부산대학교 도서관

This paper mainly studies the optimization of the patrol route of a fire inspection robot, based on analytic hierarchy process (AHP) and fire source detection. First, the A* algorithm is improved based on two aspects of the heuristic function and of the obstacle boundary setting, while the suboptimal path is obtained in MATLAB. Next, the path planned according to the improved A* algorithm is smoothed and optimized, by means of gradient descent method, Bezier curve and B-spline curve, while the index parameters are optimized by means of MATLAB simulation. In view of the simulation results, the trajectory optimization performance index evaluation system, established by five decision criteria, including running time, path length, ride comfort, no-collision effect and quadratic optimization space, is put forward. The three kinds of optimization methods are analyzed qualitatively and quantitatively, and the results show that, in the total hierarchical ranking, the B-spline curve trajectory optimization scheme has the largest weight and is more important than the other two schemes. Finally, the superiority of B-spline curve is verified experimentally.