Accelerometer-Based Multi-Sensor INS

Inertial navigation systems for airborne applications use three axis accelerometers and three axis rate gyros, and in many cases additional inputs using GPS and other sensors. The objective of the proposed work is to replace the expensive rate gyros with rate estimates based on multiple low-cost accelerometers. The error characteristics of the accelerometers, chiefly the bias stability and broadband noise, limit the accuracy and hence the application of the resulting accelerometer-based multi-sensor inertial navigation systems (AMS-INS). The technical challenges are first to predict the fundamental limitations using detailed error analysis, and then to design the hardware and embedded software algorithms needed to maximize the available performance. The proposed algorithms include robust statistical estimation, online identification of bias error terms using Kalman filtering or neural networks, and self-initialization of the rate estimates using measurements of centripetal acceleration. The Phase I effort will include a simulation test bed for AMS-INS development, single axis measurements, and both analytical and Monte Carlo error analyses. A hardware and real-time software design using 15 accelerometers, temperature sensor, and an embedded DSP controller will be performed in the Phase I Option.