【摘要】 为解决复杂工况下的机械臂避障与路径规划问题，提出一种通过降维预实验预设分段式插值的改进RRT算法。该算法基于多次降维实验拟合，得到路径中间点，通过插值不断替换目标点，将长距离多障碍的目标路径分段求解，具有较好的时间与求解优势，并且能得到研究者所期望的路径走向。经过可视化结果对比：所提算法相比RRTConnect算法减少了约71%的收敛时间，减少了约58%的迭代计算量；而相对于传统RRT*算法减少了约45%的收敛时间，减少了83%的迭代运算量。最后通过Matlab仿真验证，实现轨迹优化与关节碰撞验证，并在真实工况下通过编程控制机器人复现算法得到的路径，验证了改进算法所求解的轨迹的收敛时间短且有效，具有良好的应用价值。更多还原

【Abstract】 In order to solve the problem of obstacle avoidance and path planning for robotic arms under complex working conditions, an improved RRT algorithm is proposed by pre testing segmented interpolation by means of the dimensionality reduction. This algorithm is based on multiple dimensionality reduction experiments to fit the intermediate points of the path, and can continuously replace the target points by the interpolation to solve the long-distance and multi obstacle target path in segments. It has high advantages of time and solution, and can obtain the desired path direction of researchers. The visualization results show that in comparison with the RRTConnect algorithm, the proposed algorithm can reduce the convergence time by about 71% and the iteration computation by about 58%. In comparison with the traditional RRT* algorithm, the proposed algorithm can reduce the convergence time by about 45% and the iteration computation by about 83%. After Matlab simulation verification, the trajectory optimization and joint collision verification are realized, and the path obtained by the algorithm is reproduced by programming the robot under real working conditions, which verifites that the convergence time of the trajectory solved by the improved algorithm is shorter and effective, and has good application value.更多还原