Intelligent Strategy Recommendation Technology Research For Manned-unmanned Platform Cooperative Decision-making

The manned-unmanned platform cooperative decision-making is aimed at ensuring that "people on the loop" under the minimum control and supervision of the combatant.Based on the high-level mission instructions,the decision is made through the fusion of the perceived battlefield situation information,and is sent to the action instruction.The underlying control platform ensures that all types of platforms can reach the right place at the right time and use the right resources to perform the right tasks.This paper mainly studies the cooperative decision-making problem at the tactical level,and builds the commander experience and domain knowledge into a strategic knowledge base to provide strategic intelligent recommendation for the manned-unmanned platform autonomous decision-making and accident handling.The operational environment for game confrontation,with the help of strategic intelligent recommendation technology,can alleviate the high cognitive load on the commander caused by the increase in the number of unmanned platforms,so that it can concentrate on a more macroscopic and abstract command and decision level.To achieve better complementarity of human and machine characteristics,and greatly improve the overall intelligence and operational effectiveness of people and equipment.This paper mainly carries out the following three aspects of work and contributions:1.In view of the increase in the number of unmanned platforms,commanders can not meet the needs of decision-making and control.A genetic algorithm based on expert strategy is proposed to improve the autonomous platform decision-making and unit control capabilities of the unmanned platform,so as to recommend the most unmanned platform in the confrontation scenario.Excellent strategy to complete tactical tasks.This paper uses the real-time strategy game verification algorithm,first encodes the expert strategy into a built-in script to form a script strategy set(Portfolio);secondly,combines the adaptive genetic algorithm(AGA)into the specific experimental environment and research questions.Finally,verify the effectiveness of the algorithm by setting the confrontation scenario.The experimental results show that the proposed algorithm overcomes other current algorithms with a win rate of 73.85%,which can realize the strategy recommendation of the manned-unmanned platform in the confrontation scenario,and realize the unmanned platform autonomous decision and unit control to complete the tactics.task.2.For complex dynamic decision-making environments,expert strategies can not solve unconsidered situations and respond to emergency emergencies,and propose Bayesian-based strategy learning and recommendation techniques to improve the adaptability of unmanned platform autonomous decision-making and face sudden situations.The supervision and learning of people in the loop has realized the dynamic adjustment of the strategy.This method can take into account more complex environmental factors and make full use of expert knowledge.Specifically,the Bayesian model of the strategy in decision-making is first established.The model considers combat elements such as enemy units,tactical values,enemy bases,etc.Secondly,it learns to fill the Bayesian probability distribution table in the expert data set;Verify the accuracy and validity of the model on the experimental platform.The experimental results show that the established Bayesian model beats the intelligent method built into the platform with a 78.2% win rate.That is,the model can make intelligent recommendations for the strategy in the manned-unmanned platform decision-making,and command and control the combat unit to perform tactical actions to achieve the best operational effectiveness.3.According to the specific man-unmanned platform cooperative decision-making system,the effectiveness of the proposed strategy intelligent recommendation method is verified.Firstly,the specific application and function realization process of the intelligent recommendation method proposed in this paper are introduced.Secondly,the specific operational scenarios are set,the initial conditions of the experiment are set and the experimental environment is built.Finally,in the specific decision system,the model is verified.Effectiveness.The experimental results show that the proposed strategy intelligent recommendation technology can effectively support the coordinated decision-making of the human-unmanned platform and improve the self-determination ability of the unmanned platform.