Multi-AUVs Collaborative Target Search Based On Bio-inspired Neural Network Algorithm

As a new type of underwater equipment with high concealment,flexibility,high degree of autonomy and low economic cost,autonomous underwater vehicle(AUV)has become one of the important aids to solve underwater problems.Aiming at the problem of target search in underwater unknown environment,the main research work of this paper is described as follows:First of all,through the analysis of the research status of multi auv collaborative technology and target search technology,the research background and significance of the subject are clarified.The AUV kinematics model,sonar model and the principle analysis of the neural excitation network are constructed to pave the way for further research.Secondly,aiming at the problem of signal source search path planning in underwater unknown obstacle environment,this paper proposes a AUV multi-objective search algorithm based on improved neural excitation network.In order to solve the problem of multi-objective search and search in small obstacle environment,the neural network-based target eusop algorithm is combined with genetic algorithm and improved artificial potential field method to optimize.Based on the above two improvements,the steering angle limitation and steering cost of AUV are considered,and its efficiency and feasibility are verified by comparative experiments.Then,aiming at the underwater target search problem with unknown target location,this paper proposes a real-time perceptual map based target search algorithm.Based on the establishment of real-time perceptual map including probability map,uncertainty map and revisit prime map,and the setting of its update rules,the algorithm sets up the attractor map to guide the AUV to search the unknown corner area,and establishes the attractor update formula combining with the neural excitation network algorithm,which improves the search efficiency.Moreover,a search state map is established to guide the AUV to search the surrounding areas that are not searched.Based on the real-time perceptual map,the search revenue function is set up,so that the AUV can make the optimal search decision.The target search algorithm based on real-time perceptual map can efficiently search the target in the unknown environment.The feasibility and efficiency of the proposed algorithm are verified by the simulation experiment of the search task with random target distribution.Finally,aiming at the problem of underwater cooperative search in the unknown large-scale sea area,this paper proposes a distributed rolling time domain search algorithm.On the basis of real-time perception map,the activation mechanism of attraction source map is adjusted,and the allocation strategy of attraction source is proposed to meet the search requirements of multi AUV system.Using distributed system structure,the whole system search decision-making problem is divided into local search decision-making problem of each individual,which reduces the requirement of communication link and increases the robustness of multi AUV system.Considering the collision avoidance between AUVs,the rolling time domain search strategy model is used to make the AUV search the area with high uncertainty and high efficiency as much as possible Reduce the uncertainty of the whole search environment,and search to the target faster.The simulation results show that the method is robust and efficient in the case of normal search and fault AUV.