基于预测窗的轮式移动机器人最优避障避碰算法

针对非线性轮式移动机器人的避障以及多机器人间的相互避碰问题,提出了一种基于预测窗的避障避碰算法.首先为了便于预测碰撞的发生,通过反馈线性化将非线性的机器人运动学模型转化成线性模型;然后根据线性模型预测会导致机器人发生碰撞的所有相对虚拟加速度变化量集合,称之为加速度变化障碍.基于此,为每个机器人构造既能躲避障碍物又能相互避碰的可行加速度变化集合.然后通过优化指标函数求得最优虚拟加速度变化量,最后将其转换成机器人的实际控制量.这种算法与现有的相比,可使机器人在避障或避碰过程中的行驶方向角、线速度的变化幅值更小,角速度和线加速度的变化更为平顺,而且运行所用的平均时间更短.仿真结果演示了所提出算法的有效性和相对于已有方法的优势.

The optimal obstacle/collision avoidance algorithm for wheeled mobile robots based on prediction window

This paper proposed an obstacle and collision avoidance algorithm for nonlinear wheeled mobile robots,based on prediction window. First, for ease of predicting potential collisions, the nonlinear kinematics model of each robot is transformed into a linear model by using the feedback linearization technique. Then, based on the linear models, the set of all relative virtual acceleration changes that will lead to collision is predicted for each robot, which is called acceleration change obstacle. Accordingly, the feasible acceleration change set which can avoid collisions with obstacles and other robots is constructed for each robot. Finally, the optimal virtual acceleration change is obtained by minimizing an index function, and is further converted into the actual control input of each robot. Compared with existing approaches, the proposed algorithm can result in smaller variations in each robot’s heading angle and linear velocity, and smoother angular velocity and linear acceleration during the process of obstacle avoidance or reciprocal collision avoidance. Simulation results illustrate the effectiveness of the proposed algorithm and its advantages over existing methods.