Enhancing Gyrocompassing Performance Using Low-cost Optical Sensors

Many applications require automatic and accurate determination of true north. While this can be accomplished in a number of ways, gyrocompassing (i.e., determining the orientation of the sensor by observing the rotation of the Earth) is a popular technique for tactical applications due to the high accuracy possible and the inherent immunity to external electromagnetic disturbances. Unfortunately, accurately detecting the Earthas rotation rate requires a relatively high-performance gyroscope with well-calibrated error characteristics. This results in an engineering trade between alignment time, accuracy, and cost. In this paper, a completely passive, lightweight gyrocompassing sensor is proposed that combines a low-cost optical sensor and automatic feature tracking techniques via an unscented Kalman filter algorithm to estimate true north. The approach is validated using a non-linear Monte Carlo simulation and conclusions are drawn regarding improvements in gyrocompassing performance as a function of sensor error characteristics and imaging conditions.