Saharsh & Jagtap's CBF-Only Framework Simplifies Multi-Agent Formation Tracking

Researchers S. Saharsh and Pushpak Jagtap have published a new approach to formation tracking in multi-agent systems that could simplify how swarms of drones, ground robots, or autonomous vehicles coordinate. Their paper, submitted to arXiv on June 24, 2026, proposes a real-time control framework that uses a single Control Barrier Function (CBF)-like constraint within a quadratic programming setup. Unlike prior methods that require both a nominal formation controller and a separate safety filter, this work tightly integrates formation tracking and safety into one optimization problem. The controller relies solely on relative information from neighboring agents and requires no manual parameter tuning, making it more practical for real-world deployment.

The framework is designed for heterogeneous agents—each with potentially different non-linear dynamics—and uses a leader-follower structure to drive the swarm into a desired spatial formation while avoiding collisions. The authors validated their approach through simulations of moving formations, demonstrating that the CBF-only method can achieve stable tracking without additional control layers. The paper (9 pages, 2 figures) is categorized under Systems and Control (eess.SY) and builds on recent advances in control barrier functions for safety-critical systems. By eliminating the need for a nominal controller, this work reduces computational overhead and simplifies the design process, potentially accelerating adoption in applications like drone light shows, warehouse robots, and autonomous convoy systems.