A SYSTEM AND METHOD FOR MOTION CONTROL OF UNMANNED AERIAL VEHICLES

This patent application discloses a method and system for real-time control of an unmanned aerial vehicle (UAV) through joint optimization of positioning and control. The method involves receiving sensor measurements from onboard sensors and constructing a unified probabilistic factor graph that integrates positioning-related factors derived from the sensor data and control-related factors generated by a model predictive control (MPC) framework. The positioning factors account for spatial state estimation, while the control factors encode dynamic constraints and trajectory objectives. The unified factor graph is solved iteratively using factor graph optimization (FGO), enabling simultaneous optimization of the UAV’s state estimation and control inputs in a single computational framework. By unifying positioning and control, the system eliminates the need for manual parameter tuning to compensate for positioning uncertainties, as the FGO naturally adapts to varying state confidence levels. This integrated approach reduces positional tracking errors and minimizes orientation oscillations during flight, enhancing stability and precision. The system operates in realtime on multi-rotor UAVs, leveraging computational efficiency of FGO to maintain responsive control without latency. Key advantages include improved trajectory tracking accuracy, robustness to sensor noise, and reduced mechanical vibrations caused by orientation fluctuations. The invention enables autonomous UAV operation in dynamic environments with enhanced reliability and performance.