Receding-Horizon Nullspace Optimization for Actuation-Aware Control Allocation in Omnidirectional UAVs

A team of researchers from institutions including the University of Padova and New York University Abu Dhabi has developed a new control algorithm that solves a critical problem for advanced omnidirectional drones. These 'fully actuated' UAVs can move in any direction and rotate independently, but their complex motor systems create asymmetric dynamics that cause oscillatory, jerky commands during fast maneuvers. The team's novel 'Receding-Horizon Nullspace Optimization' method treats this as a constrained optimal control problem, using a predictive model to simulate the closed-loop system's behavior over a future time horizon.

By anticipating how the motors will respond, the algorithm can proactively smooth out the commands sent to each rotor, redistributing the workload while still achieving the exact force and torque (the 'body wrench') required for flight. The optimization is performed online using a Constrained iterative Linear Quadratic Regulator (iLQR) solver. In simulations on an 'OmniOcta' platform, this approach drastically cut motor command oscillations compared to standard quadratic programming allocators, leading to more precise tracking of both position and orientation during dynamic flight. This represents a major step toward making these highly agile drones stable and reliable enough for real-world interaction tasks, from inspection to manipulation.