Research On Maritime Distress Target Search Algorithm For Agents

The location of the distress target at sea is uncertain and time varying,which has brought great difficulties to the search task planning.In view of this problem,an unmanned agent-based maritime distress target search algorithm is proposed in this paper.The algorithm first studies the probability of containment model of distress target based on the drift characteristics of the maritime distress target,and then,a search path planning algorithm based on the probability of containment is proposed.In order to solve the time-varying problem of the target position,an optimal time window-based search algorithm is proposed to realize the autonomous planning and search for the agent,which can effectively improve the efficiency and success rate of searching the maritime distress target.The main content of this paper is as follows:1.The method of establishing the probability of containment(POC)of marine distress targets model based on Monte Carlo random particle approach and the grid method is proposed.The method first analyzes the drift characteristics of the distress target,and uses the computer to randomly simulate the drift path of the random particles according to the environmental factor data of the incident area.Then,according to the position of the particle,the area to be searched is determined,and divided into several sub-areas by the grid method.Then the POC value of each sub-area is calculated by the particle number statistics,and finally the POC model of the target is established,which provides priori information for the agent to plan a search path.2.The search path planning algorithm based on the POC model of the maritime distress target is proposed.In order to search for a sub-region with higher POC of the target as soon as possible within a limited time,the search path planning algorithm takes the POC value of each sub-region in the area to be searched as the path weight value.When planning each step for a path,only the search path with the highest weight is retained to achieve an efficient search for distress targets.3.The time window-based search algorithm is proposed.Since the location of the distress target is time varying,the time of search task planned by the POC model is limited.Therefore,the proposed algorithm updates the posterior probability of the target in each sub-region in the area to be searched based on Bayesian theory,when the search time overstep the time window.The POC model of the target is also updated and the next search task is re-planed.At the same time,the factors which influence the time window are studied,and the conditions of the optimal time window are obtained through experiments.By building a simulation platform of maritime distress target search,the task of searching for distress targets is simulated.The results show that the proposed algorithm can accurately construct the POC model of target,and rationally plan the search path as well as continuously plan the search work for the target.The search success rate can reach more than 95%under the optimal time window.