Sensor redundancy management/fault detection and isolation of the inertial measurement unit for a land-based vehicle's locating system

A sensor redundancy management/fault detection and isolation system is developed which utilizes inertial sensors, including accelerometers, gyros and inclinometers, applied to land-based vehicle locating systems. The locating system allows the central control of a mass transportation to accurately locate each vehicle's position at any instant.; The center piece of instrumentation of the vehicle locating system is the inertial measurement unit (IMU) that is composed of ten primary sensors, namely, five accelerometers and five gyros. These accelerometers and gyros are arranged in a pentad configuration to provide redundancy necessary for fault detection and isolation. The fault detection and isolation algorithm for the primary sensors is the Generalized Likelihood Test. This is aided with a parity vector compensation algorithm. A Kalman filter is designed to perform on-line estimation of normal sensor errors that affect the parity vector.; To determine the initial conditions in terms of the vehicle's attitude, two inclinometers are added to the unit as secondary sensors. Separate relationships are derived between inclinometer and accelerometer signals and they constitute analytical redundancy to be exploited for the failure detection of the secondary sensors. An error signature is developed to identify faulty accelerometer in a pentad redundant set based on inclinometer data.; A digital filter is designed to process the accelerometer signals to overcome the difficulty in comparing signals of varying frequency contents between accelerometers and inclinometers.; A field test using a non-redundant, orthogonal triad IMU was conducted at the Pennsylvania Transportation Institute's bus test track. The test data was used to synthesize a redundant data file for simulation. The simulation results show excellent performance of the fault detection and isolation system.; The major contributions of this research include: (1) Develop algorithms and design appropriate filter to allow mutual fault detection and isolation between primary and secondary sensors of varying frequency contents. (2) Apply the FDI algorithm to inertial sensors of current technology for land-based vehicle's navigation. (3) The integration of this combination of inexpensive primary and secondary sensors into a fault-tolerant IMU for safety-critical applications.