Distributed Target Tracking under Partial Feedback using Lyapunov-based Deep Neural Networks
Cristian F. Nino, Omkar Sudhir Patil, Sage C. Edwards, Zachary I. Bell, Warren E. Dixon · 2024
The target tracking problem is addressed for multi-agent systems where the target state information is only partially available to the agents via a heterogeneous measurement model. A necessary and sufficient condition, termed trackability, is provided, to indicate the feasibility for tracking a target with partial measurements. A Lyapunov-based deep neural network (Lb-DNN) adaptive controller is developed to achieve target tracking, under the trackability condition, by adaptively compensating for the uncertainty stemming from the unknown target dynamics. A Lyapunov-based stability analysis is provided to guarantee exponential target state estimation and tracking within a neighborhood of the target state.