Robust predictive control for mobile robots in uncertain environments

首先，提出了一种新的基于模型预测控制（MPC）的轨迹生成方法，以防止预测时间步之间的时间间隔发生碰撞。并将该方法推广到基于MPC的多车辆编队控制中。我们还提出了一种减少基于MPC的避撞混合整数问题产生的子问题的算法，这使我们能够减少MPC在每个采样时间的计算时间。在此基础上，针对现有多数编队控制方法在车辆数量较大时难以确定每辆车辆目标位置的问题，提出了一种基于编队模式抽象的编队控制方法。更精确地说，我们引入了表示地层位置、方位和形状的抽象变量，并提出了一种自动控制地层形状的方法，使其在保持算子给定的地层宽度的情况下沿目标路径移动。提出了一种蛇形机器人的路径跟踪控制方法，为本文提出的编队控制方法提供了基础。通过仿真和实验验证了这些方法的有效性。

First, a new trajectory generation method based on model predictive control (MPC) has been proposed to prevent the collisions in time intervals between prediction time steps. This method has been extended to the multi-vehicle formation control based on MPC. We have also proposed an algorithm to decrease sub-problems generated to solve mixed integer problems for MPC based collision avoidance, which allows us to decrease the computation time for MPC at each sampling time. Furthermore, we have focused on a problem of most existing formation control method i.e. difficulty of determining the target position of each vehicle when the number of the vehicles is large, and have proposed a new formation method based on the abstraction of formation pattern. More precisely, we have introduced the abstraction variables which represent the position, orientation and shape of the formation, and proposed a method to control the shape of the formation automatically so as to move along the target path while keeping the width of the formation given by operators. We have also proposed a path tracking control method for snake-like robot which gives the basis of the proposed formation control method. The effectiveness of these methods has been demonstrated by simulations and experiments.