Research On Route Planning Of Unmanned Vehicle Based On Multi-sensor Information Fusion

With the emergence of new technologies and new ecosystems such as mobile Internet,embedded systems and artificial intelligence,all walks of life are accelerating the development of integrated innovation based on intelligent platforms,and entering the era of rapid development of "artificial intelligence".Therefore,it is of practical application value to break the existing technical barriers of civil unmanned ships to reduce the danger degree of crew members during manual operation of traditional ships.Based on the dynamic development of the frontier of unmanned ship,this paper proposes an optimization scheme for the limitations of the existing path planning algorithm,adopts the design method of embedded control system,and builds a physical test platform by itself to study the path planning of unmanned ship and verify its practicability.The main research contents are as follows:First of all,combined with the application scenario of this paper and the application conditions of common environmental modeling technology,the grid method is selected as the graph construction method of this paper.Using the open source map software API,using the binary method to complete the rasterization of the map.The mathematical model of GPS/BDS combined positioning technology is established,and the coordinate transformation of the data collected by the combined positioning module is carried out based on the Internet map service regulations,so as to realize the initialization process of the unmanned ship in the map center and provide technical support for the subsequent path planning algorithm.Then,based on the existing unmanned ship path planning algorithm,the ant colony algorithm is selected as the underlying code.Aiming at the problems of slow iteration and local optimal solution,an improved ant colony algorithm based on artificial potential field was proposed,and its improved strategy was analyzed.Three typical inland river channel models of S-channel,bifurcated double-channel and Y-channel were created by numerical simulation software.Two different grid sizes were used to divide them,and the conventional ant colony algorithm and the improved potential field ant colony algorithm were compared and verified by comparative analysis method.The results show that the improved ant colony algorithm can effectively solve the problem that the conventional ant colony algorithm is easy to fall into the local optimal solution,and it is less affected by the change of grid size.Finally,based on STM32 microcontroller and equipped with ultrasonic sensor,GPS/BDS combination positioning module,DC motor,2.4G communication module and other hard sensors,the unmanned ship test platform was built.Using the window program development platform to complete the writing process of the upper computer.Develop dynamic obstacle avoidance rules for unmanned ships,and complete the launching test of real ships.The return data recorded by the upper computer showed that the maximum difference between obstacle test data of different distances was no more than 3 cm,and the longitude and latitude difference between the actual turning point coordinates of the unmanned ship and the turning planning coordinates of the designed route was in the level of centimeters.The unmanned ship could better follow the designed route in the case of no dynamic obstacles.It is proved that the improved ant colony algorithm is feasible and practical,and can complete the launching test of the unmanned ship physical model platform.