Multi-Robot Target Monitoring and Encirclement via Triggered Distributed Feedback Optimization

TL;DR

The paper tackles cooperative multi-robot target monitoring and encirclement by formulating the task as an aggregative optimization problem and solving it via a Triggered Aggregative Tracking Feedback framework. It combines a distributed gradient-based controller with an asynchronous, event-triggered communication scheme to handle limited information and scalability, ensuring convergence to stationary points while excluding Zeno behavior. A full-stack ROS 2 architecture is developed and validated through realistic Webots simulations and real-world experiments with aerial and ground robots, including density-based danger fields and collision-avoidance via Control Barrier Functions. The results demonstrate robust, scalable coordination for complex multi-objective tasks in dynamic environments, with Monte Carlo analyses confirming reduced communication loads and favorable convergence properties. The approach offers practical implications for autonomous surveillance, wildfire monitoring, and other cooperative robotics applications where distributed, multi-objective coordination is essential.

Abstract

We design a distributed feedback optimization strategy, embedded into a modular ROS 2 control architecture, which allows a team of heterogeneous robots to cooperatively monitor and encircle a target while patrolling points of interest. Relying on the aggregative feedback optimization framework, we handle multi-robot dynamics while minimizing a global performance index depending on both microscopic (e.g., the location of single robots) and macroscopic variables (e.g., the spatial distribution of the team). The proposed distributed policy allows the robots to cooperatively address the global problem by employing only local measurements and neighboring data exchanges. These exchanges are performed through an asynchronous communication protocol ruled by locally-verifiable triggering conditions. We formally prove that our strategy steers the robots to a set of configurations representing stationary points of the considered optimization problem. The effectiveness and scalability of the overall strategy are tested via Monte Carlo campaigns of realistic Webots ROS 2 virtual experiments. Finally, the applicability of our solution is shown with real experiments on ground and aerial robots.

Figures (13)

• Figure 1: A heterogeneous team of aerial and ground robots trying to cooperatively encircle a target while patrolling points of interest.
• Figure 2: An illustrative example of desired placement. In the three subfigures, the final configuration is shown when considering only objective O1 (a), objectives O1 and O2 (b), and all the three, O1, O2, and O3 (c).
• Figure 3: Block diagram representation of Algorithm \ref{['alg:closed_loop']}, where the exchanged data are denoted by $\hat{\Lambda}_{i}^{k} \coloneqq (\hat{w}_{i}^{k} + \hat{\phi}_i^{k}, \hat{z}_{i}^{k} + \nabla_2 \hat{\ell}_i^{k})$.
• Figure 4: Block diagram scheme illustrating the overall distributed architecture specialized for both the quadrotor and the ground mobile robot systems.
• Figure 5: Top-down view of the Scenario 1 virtual experiment. The target is depicted by a fire. The starting positions of the nano-quadrotors are depicted as small blue spheres, the final ones by bigger, numbered, spheres. Here, the red line represent the trajectory of the barycenter of the team, while the blue dashed lines represent the trajectories of the nano-quadrotors.

Theorems & Definitions (1)

• Theorem 3.1