Research On Multi-AUV Collaborative Task Allocation Method Based On Swarm Intelligence

How to achieve multi AUV allocation in the uncertain dynamic marine environment is a complicated issue about NP(Non-Deterministic Polynomial)complete optimization,and it is difficult to find a global optimal method to solve it.Therefore,it is important to quickly generate the best solution to solve the problem in order to minimize the time and energy of the tasks used by AUV while maximizing its profit.With regard to the coordination and cooperation of multi robot system,complex tasks that are difficult to describe,load balance is difficult to assign,and collaborative system modeling and control structure are difficult to be modeled,all of which may lead to the task complexity escalation,resulting in the low efficiency of multi-robot system coordination and cooperation,and make it unable to complete the task smoothly.It can be seen that multi AUV is implementing the task allocation scientifically and reasonably by making full use of all resources to guide the AUV execution task planning,reducing the system resource consumption cost to a minimum.It is a heated issue to be solved in multi AUV collaboration system.This dissertation,in view of the biomimetic task allocation method based on group intelligence technology,aims at solving the issues of multi AUV collaborative task allocation architecture,multi AUV task allocation,especially multi AUV task planning and multi AUV task replanning,etc.First of all,considering the current problems and challenges of uncertainty in the marine environment,multi AUV task allocation type and task allocation method,a multi AUV cooperative task allocation model is proposed,along with distributed multi AUV task allocation model and multi AUV dynamic task allocation model,this paper analyzes the design and construction of these three models.At the same time,in order to improve the Comprehensive ability of multi AUV task allocation,task planning and task re-planning in the mutational environment,a multi AUV collaborative task allocation architecture(MACTA)is used to help AUV execute tasks in a severe unknown environment at the lowest cost.Secondly,with regard to the uneven load and the trap of local optimum in multi AUV system,a multi AUV task allocation method is introduced,which is based on chaos optimization and quantum particle swarm optimization(CQPSO)with task allocation module in MACTA architecture.The CQPSO algorithm is applied to mapping out the elements of optimizing objective function,constraint condition and decision variable,in correspondence with the elements of swarm intelligence optimization algorithm,through the optimization of the algorithm,to solve the task allocation problem in the multi AUV task allocation.In the optimized local optimal neighborhood point iteration,logistic function is used to generate chaotic sequence,which may be chaotic in a sequence of optimal position and replace the current position,away from the local optimum and get the global optimal task allocation process AUV.Thirdly,considering the lack of autonomy in multi AUV route avoidance mission planning,a multi-AUV mission planning method based on response threshold ant colony algorithm(ACO)is researched in combination with mission planning module in MACTA architecture.Because the marine environment in which AUV is located is complex and dynamic,the task planning process will constantly change,and the AUV is required to reduce the computational cost of the task planning algorithm.Therefore,ant colony algorithm is needed with the response threshold of mission planning algorithm,by constructing ACO response planning model based on the threshold multi route AUV multiple obstacle avoidance task planning.Finally,because the work plan cannot be independently adjustable to the multi AUV task planning,implementation planning new tasks should be combined with the task in the MACTA architecture planning module,with a recedinghorizon quantum artificial bee colony algorithm based on differential evolution(DEQABC)method of replanning of multi AUV task.In order to enable multi AUV to accomplish mission planning quickly and execute tasks effectively,the multi AUV task replanning is believed to rely on the global optimization of multilevel AUV distributed task replanning.By introducing balance factors into the assignment of AUV mission planning of traditional contract net,the unbalanced load defects of traditional contract net multi AUV distributed task planning have to be improved,facilitated by the rolling horizon differential evolution of quantum bee colony algorithm in the uncertain ocean environment of distributed multi AUV dynamic task planning process,thereby realizing the implementation of AUV mission planning.In conclusion,with different group intelligent algorithms in the multi-AUV task allocation,task planning and task re-planning model,the multi-AUV collaborative task allocation(MACTA)architecture,is structured which may prove of its value as a new method and concept for multi AUV task allocation algorithm based on swarm intelligence research.