Research On Path Planning For A RUAV Based On Sparse A~* And Neural Dynamics Fusion Algorithm

An efficient path planning strategy is proposed for a RUAV flying in low-altitude environment,and its core algorithm is:the sparse A~*and neural dynamics fusion algorithm.The fusion algorithm is first based on the global optimal search of the improved sparse A~*,and the local fast adjustment of the improved neural dynamics model is then integrated into it,which achieves a rapid online path planning to make a RUAV accurately avoid local dynamic unexpected obstacles.This fusion algorithm not only reduces the complexity and time consumption of A~*optimal search,but also reduces the cost of neural dynamics path planning,which just shows it owns the advantages of both good quality and high efficiency.The main work of this thesis is as follows:1.The improved sparse A~*algorithm is proposed.There is an improvement of sparse A~*algorithm in this paper,mainly from the following 3 aspects:firstly,the cost function of classic A~*is improved to enhance the quality of the path result by setting weights,the actual cost g(n)which represents the principle of the Dijkstra algorithm is appropriately improved in the A~*cost function;secondly,the constraint of sparse A~*is optimized to improve the path planning decision-making method of a RUAV,according to the movement characteristic of 6 degrees of freedom;finally,the algorithm flow of sparse A~*is improved by simplifying the data structure and the process of sparse A~*,and designing a method of bidirectional parallel path planning execution.2.The improved bio-inspired neural dynamics model algorithm is proposed.Mostly combined with the characteristics of the degrees of freedom of a RUAV,there is a limit to the range of areas where the activity values of the neurons update,and its mode of totally update is modified to local update,then the number of the ergodic neurons is reduced,the global space and time complexity both decrease,and in this way the necessary condition for an efficient online path planning is established.3.The sparse A~*and neural dynamics fusion algorithm is proposed.The fusion algorithm proposed in this paper is an integration of the improved sparse A~*algorithm and the improved bio-inspired neural dynamics model algorithm,and there is a combination of the features of the global optimal search of A~*and the local rapid decision-making of neural dynamics model.The realization of this fusion algorithm mainly includes these two aspects:one is the fusion of the mathematical models,which is the fusion of the cost function of A~*algorithm and the state function of the neural activity value;the other is the fusion of the algorithm flows,which is the fusion of the two processes of the optimal search of A~*algorithm and the activity values state update of neurons.Finally,the path planning simulation experiments under the circumstances of global static and local dynamic obstacles in 3D cluttered multi-peak mountainous environment are designed.Compared with the experiment results of each classic algorithm and each improved algorithm,the result of the proposed fusion algorithm validates its optimal performance of both "short time" and "low cost",and especially,it can provide an efficient and quality online path planning strategy for a RUAV flying in the local dynamic environment,and has a very important significance for the“unmanned”performance improvement of RUAVs.