计算机生成兵力(CGF)系统中，通常认为CGF实体具有完全的战场感知能力，但这与实战过程中信息获取的不确定性相互矛盾。为了解决上述问题，在研究感知行为机理的基础上，对其产生过程进行形式化描述，并进一步分析了影响证据可信度的客观因素和主观因素，并结合实际战场环境和装备性能完成客观因素的量化工作，运用模糊逻辑量化了主观因素。最后利用证据理论合成法则将直接证据、间接证据以及先验态势进行合并，建立了可计算的CGF实体战场感知行为模型。通过仿真实例验证了模型的可用性。

It is generally regarded that the entity of Computer Generated Forces (CGF) has complete battlefield perception capability,which is conflicting with the uncertainty of acquiring information in actual combat.In order to solve the problem,the authors formalized the generation procedure of perception behavior on the basis of its mechanism analysis.Then main objective and subjective factors that have effect on evidences reliability were analyzed.The object factors were quantified based on actual battlefield environment and equipment performance,and the subjective factor was quantified by using fuzzy logic.Finally we combined the direct evidence,indirect evidence and transcendental situation together to establish a computable perception behavior model of CGF entity in battlefield.Simulation was executed to validate feasibility of the model.