Abstract: Due to the shortage of manual tension clamp drainage board bolt tighten for high voltage transmission line， such as low operation efficiency， labor-intensive， high-altitude and high risk， an bolt tighten live operation robot was designed which had double arms and double manipulators. The key issues of bolt tighten was analyzed theoretically during the entire operation process， and a RBF neural network model of bolt tighten process control was established. The robot bolt tighten process was abstracted into nonlinear approximation， and a robot bolt tighten status monitoring control method based on the RBF neural network was proposed. Finally the live operation result showed that the connection reliability of the robot tighten bolts is greatly enhanced by using the proposed method that has strong engineering practicability， and the operation efficiency ，operation safety and operation operability are all improved.

Key words: drainage board bolt, RBF neural networks, live operation robot, manipulator, tighten control