RVC-NMPC: Nonlinear Model Predictive Control with Reciprocal Velocity Constraints for Mutual Collision Avoidance in Agile UAV Flight

TL;DR

The paper addresses mutual collision avoidance for agile, high-speed UAVs in shared airspace by integrating time-dependent reciprocal velocity constraints into nonlinear model predictive control (NMPC). It relies on observable state information (positions and velocities) of other UAVs and augments NMPC with a PMM-based reference trajectory generator, enabling 100 Hz real-time operation with full nonlinear quadrotor dynamics. The main contributions are the time-validated RVCs derived from current states, their embedding as soft linear constraints in NMPC, and extensive simulations and real-world tests showing improved flight time (up to 31% faster) while maintaining collision-free navigation, plus robustness to latency and estimation noise. The approach demonstrates practical viability for high-speed multi-UAV scenarios without the need for full trajectory sharing or centralized planning, though it lacks formal collision guarantees and assumes relatively simple (obstacle-free) environments.

Abstract

This paper presents an approach to mutual collision avoidance based on Nonlinear Model Predictive Control (NMPC) with time-dependent Reciprocal Velocity Constraints (RVCs). Unlike most existing methods, the proposed approach relies solely on observable information about other robots, eliminating the necessity of excessive communication use. The computationally efficient algorithm for computing RVCs, together with the direct integration of these constraints into NMPC problem formulation on a controller level, allows the whole pipeline to run at 100 Hz. This high processing rate, combined with modeled nonlinear dynamics of the controlled Uncrewed Aerial Vehicles (UAVs), is a key feature that facilitates the use of the proposed approach for an agile UAV flight. The proposed approach was evaluated through extensive simulations emulating real-world conditions in scenarios involving up to 10 UAVs and velocities of up to 25 m/s, and in real-world experiments with accelerations up to 30 m/s$^2$. Comparison with state of the art shows 31% improvement in terms of flight time reduction in challenging scenarios, while maintaining a collision-free navigation in all trials.

Figures (8)

• Figure 1: Deployment of the introduced RVC-NMPC approach in a real-world scenario with 3 UAVs navigating to antipodal positions on a circle with radius 10m, and acceleration limit 30ms.
• Figure 2: Block diagram representing a single robot control and navigation pipeline for robot $i$ including the proposed approach for mutual collision avoidance for agile UAV flight.
• Figure 3: The illustration of the introduced time validity of the reciprocal velocity constraints.
• Figure 4: Qualitative comparison of trajectories generated by individual approaches in scenario involving 10 UAVs navigating to antipodal positions on the circle of radius 10m with velocity and acceleration limits 20ms and 40ms, respectively.
• Figure 5: The success rate (shown in colors of the matrix) and minimum mutual distance between uav (shown as numbers in the matrix [m]) under varying delay and frequency of messages obtained from other robots. The results for every delay-frequency pair are based on 100 flights involving 4 uav in APCX scenario.

Theorems & Definitions (1)

• Remark