Measurement Selection for Drone Swarm Cooperative Positioning Based on Fisher and Mutual Information

This study proposes a measurement selection and optimization method based on the joint evaluation of mutual information and Fisher information to address the problem of global optimal estimation degradation caused by quality differences in measurement information within cooperative positioning (CP) systems. First, an information entropy function for state estimation is constructed through Shannon's theorem, and a conditional entropy constraint model is established by integrating multi-source measurement data from the leader drones. This derivation yields a mutual information expression characterizing measurement contribution. Subsequently, a Fisher information matrix is formulated using second-order partial differential operations to enable dynamic credibility assessment of measurements. Building on this foundation, an optimized objective function is developed by fusing mutual information and Fisher information criteria, effectively mitigating inconsistent positioning accuracy induced by environmental interference and equipment failures. Simulation results demonstrate that when the leader's positioning accuracy or ranging sensor measurement precision degrades, the optimal measurement selection strategy enhances the global optimal estimation of CP algorithms. Experiments with drones equipped with MTI-630R inertial measurement unit (IMU) reveal that the proposed method compensates for the followers' positioning errors. This research establishes a novel information fusion framework for multi-agent cooperative measurement in dynamically uncertain environments, with optimization strategies extendable to distributed sensing systems such as drone swarms and intelligent transportation networks. The framework demonstrates significant potential for enhancing measurement consistency in complex operational scenarios.