CNN-Based Intelligent 3D Path Planning Algorithm in the Framework of the Improved Lazy Theta*

Abstract

AbstractIn this paper, a CNN-based intelligent path planning algorithm in the framework of the improved Lazy theta* is proposed to solve the problem of path planning in the 3D terrain environment. The key point of the proposed algorithm is that the safety factor and the total length of the path will be comprehensively considered. By considering these two factors, a short, safe and smooth path can be planned efficiently and automatically in a 3D terrain environment. In order to solve the problem of the path moving close to the edge of the obstacles and passing dangerously between multiple obstacles, CNN is used to create a continuous and safe 3D topographic map and improve the ways of node expansion to ensure the safety of the path. Moreover, a weight self-adjustment strategy is introduced to optimize the path cost function, which solves the problem of the low search efficiency. The simulation results show that compared with the ordinary A* algorithm and Lazy theta* algorithm, the path planned by the improved intelligent Lazy theta* algorithm proposed in this paper is safer and smoother, and the search efficiency is higher, which can be applied to different planning objects according to different task scenarios.Keywords3D path planningLazy Theta*Convolutional neural network