论文部分内容阅读
激光雷达已被广泛应用于移动机器人地图构建、定位以及跟踪等任务。针对激光雷达提出一种鲁棒测量模型。首先利用分段分析的方法对激光雷达的异常测量射线进行检测并滤除;然后提出逆向射线跟踪技术用于计算扫描射线的期望距离以及似然概率;最后选择高似然度的射线计算整体似然概率,从而滤除被动态障碍干扰的射线。该测量模型鲁棒性表现在两个方面:(1)滤除了异常的测量射线;(2)滤除了由于动态障碍干扰的射线。通过实验验证了算法的鲁棒性和精度。
Lidar has been widely used in mobile robot map building, positioning and tracking tasks. A robust measurement model for laser radar is proposed. Firstly, the method of piecewise analysis was used to detect and filter the abnormal measurement radiation of LIDAR. Then, the reverse ray tracing technique was proposed to calculate the expected distance and the likelihood of the scanned ray. Finally, the ray of high likelihood However, the probability, which is filtered out by the disturbance of the dynamic obstacles ray. The robustness of the measurement model is manifested in two aspects: (1) filtering out the abnormal measurement radiation; (2) filtering out the radiation disturbed by the dynamic obstacle. The experiment proves the robustness and accuracy of the algorithm.