This paper describes the design, implementation, and machine learning issues associated with developing a control system for a serpentine robotic manipulator. The purpose of the control system is to provide autonomous/teleoperative control of the serpentine robotic manipulator, as well as full robotic control during operation of the manipulator within an enclosed environment such as an underground storage tank. The controller algorithms make use of both low-level joint angle control employing force/position feedback constraints, and high-level coordinated control of end-effector positioning. Since the inverse kinematics solutions for a hyper-redundant serpentine manipulator are extremely difficult if not impossible to obtain using standard techniques, a variety of methods from evolutionary computation were employed to arrive at a suitable inverse kinematics model for the manipulator. Additional constraints, such as various joint angle bending restrictions, joint constraints within a given stage of the manipulator, and customer requirements may be easily incorporated into the machine learning process through the fitness function. This approach has resulted in a system which offers both high-level full robotic control and low-level telerobotic control modes, and provides a high level of dexterity for the operator.
展开▼
