基于能耗优化的深海电动机械臂轨迹规划

由于深海电动机械臂动力学模型较为复杂，难以基于动力学模型构建精确的能耗优化目标函数，因此，本文提出一种利用径向基函数（RBF）神经网络构建机械臂功耗模型的方法．首先，利用机械臂水下运动实验数据集训练所构建的RBF神经网络．利用基于该神经网络的功耗模型，结合机械臂关节空间轨迹规划多项式，建立机械臂能耗目标函数．然后，采用自适应粒子群优化（PSO）算法求解最优轨迹参数．结果显示，RBF功耗网络均方根误差（RMSE）为20.89 W；经过优化的轨迹的能耗比实验轨迹的能耗均值降低410.8 J（18.3%）．实验结果表明基于自适应PSO算法的轨迹规划方法实现了能耗优化的目标．

Trajectory Planning of Deep-sea Electric Manipulator Based on Energy Optimization

The dynamic model of the deep-sea electric manipulator is complex, so it is difficult to construct an accurate objective function of energy optimization based on dynamic model. Therefore, a method to establish power model of the manipulator is proposed, using radial basis function (RBF) neural network. Firstly, the RBF neural network is trained by using the experimental data set of underwater motion of the manipulator. By utilizing the power model based on the RBF neural network, the energy objective function of the manipulator is established combined with the trajectory planning polynomial of the manipulator joint space. Then, the adaptive particle swarm optimization (PSO) algorithm is used to solve the optimal trajectory parameters. The results show that the root mean square error (RMSE) of RBF power network is 20.89 W, and the energy consumption based on the optimized trajectory is 410.8 J (18.3%) lower than the average energy consumption based on the the experimental trajectory. The experimental results show that the trajectory planning method based on the adaptive PSO algorithm achieves the goal of energy optimization.