基于视觉非完整自主车的道路避障

就四轮非完整自主车提出了一种基于圆轨迹的道路避障策略·道路上的障碍按照障碍距离自主车的远近可以划分层次 ,一个层次的障碍能在自主车视场中全部出现·然后给出基于圆轨迹的避障算法 ,即自主车沿由自主车出发位姿和子目标点确定的圆弧轨迹走行·提出了移动光心得到的深度信息原理并推导了四轮非完整车的运动模型 ,为提出避障策略的基础作准备·基于圆轨迹避障策略能够很好地满足非完整自主车的运动特性 ,体现了方法的优越性·

Vision-Based Obstacles Avoidance for Nonholonomic Mobile Robot

Obstacles avoidance for four-wheel nonholonomic autonomous land vehicle (ALV) was discussed, and an obstacle avoidance method based on circle locus was put forward. The principle of obtaining the information of depth by parallel moving lens was given. The dynamic models of four-wheel nonholonomic ALV were lucubrated. The constraints of dynamics and obstacle avoidance were induced. Obstacles in the road can be layered according to the distances from obstacles to ALV. Thus, one layer of obstacles appear in the vision view of single camera. One sub-goal can be determined according to the configuration of ALV and the locations of obstacles of the layer. The path to the goal of ALV is the iteration of the paths to the sub-goal of each layer. ALV follows the arc determined by the initial configuration of ALV and the sub-goal. Knowing the start and the sub-goal reduces the alteration of the running of ALV.