Autonomous coordination of locomotion and collision avoidance in robotic mechanisms

A robotic mechanism for locomotion along rigid aerial lines and reticulated structures is presented. A biomimetic method for obstacle avoidance is also introduced The computation of collision-free trajectories requires the analytical description of the physical structure of the environment and the solution of the inverse kinematic equations. For dynamic environments with unknown obstacles, however, it is extremely difficult to get real-time collision avoidance by means of analytical techniques. The main advantage of the proposed method resides in its departing from the analytical approach, as it does not employ formal descriptions of the locations and shape of the obstacles, nor does it solve the kinematic equations of the mechanism. Instead, the method follows the perception-reason-action cycle and is based on a reinforcement learning process guided by perceptual feedback. From this perspective, obstacle avoidance is modeled as a multi-objective optimization process. Finally, the concept of locomotion strategies and their dynamic coordination have been introduced as a second level structure in the reinforcement learning process. The dynamic, perceptual-based coordination of different locomotion strategies allows the robot to perform complex navigation maneuvers.