This paper presents a low-pass filter and a complementary Kalman filter for improving the quality of UAV feedback control information by means of onboard UAV sensor fusion. The control of UAVs requires real-time feedback from onboard sensors to update the status of the UAV. Localisation is achieved by determining the UAV attitude and position. While the UAV position can be provided by sensors such as the GPS receiver, attitude estimation for UAVs is provided, mostly, by an IMU. IMU sensors are mainly the accelerometer, gyroscope, and magnetometer. These sensors have limitations in the accuracy caused by sensor drift and noise and thus needed post-processing for sensor data improvement. This paper focuses on correctly estimating roll and pitch angles using a Kalman filter and tests the performance of the estimated feedback angle information on a quadrotor platform.