Joint Channel Estimation and Cooperative Localization for Unmanned Cluster Systems

This paper considers an intelligent unmanned cluster system, where edge nodes need to communicate with a cluster head to collaborate on certain tasks, posing higher requirements for accurate channel estimation and localization of edge nodes. To achieve this, we propose a joint channel estimation and cooperative localization scheme by fusing the pilot signals and location prior information (LPI). In particular, we propose a mixed location-delay domain channel model, where the dominant channel paths are modeled using the nodes' locations, while the others are modeled using a delay-domain sparse channel prior. In this way, LPI is directly exploited to achieve high-accuracy channel estimation with limited pilot overhead. In turn, the more accurate channel estimation results are exploited to achieve cooperative localization among nodes. By combining the turbo approach and the expectation-maximization (EM) method, we propose a LPI-aided EM-based turbo compressive sensing (EM-Turbo-CS) algorithm that updates the location-delay domain channel coefficients and location/delay parameters alternately to achieve high-precision channel estimation and node localization with limited pilot overhead. Finally, the feasibility and superiority of the proposed scheme are verified through simulations.