Researchers: Xiaoming Hu, B. Ghosh (Washington University) and C.F. Martin (Texas Tech. Univ.).
Sponsor: The Swedish Research Council (VR).
Complex control systems such as autonomous mobile systems are in general multi-behavior control systems. In many applications, they also need to operate in a team. In such cases, they are also multi-agent systems. An integral part in the design and operation of autonomous systems is control and sensing. Both are difficult problems in a realistic environment and for a realistic mobile system. Viewing such an embedded system as a set of concurrent sensor-actuator subsystems leads to the conclusion that sensory data processing forms a dual and equally important part as control actuation. For mobile systems in unknown environments, the sensing problems tend to be particularly challenging, for example, the problem of observing and mapping the structure of a dynamic environment.
To summarize, in this project we would study several issues on the integration of sensing, modeling and control. They include: model abstraction for control and sensing, active sensing, sensor fusion for input reconstruction, simultaneous localization and mapping in a continuous environment, control of multi-agent and/or multi-behavioral systems. Methods proposed here are strongly motivated from systems and control theory.