Abstract:In advanced field,the requirements for machining accuracy of precision small parts are getting higher and higher.As main machining equipment,five axis parallel machine tool plays an extremely important role in ensuring the machining accuracy of finished parts by strictly controlling the motion posture of the tool on the machine tool.A five axis parallel machine tool end motion pose adaptive control technology was studied.Two CCD cameras were arranged in front of the five axis parallel machine tool as the vision system to capture the left and right images of the tool end.On the basis of image preprocessing,feature points extraction and matching,the pose parameters,including 3D coordinates and three pose data,were solved.Taking the actual pose of the tool end motion as the input,the five axis joint angle compensation closely connected with the tool end was obtained by using neural network.Using the PID controller,the position and attitude control quantity was calculated through the compensation quantity,and the position and attitude errors were continuously corrected,closing to the ideal position and attitude.The results show that applying the proposed control method,the average error of the ideal pose is smaller,and the fluctuation of the average error is the smallest,and the fluctuations of the 3D position coordinates are only from -0.02 mm to 0.015 mm,-0.01 mm to 0.02 mm,-0.02 mm to 0.02 mm;the rolling angle,pitch angle and azimuth angle fluctuate only between -0.02° and 0.03°,-0.01° and 0.04°,-0.02° and 0.01°,which show that the proposed control method has higher accuracy and control stability.