基于RBF神经网络的无人驾驶电机车避障控制  

Obstacle Avoidance Control of Unmanned Electric Locomotive Based on RBF Neural Network

作　　者：王玉有 余红云 吴海江 WANG Yuyou;YU Hongyun;WU Haijiang(Yunnan Diqing Nonferrous Metals Co.,Ltd.,Diqing 674400,China)

机构地区：[1]云南迪庆有色金属有限责任公司,云南迪庆674400

出　　处：《机械与电子》2025年第1期47-52,共6页Machinery & Electronics

摘　　要：无人驾驶电机车工作现实环境复杂多变,包括各种静态和动态障碍物、光线变化、天气条件等,这些因素都会影响环境感知效果,从而导致避障效果下降,为此提出一种基于RBF神经网络的无人驾驶电机车避障控制方法。通过自适应方向搜索算法成功筛选出道路边界的候选点并对其展开曲线拟合处理。在曲线拟合处理基础上对道路边界内允许通行的区域展开点云聚类分割,从而精确获取道路内障碍物的位置及距离信息,结合RBF自适应补偿控制器与RBF鲁棒优化控制器实现无人驾驶电机车避障控制。实验结果表明,所提方法可以获得精准的无人驾驶电机车避障控制结果,实际应用效果好。The working environment of unmanned electric locomotives is often complex and variable,including various static and dynamic obstacles,light changes,weather conditions,etc.These factors can affect the perception effect of the environment,leading to a decrease in obstacle avoidance effect.Therefore,an RBF neural network-based obstacle avoidance control method for unmanned electric locomotives is proposed.Candidate points for road boundaries are successfully selected through adaptive directional search algorithm and curve fitting processing is implemented on them.On the basis of curve fitting processing,point cloud clustering segmentation is carried out on the allowed areas within the road boundary to accurately obtain the position and distance information of obstacles on the road.Combined with RBF adaptive compensation controller and RBF robust optimization controller,obstacle avoidance control for unmanned electric locomotives is achieved.The experimental results show that the proposed method can obtain accurate obstacle avoidance control results for unmanned electric locomotives,and the practical application effect is good.

关 键 词：RBF神经网络 无人驾驶 电机车 避障控制 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] U463.6[自动化与计算机技术—控制科学与工程]