Haider, S. & Raza, S.A. 2015, 'Complexity Reduction of Influence Nets Using Arc Removal', Journal of Intelligent and Fuzzy Systems, vol. 28, no. 4, pp. 1849-1859.View/Download from: UTS OPUS or Publisher's site
The model building of Influence Nets, a special instance of Bayesian belief networks, is a time-consuming and labor-intensive task. No formal process exists that decision makers/system analyst, who are typically not familiar with the underlying theory and assumptions of belief networks, can use to build concise and easy-to-interpret models. In many cases, the developed model is extremely dense, that is, it has a very high link-to-node ratio. The complexity of a network makes the already intractable task of belief updating more difficult. The problem is further intensified in dynamic domains where the structure of the built model is repeated for multiple time-slices. It is, therefore, desirable to do a post-processing of the developed models and to remove arcs having a negligible influence on the variable(s) of interests. The paper applies sensitivity of arc analysis to identify arcs that can be removed from an Influence Net without having a significant impact on its inferencing capability. A metric is suggested to gauge changes in the joint distribution of variables before and after the arc removal process. The results are benchmarked against the KL divergence metric. An empirical study based on several real Influence Nets is conducted to test the performance of the sensitivity of arc analysis in reducing the model complexity of an Influence Net without causing a significant change in its joint probability distribution.
Raza, S., Haider, S. & Williams, M. 2013, 'Robot reasoning using first order bayesian networks', Lecture Notes in Computer Science, International Symposium on Integrated Uncertainty in Knowledge Modelling and Decision Making, Springer, pp. 1-12.View/Download from: UTS OPUS or Publisher's site
This study presents the application of first-order Bayesian Networks (FOBN) to model and reason in domains with complex relational and rich probabilistic structures. The FOBN framework used in this study is 'multi-entity Bayesian networks (MEBN). MEBN ha
Raza, S., Haider, S. & Williams, M. 2012, 'Teaching coordinated strategies to soccer robots via imitation', 2012 IEEE International Conference on Robotics and Biomimetics, ROBIO 2012 - Conference Digest, IEEE International Conference on Robotics and Biomimetics, IEEE, Guangzhou, China, pp. 1434-1439.View/Download from: UTS OPUS or Publisher's site
Developing coordination among multiple agents and enabling them to exhibit teamwork is a challenging yet exciting task that can benefit many of the complex real-life problems. This research uses imitation to learn collaborative strategies for a team of agents. Imitation based learning involves learning from an expert by observing him/her demonstrating a task and then replicating it. The key idea is to involve multiple human experts during demonstration to teach autonomous agents how to work in coordination. The effectiveness of the proposed methodology has been assessed in a goal defending scenario of the RoboCup Soccer Simulation 3D league. The process involves multiple human demonstrators controlling soccer agents via game controllers and demonstrating them how to play soccer in coordination. The data gathered during this phase is used as training data to learn a classification model which is later used by the soccer agents to make autonomous decisions during actual matches. Different performance evaluation metrics are derived to compare the performance of imitating agent with that of the human-driven agent and hand-coded (if-then-else rules) agent.
Stanton, C.J., Ratanasena, E., Haider, S. & Williams, M. 2012, 'Perceiving forces, bumps, and touches from proprioceptive expectations', Lecture Notes in Computer Science, Robot Soccer World Cup, Springer, pp. 377-388.View/Download from: UTS OPUS or Publisher's site
We present a method for enabling an Aldebaran Nao humanoid robot to perceive bumps and touches caused by physical contact forces. Dedicated touch, tactile or force sensors are not used. Instead, our approach involves the robot learning from experience to generate a proprioceptive motor sensory expectation from recent motor position commands. Training involves collecting data from the robot characterised by the absence of the impacts we wish to detect, to establish an expectation of normal motor sensory experience. After learning, the perception of any unexpected force is achieved by the comparison of predicted motor sensor values with sensed motor values for each DOF on the robot. We demonstrate our approach allows the robot to reliably detect small (and also large) impacts upon the robot (at each individual joint servo motor) with high, but also varying, degrees of sensitivity for different parts of the body. We discuss current and possible applications for robots that can develop and exploit proprioceptive expectations during physical interaction with the world
Haider, S., Abidi, S.S. & Williams, M. 2012, 'On evolving a dynamic bipedal walk using Partial Fourier Series', 2012 IEEE International Conference on Robotics and Biomimetics, ROBIO 2012 - Conference Digest, IEEE International Conference on Robotics and Biomimetics, IEEE, Guangzhou, China, pp. 8-13.View/Download from: UTS OPUS or Publisher's site
The paper presents a Partial Fourier Series (PFS) based bipedal gait in sagittal and transverse planes. The parameters of the Fourier series are optimized through Evolutionary Algorithms (EA). In addition to evolving the two walks (forward and turn) separately, the paper demonstrates how the combination of the two enables a dynamic and adjustable walk. The stability of the walk is ensured through an effective use of the built-in gyroscope sensor. The evolved walk has been tested on the simulated version of the humanoid Nao robot and is being used within the RoboCup Soccer 3D Simulation competition