基于视觉神经元PID的AGV轨迹跟踪控制器

AGV Trajectory Tracking Controller Based on Visual Neuron PID

为了提高视觉自动导引车(automated guided vehicle,AGV)的轨迹跟踪精度和控制器系统鲁棒性,提出了一种基于视觉神经元系统处理偏差数据的轨迹跟踪控制器。首先建立含有神经元的图像数字化和传感器辅助数据的处理系统,然后建立AGV视觉神经元运动学模型,将航向偏差和辅助数据误差作为视觉神经元PID纠偏路径控制算法跟踪控制器的输入变量,利用无刷直流电机数学模型和AGV纠偏运动学模型建立最优控制算法,最后通过纠偏电压导入双通道PID实时调节驱动轮的差速以实现AGV能按更快速和最优化的路径纠偏。实验结果表明,利用视觉神经元PID纠偏路径控制算法,体现为寻轨时间更短,转向突变时反应更机敏;在横向偏差1 m的直线行驶中,回归直线用时提高30%,其左右横向误差均不超过1.7 mm,误差减小了68%,AGV做出迅速反应和提高导航精度时,也能满足跟踪控制器的鲁棒性。

In order to improve the trajectory tracking accuracy and controller system robustness of automated guided vehicle(AGV),a trajectory tracking controller was proposed based on the processing deviation data of visual neuron system.Firstly,an image digitalization system containing neurons and a sensor-assisted data processing system were established,and then a kinematics model of AGV visual neurons was established.The heading deviation and auxiliary data error were taken as the input variables of the tracking controller of the PID correction path control algorithm of visual neurons.The mathematical model of brushless DC motor and the kinematic model of AGV deviation correction were used to establish the optimal control function.Finally,the differential of the driving wheel was adjusted in real time by introducing the two-channel PID through the deviation correction voltage so that the AGV could correct the deviation in a faster and optimized way.The experimental results show that the visual neuron PID correction path control algorithm can shorten the time of track finding and respond more sensitively to turn mutation.In the straight line driving with 1 m lateral deviation,the time of regression straight line is increased by 30%,the left and right lateral errors are less than 1.7 mm,and the error is reduced by 68%.When AGV makes rapid response and improves navigation accuracy,it can also meet the robustness of tracking controller.