Research On Kinematics Of MOTOMAN-UPJ Manipulator

The Inverse Kinematics Problem (IKP) is one of the hot topics in manipulator field.According to the structure characteristics of the MOTOMAN-UPJ manipulator, this paper puts forward a kinematics model and deduces a forward kinematics equation by using D-H transformation matrix method. After that, an inverse kinematics equation that needs the inverse matrix multiplication only once is presented in this paper. Compared with the conventional methods, the method reduces the work of calculation greatly. This paper considers the problem that sometimes the results of the inverse kinematics equations are not real. The effect to each position parameter to the robot manipulator inverse kinematics is discussed. An inverse kinematics equation is confirmed and the situation of missing results is avoided.The RBF neural network used in inverse kinematics is trained by using some forward kinematics results as training data set. Because of its local approaching ability, the inverse kinematics problem of the manipulator can be transformed into the weight-training problem. The mapping from joint-variable space to operation-variable space is realized. The computation is reduced greatly to the conventional methods, and the problem of real time is resolved by accelerating the convergence to the BP neural network. This paper studys numerical example, and gives a contrast for training result with expecting result. The results show that the strategy is efficient, which provides theoretical base for its optimal design and application, and make kinematics simulation and control become more accurate and reliable.At last, after knowing its programming environment and operating process familiarly, the pertinent kinematical experimentations of manipulator are designed combing its structure and movement characteristics. The experimentations are done to validate the feasibility of the two kinds inverse kinematics equations of the manipulator.