This paper focuses on obstacle-negotiation planning and motion control for the brachiation inspection robot and studies the key problems such as mechanical design, control system developing, kinematic analysis, and behavior planning. The main contributions of the paper are shown as follows: Firstly, the previous research work of inspection robot is introduced. And the research evolvement of fuzzy control and genetic programming algorithm is presented. The mechanical architecture, system function and control system of some typical inspection robots are illustrated. Furthermore, the key technical dificult points during the design of brachiation inspection robot are analyzed. Secondly, the paper detailedly illustrates the mechanical architecture design, control system structure, sensor design, hardware architecture, and software structure of the brachiation inspection robot. Thirdly, the obstacle-negotiation control system based on the structure of hierarchical behavior controller is presented. The action sequences for all kinds of obstacles are analyzed. And the behavior controller, the behavior coordinator and the evaluator are designed in the obstacle-negotiation control system. Furthermore, the kinematic analysis of brachiation inspection robot is considered. Fourthly, the paper establishes the simplified dynamic models of the promotion behavior and the turning behavior for the brachiation inspection robot. Furthermore, the key control problems for the above two behaviors are analyzed.By the method of fuzzy control, the behavior controllers for the promotion behavior and the turning behavior are designed. The obstacle-negotiation experiment results show the validity of the algorithmic research on the behavior controller. And the other behavior controllers of brachiation inspection robot are also discussed. Fifthly, the paper presents the behavior coordinators based on such obstacles as counterweights, strain clamps and suspension clamps. By the mapping method of configuration space and the algorithm of map searching based on genetic programming, the optimized route of inspection robot is obtained. And the route denotes a node sequence in the configuration space. After the coordinate descriptions of these nodes in the sequence are conversely mapped into the physics space, the sequence of position parameters can be acquired in the physics space. And according to the parameter sequence, the inspection robot can realize the control from the initial point to the target point. The obstacle-negotiation experiment results show the validity of the algorithmic research on the behavior coordinator.
修改评论