Research & Papers

RCOA extends convex obstacle avoidance to 3D UAV navigation

New convex formulation enables drones to fly through narrow gaps at 30+ Hz.

Deep Dive

The Relaxed Convex Obstacle Avoidance (RCOA) formulation, introduced as the first fully convex optimal control approach for obstacle avoidance, has been extended to three-dimensional environments and applied to Unmanned Aerial Vehicle (UAV) navigation by researchers Ricardo Tapia and Iman Soltani. A key innovation is RCOA's unique property derived from convergence analysis: it maintains obstacle avoidance effectiveness even when obstacles lie beyond the controller's prediction horizon. This new extension also moves beyond point-mass representations by incorporating realistic vehicle geometries, enabling collision avoidance between 3D objects rather than simple points.

Numerical simulations demonstrate that RCOA delivers computational performance matching or exceeding current state-of-the-art methods. Critically, the approach enables a Nonlinear Model Predictive Controller (NMPC) to execute aggressive maneuvers through narrow passages with reduced prediction horizons while maintaining real-time feasibility at frequencies exceeding 30 Hz. This makes RCOA particularly attractive for agile drone applications in cluttered environments, such as search-and-rescue or industrial inspection, where fast, safe navigation is essential.

Key Points
  • First convex optimal control formulation for obstacle avoidance that works beyond prediction horizon
  • Extended to 3D UAV navigation with realistic vehicle geometries, not just point-mass models
  • Achieves real-time NMPC at over 30 Hz, enabling aggressive maneuvers through tight spaces

Why It Matters

Enables safer, faster drone navigation in cluttered environments with real-time convex optimization.

📬 Get the top 10 AI stories daily