Is this project an undergraduate, graduate, or faculty project?
Undergraduate
Project Type
individual
Campus
Daytona Beach
Authors' Class Standing
Nick Wilson, Sophmore
Lead Presenter's Name
Nick Wilson
Lead Presenter's College
DB College of Engineering
Faculty Mentor Name
Dr. Bryan Watson
Abstract
As multi-agent systems grow increasingly complex, physical robotic swarms are essential to bridge the gap between limited software simulations and real-world application. The BID4R STARS swarm provides a physical platform for testing diverse algorithms, yet its success relies heavily on fundamental agent capabilities—most notably, obstacle avoidance. Preventing intra-swarm collisions is critical to avoid hardware damage and experimental disruption. To address this challenge, this project implements the Reciprocal Velocity Obstacles (RVO) algorithm onto the STARS swarm. While standard collision avoidance algorithms often overcorrect and induce oscillatory agent movement, RVO factors in the velocity and anticipated responses of all agents involved in a potential collision. By assuming reciprocal reactions between robots, RVO generates significantly smoother motion trajectories. The successful integration of RVO will reliably prevent collisions while maintaining fluid swarm movement, establishing a robust, damage-free testing environment. This foundational improvement ensures the ongoing viability of the STARS swarm and enables future physical research on advanced multi-agent solutions.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Included in
Multi-Vehicle Systems and Air Traffic Control Commons, Navigation, Guidance, Control and Dynamics Commons, Robotics Commons
Improving Multi-Agent Swarm Collision Avoidance using Reciprocal Velocity Obstacles
As multi-agent systems grow increasingly complex, physical robotic swarms are essential to bridge the gap between limited software simulations and real-world application. The BID4R STARS swarm provides a physical platform for testing diverse algorithms, yet its success relies heavily on fundamental agent capabilities—most notably, obstacle avoidance. Preventing intra-swarm collisions is critical to avoid hardware damage and experimental disruption. To address this challenge, this project implements the Reciprocal Velocity Obstacles (RVO) algorithm onto the STARS swarm. While standard collision avoidance algorithms often overcorrect and induce oscillatory agent movement, RVO factors in the velocity and anticipated responses of all agents involved in a potential collision. By assuming reciprocal reactions between robots, RVO generates significantly smoother motion trajectories. The successful integration of RVO will reliably prevent collisions while maintaining fluid swarm movement, establishing a robust, damage-free testing environment. This foundational improvement ensures the ongoing viability of the STARS swarm and enables future physical research on advanced multi-agent solutions.