Research On Trajectory Planning Algorithm And Application For Four DOF Manipulator

With the upgrading and updating of intelligent manufacturing and the ra pid development of robot technology, the higher requirements on industrial robot’s running stability, running accuracy and planning efficiency are put forward. Besides, trajectory planning algorithm is the core of the manipulator’s running stability and running accuracy. In order to complete a given task in the Industrial robot workspace, we must be accurate in the planning of the robot moving point displacement curve, the speed curve and acceleration curve. Therefore, studies on the efficient trajectory planning algorithm are of significant theoretical and practical value.First of all, The research object of this article is SCARA four degrees of freedom manipulator. The manipulator each link coordinate system was established based on D-H representation. Determine the D-H parameters of each link and joints variables. The pose matrix between the links are obtained. In the end, we got the homogeneous transformation between the fixed base and end-effector. SCARA robot’s forward kinematics equation is deduced. Besides, a 3 d simulation model of SCARA robot is established by using robotics toolbox for Matlab.Secondly, the inverse kinematics problem that is to calculate the joint variables according to the robot end-effector position and posture. In order to improve the calculating speed and precision, a method of calculating manipulator inverse kinematics solution based on ANFIS is presented in this paper. According to the cartesian trajectory selecting starting point and end point and intermediate point, obtained the corresponding joint variables in sequence. Then get the angle curve in the joint space by using interpolation method. We randomly select sample points on the joint curve. Then we use the data training and validation of adaptive neural fuzzy inference system. We carried on comparison and analysis with traditional method based on BP neural network. The results show that ANFIS ’s accuracy and calculation time are superior to the BP neural network on calculating the inverse kinematics problem.Thirdly, in view of the manipulator joint space trajectory planning problems, this paper uses the cubic spline interpolation algorithm and cubic triangular Bézier spline interpolation algorithm to solve the trajectory planning of joint space, and gives the detail construction process of the algorithm. For the cubic uniform b-spline interpolation algorithm curve not going through the first and the last feature point, the manipulator trajectory planning method based on cubic non-uniform b-spline interpolation algorithm is presented in this paper. By this method, the starting point and the termination point is overlapped with the Feature polygon point.Finally,In a SCARA robot experiment platform, the study test the three methods of trajectory planning. We got the actual trajectory of the manipulator and compared with the target trajectory. The paper showed the overall error curve and fragmented error curve, and analysed the causes of error. The experimental results verify the effectiveness of the algorithm proposed in this paper.