UAV Tracking Moving Target Mission Planning Based on TW-SAC Algorithm

Aiming at the problem of UAV tracking moving target mission planning in three-dimensional complex environments, this paper proposes a three-way decisions and soft actor-critic (TW-SAC) fusion intelligent algorithm. Considering the needs of reward maximisation and threat minimisation for UAV tracking moving targets comprehensively, the reward function and threat assessment model for UAV tracking moving targets are constructed based on the SAC algorithm. At the same time, the three-way decisions is introduced into the SAC algorithm, and the uncertainty of the three-way decisions is introduced to set different reward weights for the reward function, so as to construct a UAV intelligent tracking strategy based on the TW-SAC algorithm. Simulation validation shows that the fusion intelligence algorithm converges faster, locks onto the moving target more frequently, and has better robustness than the single SAC algorithm when it is applied to the task planning process of tracking a moving target by a UAV in 3D space.