Adaptive locomotion control system for robots with arbitrarily modular design

Abstract

The paper proposes an adaptive control system for modular hyper-redundant systems that can learn to solve the control problem of robots with arbitrary designs from a given class. The proposed model uses logical-probabilistic knowledge discovery methods to find effective control patterns in an array of system’s environment interaction statistical data. For making the system independent on the chosen robot design, it was proposed to include design information in the training data. Using this information during the training process allows the control system to tune in to control the robot, regardless of its design. The effectiveness of the approach is demonstrated by the example of training virtual models of robots to move forward.