Liu Zaixin

Research on motion control algorithm for mirosot soccer robot

Based on nonholonomic restriction condition, the kinematics model of wheeled soccer robot is derived. Based on dynamic restriction condition, the dynamic model of wheeled soccer robot is derived with considering the dynamic of driving electric motor. Analysis shows that the line acceleration and angle acceleration of soccer robot are related to current line velocity, angel velocity, given line velocity and given angel velocity, etc. A new motion control algorithm — proportion differential coefficient and second step differential coefficient (PDD2) is put forward with considering the second step differential coefficient of distance error and angle error. The simulation experiments of motion control are done with proportion method, PD method and PDD2 method. The experiments show that PDD2 algorithm enables the soccer robot to move to a moving target point rapidly, accurately and steadily.

Tooth contact parameters analysis of the non-backlash double-roller enveloping hourglass worm gearing

The tooth contact parameters such as lubrication angle, induced normal curvature, entrainment velocity, et al. of the nonbacklash double-roller enveloping hourglass worm gearing and influencing factors are studied systematically. The averages of the lubrication angle, the induced normal curvature, the entrainment velocity and the rollers rotation angle which located on a number of mesh lines are taken as meshing performance index. Based on numerical computation, the influences of the roller radius, the roller installation offset distance, the throat diameter coefficient and the drive center distance on the tooth contact parameters of the non-backlash double-roller enveloping hourglass worm gearing are described by using the contours. The results show that a superior performance worm gear can be obtained by choosing rationally geometry parameters when the center distance is between 250~450mm.

Robot autonomous navigation based on multi-sensor global calibrated

Aiming at the robot problems such as little relevance, low accuracy of the simultaneous localization and map building (SLAM) and easy locking, lack of initiative of the navigation system, the multi-sensor vision system is introduced, and then unifying the data of each sensor by world coordinate system of global calibration based on the local calibration of each vision sensor module, a serial of local maps are combined into a global map by the derive of Least-Square (LS). Preprocessing of the global map data is done by the genetic programming (GP) arithmetic and inference is done with the delta fuzzy rule to plan the best routine to achieve robotic autonomous navigation. Simulation results show that the robot can create accurate and complete map of the environment and bypass the obstacles agilely to reach the destination smoothly and reliably with the map. Thus the feasibility and effectiveness of this strategy is verified.