A denoising and drift-control approach for UAV trajectory tracking

As part of dead reckoning, inertial navigation has been developed for many years in locating the vehicle without GPS. We employ the technology on a quadrotor platform to track the trajectory of the vehicle. In this paper, a new method was designed to make the tracking result more precisely. To achieve this, we firstly smooth the sensor data by minus offset dynamically and adopt Kalman Filtering algorithm. Then we detect velocity and correct the wrong value in time to decrease drift in displacement. This is an important step in locating the vehicle simply by odometry.