Learning and adaptation of sensory perception models in robotic systems

Models of perception are an important element in the control of sensory perception in autonomous systems. The performance of a perception controller will depend on how well the models reflect the time-varying performance characteristics of sensors and data processing algorithms. A novel approach to achieving high quality models through real-time adaptation is presented. Models reflecting observation uncertainty are adapted in accordance with online sensor performance using a radial basis function approach modified to allow real-time operation