基于神经网络模型的拱泥机器人控制

拱泥机器人是为打捞沉船的关键工序—水下攻打千斤洞而设计的特种机器人.分析了拱泥机器人在水下泥土环境中受力情况,建立了基于MATLAB的拱泥机器人头部的动力学模型.针对拱泥机器人工作环境的不确定性,将传统PID控制与神经网络相结合,建立了一种应用BP神经网络实现PID参数自整定的自适应PID控制模型.控制器由PID控制器、神经网络NNC和NNI组成.神经网络NNC能够根据拱泥机器人动态特性的变化,自动整定PID参数,从而改善了控制性能.神经网络NNI为NNC提供学习的梯度信息.计算机仿真结果表明该方法是有效的.

Control of move-in-mud robot based on neural network model

The move-in-mud robot(MMR) is a new type of special-use underwater robot. It can perform hole excavating work to benefit sling passing in the planned trajectory of the underwater mud. The forces on the MMR under the ocean's mud were analyzed and a dynamics model of the MMR's head was established. Aimed at the strong uncertainty of the MMR's working environment, a kind of adaptive PID control method using BP neural network was proposed. The BP neural network was used to modify the parameter and gives the MMR adaptive capacity. The neural networks model for the head of the MMR was established to provide grads data for the BP neural network learning. The simulation results reveal the effectiveness of the neural control strategy.