Path planning technology which is the core of robot control is being widely studied in the field of automatic control system. The traditional path planning technology focuses on the navigation of point to point, which is a single mission mode and unable to meet the needs of more complex tasks. A new path planning method based on linear temporal logic, which is developed in recent years, can deal with relatively complex tasks with timing property. But in the existing research results of this field, the existence of an effective multi-point circuit problem solution has not been found. Therefore, on the premise of meeting the requirements of mission desired described with linear temporal logic language, planning an optimal multi-point patrolling path is of great significance.In this dissertation, the problem of multi-point patrolling path planning under the limit of complex tasks, which is described with linear temporal logic, is being solved. First, deal with the case of a single robot planning problem. The environment is modeled as a transition system and the patrolling task is described with linear temporal logic formula, an extended Product automaton combining the transition system and the linear temporal logic formula is constructed by implementing circular shift algorithm such that the network topology with complete path information could be established, the Dijkstra algorithm is utilized to search the optimal path in the network of the extended Product automaton, and thus, the optimal path satisfying the task requirement is correspondingly obtained. Then, a promoting from a single to multi-robot using the global transition system approach is being proposed, and integrate the information of a number of robots with different motion abilities in the environment, propose the optimal patrolling path planning algorithm for multi-robot, to obtain the team path for each individual, and such that minimize the cost of all individuals’runs. Finally, design a robot navigation system with two-layer structure, which uses the path planning method based on linear temporal logic as the layer of global planning, a path track controller based on fuzzy logic control as the on-line control layer.Simulations are both given to single robot and multi-robot cases, results show that a single robot path planning algorithm can achieve the optimal multi-point plan included in mission requirement, multi-robot optimal path planning algorithm can achieve optimal multi-robot motion planning of different motion capacities. In addition, as to the robot navigation system, simulation is utilized to firstly verify the reliability of the fuzzy controller designed, followed by realizing the navigation system on the experimental platform, the experimental results show the effectiveness of the system. |