Orientation Sensing for Helicopter UAVs under Strict Resource Constraints

For flight control of helicopter UAVs, it is essential to have high-quality 6-DOF attitude measurements. In the course of the miniaturization, the need for reliable orientation sensors conflicts with strict resource constraints. This paper presents the algorithms for orientation calculation used on-board TU Berlin’s helicopter UAV MARVIN as an example of “low-resource” orientation measurement. While rotation rate sensors provide short-term data, acceleration and magnetic field sensors are fused for delayed on-line calibration, using GPS to eliminate kinematic acceleration effects. All maths are performed using 32 bit fixed point arithmetics, involving certain “crude” simplifications. The algorithm’s performance and computational complexity are compared to a full Kalman Filter fusion of the sensors, justifying the use of the specialized and simplified approach.