中文摘要：

针对双臂机器人的手臂运动控制问题，研究了其神经网解耦和路径规划算法。首先，对其双臂机械结构进行了分析，总结了各个关节对机器人末端位姿的影响；其次，通过MDH法建立了连杆坐标系，并给出了机械臂的运动学模型，进而依据手臂末端位置与姿态解耦的特点，通过三层BP神经网络与解析法相结合的方式研究了解耦逆运动学算法，避免了蚁群算法收敛较慢的问题；最后针对S曲线介绍了规划过程中各个参数的求解和适配方法，提出一种基于数值积分的B样条求解方法实现曲线拟合，优化空间曲线的平顺性；最后，对运动学算法与轨迹规划算法的有效性进行了验证。

英文摘要：

The neural network decoupling and trajectory planning algorithm are studied for motion control problem of dual-arm robot. Firstly, the mechanical structure of the arms is analyzed and the influence of each joint on the end pasition of the robot is summarized. Secondly, the connecting rod coordinate system is established by MDH method, and the kinematic model of the manipulator is given. The decoupling inverse kinematics algorithm is studied by means of three-layer BP neural network and analytic method, which avoids the problem that the ant colony algorithm converges slowly. Finally, the process of the simulation is introduced for the S-curve. The optimization of the spatial curve is carried out by using the B-spline solution method based on numerical integration. Finally, the validity of the motion algorithm and the trajectory planning algorithm is verified by the method of solving and fitting the parameters.