机器人神经网络自适应控制下的新能源电力场站巡检技术

New energy power station patrol technology based on robot neural network adaptive control

为优化新能源电力场站巡检效果,研究巡检机器人神经网络自适应控制下的新能源电力场站巡检技术。将巡检机器人移动坐标系变换到世界坐标系中,利用Lagrange方程建立巡检机器人动力学模型;以CCD摄像机为核心构建环境图像采集系统,采集并归一化和均值化处理图像信息;建立神经网络模型,结合梯度下降法则明确控制目标函数,训练网络误差,输出最终控制量,将此量作为依据实现巡检机器人自适应巡检。应用分析证明,所提技术控制精度高,在任何巡检模式和环境下都能规划出合理的巡检路径,具有较好的巡检效果。

In order to optimize the patrol effect of new energy power stations,the patrol technology of new energy power stations under the neural network adaptive control of the patrol robot was studied.The moving coordinate system of the inspection robot was transformed into the world coordinate system,and the dynamic model of the inspection robot was established by using Lagrange equation;The environment image acquisition system was built with CCD camera as the core to collect,normalized and average image information;The neural network model was established,and the gradient descent rule was used to define the control objective function,train the network error and output the final control quantity.This quantity was used as the basis to realize the adaptive inspection of the inspection robot.The application analysis showed that the proposed technology had high control accuracy,could plan a reasonable patrol path under any patrol mode and environment,and had a good patrol effect.