LAUSR

Abstract To overcome the issues of inflexible interactive mode, low negotiation efficiency and poor dynamic response capability of satellite clusters, we propose a new multi-dimensional and multi-Agent clusters collaboration model (MDMA-CCM), including task pre-processing, allocation, scheduling and re-planning under failed conditions. First, the task clustering based pre-processing method is adopted to improve the observation efficiency by combining potential targets. Second, the Contract Net Protocol (CNP) based secondary allocation strategy is developed to effectively increase the observation benefit and reduce the impact of task conflicts through announcement, bidding and awarding. Third, a new Interactively Re-planning Method (IRM) is proposed to effectively reduce the loss by re-inserting tasks in case of failure. Simulation results show that the new planning model not only has obvious advantages in calculation time, but also outperforms the centralized multi-Agent system in terms of benefit value and completion rate.