Integrated Real Time Trajectory Optimization and Chance-Constrained Model Predictive Control of UAVs 

In this work, we developed a parallelized framework for real-time trajectory optimization and collision-chance constrained Model Predictive Control of quadrotor UAVs.
🔍 Key Highlights:
Real-Time Trajectory Optimization: Minimizing maneuver time while considering the UAV's dynamics, bounds, and terminal constraints.
Collision Avoidance: Imposing a safety sphere around obstacles with unknown dynamics and designing an optimal trajectory to avoid them.
Chance-Constrained MPC: Designing acceleration commands, desired orientation, and angular velocity to constrain collision probability within a user-defined safety margin.
Parallelized Architecture: Parallelizing two optimizations, i.e., planner and MPC, to be running in real-time.

Наука