A GPS aided Full Linear Accelerometer Based Gyroscope-free Navigation System

This paper reports on a GPS aided, full linear accelerometer based, gyroscope-free inertial navigation system (INS), which permits an almost arbitrary placement and orientation of the acceleration sensors within the moving body. The objective of this approach is to reuse different accelerometers already existing in e.g. a car and thus reducing the costs of an INS. Besides navigation, this system can also be used for car safety applications like roll-over detection. In this field, calculation of the center of rotation in both, the body and the navigation frame of reference, is of great interest, which is an additional benefit of the system. The INS itself consists of at least 12 sensitive axes being necessary to calculate the entire body movement. In contrast to previous approaches, this is done by a simple matrix multiplication reducing the calculating effort. Simulations as well as measurement results verify this concept. Due to the sensor noise and the necessary integration steps, the variance of the position and orientation, meaning the uncertainty, increases with time. This problem is solved by coupling the INS with the long time stable GPS through a predictive filter. In contrast to a conventional gyroscope based INS, this approach offers additional information about the body movement, which is used in the filter to improve the overall performance. Due to the nonlinear behavior of the orientation, an Unscented Kalman Filter using a third order Wiener Process Acceleration Model is used. The adaptation of the covariance matrix is done by a fictive trajectory. With a GPS standard deviation of 10 m having a 1 Hz update, a sensor noise of 5 mg for a 100 Hz bandwidth and a simulation interval of slightly more than 4 minutes, the simulated position and orientation standard deviation are 4 m and 2°, respectively.