基于双目立体视觉伺服的智能车室内弯道控制

摘要

In order to improve the robustness of the cornering location of intelligent vehicle,the feature information of the corner are extracted and matched to create the path. The judgment of the vehicle steering at corner and the off-course angle error are obtained according to the frontground stereoscopic vision perception information from the vehicle. The turning control of intelli-gent vehicle is realized by using BP control strategy,which avoids the influence of nonlinear factor that is insurmountable for the traditional PID control to overcome when the parameters of the controlled system is changed. The control difficulty caused by unknown environment and nonlinear system was solved. The simulation results show that the method is effective in turning control.%为了提高智能车弯道中定位的鲁棒性，将室内弯道特征信息进行特征提取与匹配来建立路径，利用自身前景立体视觉感知信息获得车体弯道转向的判断和偏航角度的给定，并采用BP控制策略对智能小车实现转弯控制，避免了传统PID控制方法因被控系统参数变化时难以克服系统中非线性因素影响，解决了未知环境与非线性系统对控制带来的困难。仿真实验结果表明方法在转弯时控制的有效性。