implemented

42:28

in an actual station so UTS were

42:30

incredibly lucky that we have a unit

42:33

called rapido they're very new and what

42:36

rapido does is it takes the research

42:38

from unruly academics like myself and

42:41

and others and then polishes it and

42:43

turns it into something that is more

42:45

like a finished product so we're able to

42:47

work with industry right along the

42:49

entire development chain when we're

42:52

developing new technologies and and the

42:54

like so with those with that in mind

42:58

what I'd like to do now is invite Alan

43:01

Olimpia vich to to join us so Alan is

43:06

from the Centre for autonomous systems

43:08

and Alan's a specialist in I think it's

43:13

the other one we want to go to Alan's

43:15

our specialist in robotic sensing oh and

43:18

we're seeing some of it now ok so Alan

43:21

can you just talk us through what we're

43:23

seeing so we've got these this group of

43:25

people here now

43:28

we're crowding thank you all right all

43:34

right so we definitely gone beyond

43:36

clipboards we have a 3d sensor we have a

43:45

3d sensor that's actually mounted on

43:46

that pole there which is close to the

43:49

commercial ready type and it's capturing

43:53

a scene which is in front of these train

43:54

doors customers moving about and some of

43:58

the data obviously the sensor itself we

44:01

have developed the possession algorithms

44:02

that are tailored to extract data that

44:04

can be of use for rail service operators

44:08

some of the data is actually portrayed

44:10

here on the screen above we have the

44:14

people appearing in white there like

44:15

white silhouettes on the scene and below

44:18

them are tracks so if you to move if you

44:21

were to move about the there's a track

44:24

history that kind of goes and follows

44:25

you behind behind you so the type of

44:28

information that we can get from this is

44:32

is the amount of passengers that are

44:35

present we can get also the amount of

44:37

people boarding and alighting in time we

44:40

can look at crowding we know someone's

44:43

interfering with the door and we can

44:44

also extract when the doors have opened

44:45

then doors have closed and all the

44:47

associative times around that okay also

44:50

if passengers are just standing at the

44:51

door naturally boarding that can be also

44:53

identified so usually when a train pulls

44:56

into a station most passengers on a Town

45:00

Hall I've noticed are usually standing

45:02

to either side of the the doors aren't

45:04

they okay so sometimes we have Stuart

45:10

Warren from repeat I said he will turn

45:12

kind of the viewer round right okay so

45:14

so we can see who's quite John who's

45:17

there now closest and in in the middle

45:19

of the doors so as a as a rail operator

45:23

you you could get this information about

45:24

whether someone's obstructing the doors

45:26

where people be Jupiter around it and if

45:28

you want to implement a strategy to

45:30

alter how people move about you can have

45:32

a ground truth like you can actually

45:35

compare whether there's been any

45:36

difference in what you've done

45:38

because what I find really interesting

45:40

about this is that you are able to like

45:42

rotate the image and so therefore you're

45:44

able to get these incredibly accurate

45:46

counts on people but you can also see

45:48

the actual behavior like it's almost

45:50

like you can see people's personalities

45:53

but I can't identify who these people

45:57

are as individuals so the we leverage

45:59

the head and shoulder signature we call

46:03

it which is dimensions of each

46:04

individual there they're only to kind of

46:07

do that associational mm-hmm follow up

46:09

follow people around the train doors and

46:11

they allow us to separate people amongst

46:13

the hundred but they're not enough that

46:15

we can then find them on the street or

46:16

anywhere else yeah

46:18

and obviously this information you can't

46:19

then reuse anywhere and find these

46:22

people so it allows a level of privacy

46:25

so we can see people moving and we can

46:28

see where they are in relation to the

46:30

the Train door a bit of cartwheeling

46:34

there bit of air guitar all right and we

46:36

can also see people who are not moving

46:39

as well can't we so we can see people

46:41

who might be a bit obstructing so we can

46:43

have a side view and you'll be able to

46:44

see us hopefully three people here

46:48

courses ok so the trade you you look

46:50

like you're being a bit obstructive

46:52

there how are they my goodness all right

47:02

no no ok this is getting really bad I

47:05

think we need some help here in

47:08

particular I think we need some customer

47:10

service attendants so

47:15

[Music]

47:19

we definitely need to sort this unruly

47:22

mob out all right so what do things all

47:51

righty all right wonderful okay

48:18

you can come out now all righty

48:23

so Simon can I ask you like you can see

48:26

this stuff now I think this is the first

48:28

time that you've seen this and we were

48:30

speaking the other day about all this so

48:33

if you had more information about where

48:36

people were located and what was

48:38

happening outside train doors would that

48:41

be useful to you as a as an operator

48:43

absolutely so my job is to look after

48:47

one door I can't see all the doors along

48:51

the train all 16 doors so if we have a

48:53

technology here where we could spread

48:54

people across the platform much more

48:57

evenly to use all available doors and

48:59

encourage them from using some sort of

49:01

app on our phone and move them along

49:05

much better so it's sort of like giving

49:08

you eyes that can see further than what

49:10

you can with your own eyes so far and

49:15

there's always lots of customers in

49:16

between but if I could see something

49:18

that's gonna tell me to encourage others

49:20

to move on I would try and encourage

49:22

them hmm okay thanks a lot for that

49:25

thank Simon and thank you Andrew all

49:27

right

49:30

okay so I'd now like to ask Mike ailing

49:34

from downer to join us Mike's the

49:37

general manager of intimate of

49:39

innovation at downer and you're kind of

49:41

responsible for a lot of these this

49:43

stuff that's going on so we've had it's

49:48

been great working with Sydney trains

49:50

but we've also been lucky to have

49:51

another project with down a rail that's

49:53

also been supported by the round

49:55

manufacturing CRC and so you're much

50:00

further along that that development

50:02

paths so I was just wanting to check in

50:06

with you to ask but where where are you

50:07

up to because I'm actually out of the

50:09

project now cuz I'm only do the early if

50:12

I don't do the later stuff so what

50:13

what's happening I think you had a

50:15

earlier slide so we're now moving into

50:17

the commercialization phase of this

50:18

project so taking all the great research

50:21

that you guys have done and that one's

50:22

done and turning all the other research

50:25

and all the all the algorithms turning

50:28

that into some survival data so it would

50:31

we're in the commercialization phase and

50:33

again it's great that we're actually

50:35

using the rapido the UTS function to

50:38

help us commercialize that so can't

50:41

think things that we're doing at the

50:42

moment is to two aspects will be the

50:45

hardware so they they'll see the cameras

50:47

hopefully we're putting a cover around

50:49

it because it's gonna be we need to

50:51

protect it so they're doing a whole lot

50:52

of covers and trying to turn that into a

50:55

manufacturable product so obviously

50:58

we're gonna need quite a few of them for

50:59

the platforms so we're doing all that

51:01

and also looking at the the sensing

51:05

cameras are making sure they work in you

51:07

know or normal environments obviously

51:09

underground and also outside as well

51:11

because sometimes you get your issues

51:13

with sun glare and other things so we're

51:15

working through that then there's a

51:17

software aspect to it which is turning

51:20

all that data and all those blobs are

51:22

one of the bigger blobs moving around

51:25

and turning that into verbal information

51:28

and we look we've got two use cases one

51:29

you know people like Simon we're turning

51:32

that into a product that you can have on

51:34

a device on the platform that can then

51:36

tell you where the density of that those

51:39

passengers are

51:40

and also linking that with some of the

51:43

data we taking off the trains to tell

51:44

you when the next trains coming in which

51:46

switch carriages are full which ones are

51:49

half full and which ones are empty so

51:50

you can try and distribute passengers

51:52

along they along along the platforms so

51:54

strongly about that minute and then the

51:57

second part is really then from an

51:59

operation center perspective turning

52:01

that into you know metrics and some some

52:04

dashboards about you know trends dwell

52:06

time trends that type things so pretty

52:09

exciting we're turning air into a

52:11

minimal Viable Product and that's what

52:13

we're working with at the moment I still

52:18

don't know I mean I haven't caught up

52:20

with this for a while it's it's it's

52:22

been it's been an interesting journey

52:24

because it this is now getting here for

52:25

master perspective the exciting bit

52:27

where we can turn that into a viable

52:28

product and lovely seeing that research

52:31

being turned into something that can be

52:32

utilized and you're working with them

52:35

steward and the guys we're hoping to get

52:38

something out sort of February next year

52:40

the first minimal Viable Product and

52:42

then to start launching it probably in

52:44

March April next year so that's the plan

52:47

that is actually quicker than I thought

52:48

there you go I'm genuinely quite someone

52:51

that wasn't in the script okay okay so

52:55

do you want to add anything else no

52:56

that's good all righty okay so thanks

53:00

Mike so what I'm going to do now is

53:02

we're now going to click forward to

53:05

another part of the robotic system we're

53:08

now going to start looking at a bit of

53:11

the actuation and I'm going to ask dr.

53:14

Nathan Koechner to join us Nathan's an

53:17

adjunct professor at the Institute for

53:19

sustainable futures where I'm also

53:21

located and I think what we might do

53:24

guys is if we can click back to the the

53:28

sensor that would be that would be

53:30

really good so Nathan what is this walk

53:38

us through this um so like many of us

53:40

said and I can't Anthony said it's all

53:42

about trying to change customer behavior

53:44

that can take many shapes and forms this

53:47

is just a really nice visual way to

53:48

communicate that idea so this is

53:51

actually a digital sign

53:54

built into the sensing system we've seen

53:57

before and built into our actual Waratah

53:59

door so you can imagine a system like

54:02

this in this particular case where maybe

54:04

I'm somewhere I shouldn't be in front of

54:06

the doors and I start to get the message

54:08

don't be there or maybe when the the

54:11

trains ready for me to go on it turns to

54:14

a green arrow or something that

54:16

encouraged me in the idea is just to

54:18

plug this in to normal operations but

54:21

like I started with it's just one of

54:23

many different possibilities this is

54:26

sort of like a little like a subtle cue

54:28

and we've used the term nudge quite a

54:30

bit so this is a sort of is very rough

54:32

and ready but this is the the

54:33

conceptually this is what we're looking

54:35

at isn't it so the best examples of us

54:37

using this are the hardest to shine room

54:39

like this so we've run our life in

54:41

Sidney trains in Town Hall I recently so

54:44

in the interface between Kiev in town

54:46

hall and Town Hall our kite and hole

54:49

obviously it's just a really soft gentle

54:53

note on the shoulder or you think about

54:55

it if you're walking along and you get

54:56

to a decision point and an hour goes

54:59

this way just as you appear just as you

55:01

get into the field of influence then you

55:04

know the one in ten person fears a

55:06

little bit to the left the follow that

55:08

by the time they go 20 meters down the

55:10

track maybe there are despite of stairs

55:11

and sort of that flight of stairs if

55:13

we've got a real example like at all

55:15

does where are stairs 10 if we have more

55:18

people using this entrance there would

55:20

be less problem on the concourse then a

55:22

bit of technology to do that Woodhull

55:24

the really interesting thing is when we

55:27

do it in places like Town Hall no one

55:30

notices no one realizes nor complains

55:32

about it no it says what happens there

55:34

were you conscious the numbers they did

55:35

it so they're going there I didn't said

55:38

I don't know what you're talking about

55:39

coach the numbers and they all went the

55:42

way we told them to go so we really have

55:44

hit that level of being of the influence

55:46

on the surreptitious level to get a

55:49

desired outcome without our overloading

55:52

so it does work

55:54

unfortunately it works unfortunate

55:55

because I'm research everyone things

55:57

work we have to do extra work it's

55:58

better when it's a good idea it doesn't

55:59

actually work

56:02

excellent good one okay so I mean

56:06

another couple of quick question so this

56:09

is I mean I think this is a kind of a

56:11

crude thing but it's sort of your green

56:13

and red we all understand that what I

56:15

mean there's this form but what other

56:18

sorts of forms could actuate in like

56:20

this come from because we talked about

56:21

nudging and it's supposedly subtle and

56:23

cue and it's I've stopped the bursting

56:26

it's crude I'm research of the very

56:28

fuzzy blue end of the spectrum just to

56:29

get an idea out there and they doing

56:31

someone like Mike jumped on boys down to

56:33

make trains and I started and built

56:35

whole train doors it wouldn't look like

56:36

that they would still work hard to

56:39

answer your actual real question it

56:41

doesn't have to be a lot it could be

56:43

ticket gates it could be barriers it

56:45

could be things on the floor it could be

56:47

the waste doors work I think the best

56:50

example for this is if anyone seen a

56:52

Pixar animation or a Disney animation

56:54

they can make a broomstick have life so

56:57

anything that can move and tell a story

56:59

and where people will respond to stories

57:02

if we're going along and the thing

57:03

that's wiggling tells us the story of

57:04

don't be here then we don't be there so

57:08

that that's that raises another point

57:10

because I mean it's good that we've got

57:11

the senses up so the the other thing

57:14

that you've often talked about how I'm

57:15

um

57:16

the other thing that you've talked about

57:17

a lot is that this type of actuation it

57:20

doesn't work if it's just going all the

57:22

time so you it's got to be subtle and

57:25

and then you run it through that and the

57:27

relationship in particular between what

57:29

why the senses are really important to

57:31

be connected to and the connections they

57:33

have with the actuation so it's the

57:36

classic we all know and we're all

57:38

fortunate because we are people and

57:39

we've gone through our lives being one

57:40

if you you enter an environment this is

57:43

constant noise you just ignore but if

57:46

you're standing somewhere and something

57:47

falls over you just instinctively go oh

57:49

it wasn't me so it's that little bit of

57:52

the information being contingent to you

57:55

having done something that means you pay

57:57

attention to information that really

57:58

meant something that's sign change

58:00

because I did something that fish moved

58:02

because I approached it it means you pay

58:04

attention to it

58:05

simplest way is more people looking at

58:07

the information the more people are

58:08

going to follow it the more people will

58:10

thinks that informations personalized or

58:12

individualized to them the more

58:13

we'll take it on board then it's a

58:15

matter of knowing what's happening in

58:18

the world in real time and they are the

58:20

connector to if we know where the people

58:22

are we know when they entered we know

58:23

when we want to tell them the

58:24

information a bit like Simon and you can

58:26

do if they can see what's going on they

58:28

know when to tell customers things then

58:30

we told them the information they need

58:31

to know then and only then which means

58:34

they instinctively listen to it take it

58:37

on board rather than just ignoring it as

58:38

yet another sign another exit sign we've

58:41

seen them everywhere so why do you here

58:42

there's one last thing I just want to

58:44

quickly talk about is the difference

58:46

between a robotic system and what you've

58:49

called a device mesh so I can say I get

58:54

what you're saying here so if I'm here

58:55

for example and I'm being naughty and

58:59

that that's then telling me that I'm

59:01

being naughty so there's no really

59:04

elaborate cognition system involved

59:06

there is there no so it's really a

59:08

continuum in a way if we had a widget

59:12

maybe it's fear of influence or

59:14

connectivity is only this big so maybe

59:16

it's a proximity kind of sensor but as

59:19

soon as that has a bit more of

59:20

intelligence like maybe it likes me so

59:22

let's mean here but doesn't like you so

59:23

it doesn't like you there that maybe

59:25

becomes a device the device mesh is the

59:28

concept of that's a useful device in and

59:31

of itself but it also connects to other

59:33

devices around so you imagine here if

59:36

this device were saying too crowded

59:38

everyone get away and its sister devices

59:40

at the top of the stairs saying don't go

59:42

down these stairs it's a bad entrance

59:44

it's busy down there go somewhere else

59:45

instead

59:46

then each sister device was saying at

59:48

entrance point the Concours don't go

59:50

this way go down the other way because

59:51

it's too crowded and there that's when

59:53

we start to stuck the devices into the

59:54

mesh start to get or blow the line

59:57

between the micro system micro

60:00

activation system and the macro a

60:02

Croatian system we can imagine we can

60:04

actually step over the precinct boundary

60:07

with the device mesh transition from

60:10

maybe something in a bowl ad or the

60:13

concourse to something on a mobile phone

60:15

or something on our news boards okay so

60:19

but but the big difference is that it

60:21

they're these individual things and

60:24

they're doing a job they helping the

60:25

situation make

60:26

it better by theirself they're

60:29

autonomous agents I can't be useful in

60:31

and of itself if it gets disconnected or

60:33

I fight it it still has viable valuable

60:35

function they're more powerful together

60:38

sure but they're very useful by themself

60:40

okay back in scale okay I hear what

60:43

you're saying

60:44

all right well that's good I think what

60:49

we need to do now is just look at a very

60:51

different type of sensing which could

60:53

lead to some other actuation devices but

60:57

thank you very much for that all right

60:59

so what I'd like to do now is ask

61:01

professor Sean hey to join us so Sean is

61:06

from the global Big Data technology

61:08

center and Sean's going to show us a

61:11

different form of sensing and I guess

61:14

the the main difference here short is

61:16

that what you and the folks have been

61:18

doing that have been working on these

61:20

these sensors and we might go over to

61:23

this other almost sensing so it's the

61:26

different camera now here we are yeah

61:29

okay so this is this is very different

61:31

and we're more at the you're more at the

61:32

proof of concept stage rather than the

61:34

product ization it don't you yeah agent

61:36

Lee Thank You Marshall

61:38

we want to actually to take the least

61:42

cost planning approach and see whether

61:45

there was any existing infrastructure to

61:48

use and a lot of problems or one of

61:51

their problems kept coming out from our

61:54

workshop and also from today is the

61:57

congestion so we have find ways and you

62:02

know being looking at ways to create

62:05

congestion Hitmen and they can till the

62:09

congestion levels and in your time at

62:12

the platform at concours at anyway hmm

62:16

okay so I've forgotten all right okay so

62:20

did you want us to go and maybe get some

62:23

more volunteers to show how these work

62:25

or did you want to run us through some

62:26

of the IDs yeah don't this work is many

62:31

done by warm by PhD student called

62:34

hungry who is here and yes so we just

62:37

want to tell the ideas

62:39

I'm going to show you our real system

62:41

actually implemented and in this system

62:45

you can see we have camera there

62:47

so pretending this one is the CCTV

62:49

camera so the idea is okay so a CCTV is

62:53

to send people okay within the green box

62:55

here and then when we have more and more

62:57

people coming coming up here so

62:59

hopefully we can build up the

63:01

conjunction map or the Hitmen and then

63:04

when the number of people reach the

63:07

threshold okay so we set a special to be

63:10

8 right so and when the number of

63:12

people's reach the stretcher

63:13

then we will create the warnings you

63:16

know or I saw which you can hear so

63:18

hopefully you know we can say see the

63:21

right hand side the blue one okay stick

63:23

and told him it so at this moment we

63:24

don't have many people we have one image

63:26

showing myself here so it won't seem

63:28

very heavy so yeah so being would be

63:32

nice you know if we can have more from

63:34

the audience okay guys we can greatly

63:36

build up the heat map yeah so as you can

63:39

see right so on yeah so we want people

63:41

to be in the green box alright so and

63:43

then now at this moment this technique

63:45

is more or less based on their face no

63:46

condition so when the number of people

63:49

reach age okay so we hear the voice okay

63:51

so say cloudy cloudy then we would like

63:53

to have people move out of the outside

63:56

black line and then the number of the

63:58

people will reduce to zero at this

64:00

moment right so we hate yes so what so

64:04

this is the whole idea so I'm you know

64:06

so we can be until when you know the

64:09

location okay so it's crowded and you

64:12

know we can make kind of advice okay to

64:16

tell people a way to move and you know

64:18

to let say for simple place or kind of

64:20

thing okay so I mean one of the things a

64:23

lot of us here who were transport people

64:25

we know about and think about is level

64:27

of service criteria and there's levels

64:29

and measures for that so this system

64:32

could basically give a live level of

64:36

service measure um by looking at the

64:39

densities over different areas you

64:41

exactly got this one X is very different

64:44

from the one just mentioned by Ellen

64:46

them might okay so we are not using 3d

64:49

census so all we are using here is 2d

64:53

with 2d sensor we can only get to the

64:56

threat visual kind of pictures or images

65:01

so will not be able to easily tell the

65:05

location of each individual person I saw

65:07

as you can see you know we can still

65:10

love detail where people saw so it

65:12

doesn't mean we cannot do it or I saw

65:14

just mean motive code yeah so for that

65:16

big group behavior yep this this is good

65:19

enough for some some aspects of it more

65:22

this one this is webcam yeah it's good

65:24

enough for now so I'm for the Lea system

65:26

we are using the existing infrastructure

65:28

CCTV cameras at the local station for

65:31

the for the testing so we are using

65:33

machine learning technique so which one

65:35

condition the audience I can so we wish

65:39

the CCTV cameras we have a wider view

65:41

and then we can take videos of many

65:45

people about 100-200 people's in the

65:48

scene and then we can still estimate

65:50

number of people's on the platform on

65:53

any areas okay so in the change station

65:56

so do you want to do you want to take us

65:58

through some of those yeah yeah before

66:04

showing him our CCTV base and density

66:07

estimation human density estimation and

66:10

also DeMuth I need to show one of the

66:12

work done by another team okay in our in

66:16

your buffer Cote so which is led by

66:18

Professor watch a mouse so this way is

66:21

based on Wi-Fi okay not it in CCTV so as

66:24

you can see okay based on Wi-Fi I'm the

66:26

thing has already also created a heat

66:28

map and this already taking into account

66:31

obviously CCTV data as well although

66:33

Hema credit using our CCTV data and you

66:39

know so se concedes and you know Wi-Fi

66:42

actually can not actually tell the

66:44

number of people because not everybody

66:46

would turn on the mobile phone so I can

66:49

I see her and so audience here probably

66:51

20 percent 30 percent of people thing on

66:53

the mobile so yeah we can only detect

66:56

people who already think of the cenotes

66:59

so but nevertheless okay so probably

67:01

this is not very clearly seen so using

67:04

our TTS ect

67:06

the camera okay based on one single CCTV

67:08

camera this is the concourse areas in a

67:10

different station so in the middle one

67:12

you can see ok the kind of Testament

67:15

which is not him

67:16

the Rohan sized him at the dimension is

67:18

more look more like the Unreal dimension

67:21

of the of the concourse but one camera

67:25

may not cover the whole whole area so in

67:27

a different station platforms 2 or 3 & 3

67:31

okay you can see one camera see if I can

67:34

only cover okay so about 1/6 of the

67:37

areas in a dear and open station so

67:39

that's why we need to combine the view

67:41

of different cameras together

67:43

so you see c9 and CA and c7 c9 see if I

67:47

see now basically you know back to back

67:50

ok next to each other one covering

67:52

platform to c9 and see if I come in and

67:55

perform part of platform 3 and then we

67:58

have c8 ok so we've been so covering

68:01

more part of the platform's and c7 okay

68:05

for viewing in different directions and

68:07

then you can see the combined result at

68:10

the at the bottom so this work Orchestra

68:12

sister fossa plot we have two different

68:14

approaches actually this approach is

68:16

many done by my pasty skin and Helia so

68:19

who is also here so I'm going to show

68:21

you actually another one you and the

68:24

audience so I'm done another wise more

68:26

intelligent approach so this way is

68:29

based on my tip Nani okay so everybody

68:31

now no did noni okay so it's very smart

68:34

if they're also very actually technique

68:36

so from this one so you also see the

68:38

combined result ok of the different

68:42

station for platforms on 2 and 3 and

68:45

then in the middle part you see the him

68:47

Ezra alright so in the bottom chart okay

68:50

you see the accuracy so the blue ones

68:52

telling the estimated number okay and

68:56

and the real one I think is the no the

68:58

blue one provides the clown shoes and

69:01

the red one STM is the estimate number

69:04

so you can see the number go from over a

69:07

hundred to to a few people

69:10

okay on the platform we can actually

69:12

model s quite actually estimate the

69:15

number so that

69:17

the difference is not very much on

69:19

average so you're getting counts and you

69:21

get reading relative positions as well

69:23

yeah we can also get a little position

69:26

but if you look at the heat map right so

69:28

also we can see okay total the areas you

69:32

can see the red color okay that tail is

69:34

quite quite kind of in a crowded area

69:37

over there okay on upper floor so if you

69:40

see the blue areas that mean nobody over

69:42

there okay like you know the view on c7

69:46

okay at this moment we see we see nobody

69:48

okay yep so there's still some more to

69:51

go with this do how to apply it

69:53

yes so we need updates more because we

69:56

record this data only in July sometimes

69:59

so using deep learning we need to have a

70:01

lot of data to you know to collect and

70:04

then name for the training so then we

70:06

can have more accurate result at small

70:08

metal result is quite present already so

70:11

we you know we still be able to to see

70:13

more a lot more accurate result and then

70:17

you know and then we go to the

70:18

application so for sure yeah okay thanks

70:21

for that

70:21

thank you much Elaine thank you yeah all

70:23

right okay so um oh yeah the clicker

70:27

I've got it back alright so what I'd

70:31

like to do now is ask dr. Chanyeol you

70:35

to come and join us so I wanted to ask

70:39

you a channel what are your thoughts

70:41

about connecting all of these things so

70:44

Nathan that I had a quick discussion

70:45

about the device mesh where it's all

70:47

relatively unconnected but you're

70:49

looking at cognition which is about

70:51

collecting or connecting those things so

70:53

what tell us a bit about that I guess

70:55

I'll start from the very top view so

70:57

with this all new technology now we know

71:01

more about the micro level passenger

71:03

flow and which means that that we now

71:06

know where people are and where they are

71:08

moving to in real-time and there's one

71:11

others also be information that we

71:12

actually have to talk about which is the

71:14

Opel cars

71:16

with this new Opel Adam system that we

71:20

there's recently implemented we now can

71:22

track the macro-level passenger flow

71:24

from stations to stations in this video

71:28

what we see is the

71:30

passion TM crowdedness of each station

71:33

and how many people are actually moving

71:35

from one station to the other so each

71:38

circle represents station and the size

71:40

in the color represents how crowded and

71:43

how many people there are and the lines

71:45

represents like each people traveling

71:47

between the stations so yes yes yes I

71:55

was hoping you wouldn't do this to us I

71:57

sort of met I'm just a planner I can't I

72:02

can't okay I can't cope okay okay so now

72:10

the question now the question is how do

72:12

we utilize how do we fully utilize all

72:15

this information to improve customer

72:17

experience and in order to do so we

72:19

should be able to digest all this

72:21

information and then understand in our

72:24

brain like how everything works

72:26

and my question is can we actually do

72:28

this can we actually under can we

72:30

actually digest the oldest information

72:31

as human and then come up with something

72:34

smart like smart plan that optimizes

72:36

like a network throughput or something

72:38

so in this scoping study we started off

72:41

by mathematically formulating the

72:44

passenger flow for each station so here

72:46

our station is decomposed into different

72:50

parts so here we have concourse

72:51

platforms and gig education stuff and

72:54

then we describe the relationship

72:57

between each part of this station so I

73:01

still think the maths is just too much

73:03

because I don't want to know this is

73:05

exactly what I expected so this shows

73:07

you that the system we are dealing with

73:09

is very sophisticated it is very complex

73:12

and most of us most of us including

73:14

myself I mean I wrote this but yeah I

73:16

don't really understand like exactly

73:18

what is going on so this is the point so

73:20

which means that this cognition problem

73:22

we are trying to solve is very

73:24

non-trivial and it's very hard to solve

73:26

so as human I don't know how to solve

73:29

this for now not yet not yet okay so

73:32

let's take two station views so let's

73:35

say this is actually put some numbers

73:40

but nothing so we have two stations here

73:43

and now we can like connects make some

73:46

like like to draw some lines to

73:47

represent the dependencies between two

73:50

stations like how each component in a

73:52

station is interconnected with the other

73:55

part in other station so this is simple

73:57

enough this is what happens if we have

74:00

ten stations so it is impossible for

74:02

human as a human to understand each bit

74:05

and what is going on as a like a whole

74:08

so but but the problem is we have more

74:11

than 200 stations in New South Wales and

74:13

then I actually try to put like 50

74:15

stations but it was impossible

74:16

everything was black so I kind of get

74:20

this one that I said to get the vibe

74:22

with this well I couldn't read the math

74:24

yes so if you just look at like one by

74:26

one that you might understand but if you

74:29

have to understand the ant dependence

74:30

between every single variable that is

74:33

the point this so if this one

74:34

illustrates along with this some model

74:37

that we have to actually understand

74:38

everything in order to make some smart

74:41

decision yeah so it is impossible for us

74:46

to understand everything so and what is

74:48

even harder is to actually come up with

74:50

some smart decision so instead of

74:54

instead of having human to do the two T

74:58

to do these kind of jobs we would like

75:00

to have a like a smart cognition system

75:02

that works like an Oracle so we would

75:05

like a system like this one - OH

75:10

to oversee the whole network the system

75:13

that gets the real data at the the real

75:18

time data from CCTV and Wi-Fi and all

75:20

these sensing things and then make a

75:22

smart decision based on this information

75:24

and we also want this system to forecast

75:28

and predict what's what is likely to

75:31

happen and act before that happens

75:33

rather than simply reacting to what is

75:35

locally happening so if I'm if I'm

75:40

hearing what you're saying it's sort of

75:41

like with the Apple car data that you

75:44

were talking about earlier it might be

75:46

the case that if you've got a really

75:48

smart cognition system in

75:51

there might be something that goes on

75:52

it's a Strathfield station two hours in

75:55

the past that is then going to affect

75:58

what goes on at townhall station and the

76:01

system has actually learned how to

76:03

recognize yes so based on this Opel data

76:06

which which we accumulate over time we

76:08

can learn how like how things act

76:13

differently based on what we have

76:15

currently and based on this real-time

76:17

data we can act we can predict what's

76:19

going to happen and act in advance or to

76:22

find like a better solution that

76:24

optimizes that resolves this situation

76:27

kind of thing so so in the scoping study

76:31

we have identified like three most

76:33

important problems that we think is like

76:35

critical as part of this cognition

76:37

system so first is the interchange

76:39

problem if a passenger is traveling from

76:42

one station to the other where how do we

76:45

estimate where this passenger is gonna

76:47

transfer like at which station a

76:49

forecasting problem if we see a large

76:52

crowd of people entering a station at a

76:54

certain time how would it influence the

76:56

rest of the network and the lastly I

76:59

think this is really important problem

77:00

planning problem if we alter the

77:03

timetable how would it affect the

77:05

passenger flow and ultimately can we

77:08

find an optimal timetable that maximizes

77:11

the network throughput as well as the

77:13

passenger satisfaction okay I think

77:19

that's about it yeah I hope it is

77:21

there's no more mess

77:31

wherever you are and wherever you go in

77:34

the world may you always travel well

77:35

thank you very much

77:37

[Applause]