This activity focuses on the design, development and implementation of robotic systems for research on control, automation and machine intelligence. Mobile robot platforms and robotic manipulators have been designed and constructed, and research is underway to enhance the sensing and control capabilities for obtaining increased accuracy and autonomy. Control algorithms have been proposed for effecting trajectory tracking of mobile robots in conjunction with obstacle avoidance. In addition, novel neural network based control algorithms have been proposed for stochastic-adaptive dynamic control of mobile robots. In this work the controller estimates the nonlinear dynamics of the robot in real-time, so that it can maneuver successfully even in situations when these are uncertain or time-varying. In contrast to other adaptive controllers hitherto proposed for mobile robots, these algorithms fuse estimation and control via the concept of dual-adaptive control. This results in an improved control performance. These algorithms have been validated on a real mobile robot, and are being implemented on a robotic manipulator in our control laboratory.
10 December 2013