Wind Compensation in Drones using PID Control Enhanced by Extended Kalman Filtering

Several sectors, including relief, precision farming, logistics, and security, have started using drones. Their performance is often impacted by external factors, and among these, winds are the most impactful as they negatively affect aerodynamic efficiency, flight stability, and trajectory accuracy. Although H-infinity controllers could address the problem as a solution for disturbances, the amount of computation is large and therefore cannot fit into light-weight platforms such as drones that need lean and fast processing in real time. We propose a lightweight control scheme consisting of an Extended Kalman Filter (EKF) and a PID controller to address wind disturbance compensation. The drone dynamics are nonlinear; the EKF acts as a refining mechanism that improves input from noisy IMU/GPS sensors while also modeling wind forces as a state variable for real-time estimation of disturbances. This estimation is then fed in a feedforward manner into the controller that proactively rejects wind disturbances before they reach flight level. Simulated investigations on a 1.5 kg quadrotor subjected to sinusoidal and random wind disturbances have shown that PID+EKF has reduced the root-mean-square trajectory error by about 45% compared to that with the PID controller alone and has also reduced the velocity settling time to almost half of that with the PID controller.