Abstract:

Computer generated forces (CGFs) are virtual combat force objects created by computers and critical elements in the field of military simulation. Deceptive path planning is a basic method of deceptive behavior, which is important for improving the intelligence and competitiveness of CGFs. However, the current combination of deceptive behavior and military simulation is insufficient, and classical path planning methods cannot effectively take advantage of the partial observability of the battlefield and achieve better deceptive effects. To solve these problems, we propose four new deceptive path planning methods by re-defining a single circular fog road network based on road networks and four classical deceptive strategies. In order to make a visual comparison between the new and classical methods, we conduct a red-blue adversarial scenario experiment on the KD-HLA-RTI-based tank combat platform. The red tank plays the role of the observed and conducts path planning according to different strategies, while the blue side, as the observer, identifies the real goal of the red tank in real time and deploys defense resources to it. The experimental results indicate that compared with the classical methods, the new ones can improve the time efficiency and deceptive effectiveness more effectively for the observed under partially observable conditions of the battlefield environment. Subsequent multiple simulation experiments demonstrate that the third approach exhibits better robustness to battlefield environmental changes in terms of improving the deceptive effects.

Key words: compater generated forces(CGF), deceptive path planning, goal recognition, fog of war, tank combat