基于Agent的人机组合行为建模方法研究

Agent-based Modeling of Behavior Representation for Human-aircraft Combination

为了仿真人类飞行员操纵飞机进行作战的复杂行为,提出了一种基于Agent的人机组合行为建模方法。通过建立一个以人为主导、模块化、层次化的多Agent模型体系,将飞机平台Agent模型分解为多个能反映实际装备行为的子系统Agent模型,并将飞行员Agent模型细分为信息处理、态势评估、决策规划、行动执行四层混合式结构,以综合运用模糊推理、动态贝叶斯网络、模糊Petri网等方法来建立其行为模型。最后在基于vsTASKER的虚拟仿真环境下对人机组合模型进行了测试,目前的研究结果验证了该方法的可行性。

To simulate the complex behavior of a combat aircraft operated by a human pilot, an agent- based modeling method for human-aircraft combination was introduced. By building a human-oriented, modular, hierarchical and multi-agent based model framework, the agent model for aircrafl platform was divided into many sub-agent models which could represent the real subsystems, and the pilot agent model was represented by a hybrid structure composed of information processing, situation assessment, decisions and planning and action execution. So the behavior of the pilot agent could be modeled by employing fuzzy reasoning, dynamic Bayesian network, fuzzy Petri-net in the same framework. Finally the human-aircraft combination model was integrated into a simulation environment based on vsTASKER. The ongoing work and test results indicate that this approach is feasible.