基于神经网络的农用车辆自动跟踪控制

Automatic On-tracking Control of Farm Vehicle Based on Neural Network

跟踪控制器是农用车辆自动行驶控制的重要组成部分。将神经网络控制技术应用于车辆的自动行驶控制中,使得控制器具有良好的自学习功能,提高了控制器的环境适应能力和现场处理能力。在牧草地上的实车实验结果表明:车辆沿直线路径自动行驶时,95%的偏差绝对值小于5cm。

On-tracking controller plays an important role for autonomous farm vehicles. The neural network control method was applied into the control of autonomous farm vehicles, which enabled the controller with sound self-learning function. The environment adapting and on-the-go process ability of the controller were improved. The test results showed that 95% of the absolute errors were less than 5 cm when the vehicle was tracking on rectilinear path on grassland.