Cooperative Navigation of Mobile Radar Sensors Using Time-of-Arrival Measurements and the Unscented Kalman Filter

Multiple-input multiple-output (MIMO) mobile distributed radar networks have the potential to dramatically enhance the state-of-the-art in radar sensing capabilities. One of the main challenges in implementing practical mobile radar sensor networks is providing the navigation accuracy necessary to meet the strict positioning requirements of radar sensors, particularly at high frequencies. This challenge is exaggerated in circumstances when an absolute reference such as the global positioning system (GPS) is not available. Recently, concepts have emerged in the area of cooperative navigation, where a navigating network of platforms uses inter-node measurements of range to correct for errors in inertial measurements, enabling high-accuracy navigation even when GPS is unreliable. Although these techniques have been studied extensively in theory, the presence of practical implementations in the literature is sparse. In this paper, a practical algorithm is presented for cooperative navigation. The algorithm is based on the unscented Kalman filter (UKF) due to its simplicity, moderately low computational burden, and robustness to the nonlinear navigation process and measurement model. The algorithm is explained with all the required information to implement it in a system. Simulated results of the algorithm are shown for a variety of circumstances, demonstrating that the proposed estimation technique can practically and significantly enhance the accuracy of navigation in a network of sensors with and without access to GPS. Finally, to demonstrate the utility of cooperative navigation capabilities for mobile radar networks, a simulation study is conducted. In this study, a synthetic aperture radar (SAR) simulation is performed on a swarm of simulated radar platforms to demonstrate the substantial increase in radar imaging performance enabled by the cooperative navigation scheme, which leads to smaller amounts of image translation and smearing than the non-cooperative case.