From Inertial Measurements to Smart Cities

Autonomous vehicle technology is advancing rapidly. Their control capabilities often rely on high-bandwidth state estimation incorporating inertial measurements. High-performance state estimation incorporates inertial measurement error models through the process of state augmentation to enable on-the-fly instrument calibration. The feature article Inertial Measurement Unit Error Modeling Tutorial by Jay A. Farrell, Felipe O. Silva, Farzana Rahman, and J. Wendel provides a tutorial describing the process and issues related to developing a state-space model for the stochastic errors affecting an inertial measurement unit (IMU). The starting point is the instrument error characterization data sheet provided by the manufacturer, which is typically either an Allan standard deviation graph or the Allan variance parameters extracted from that graph. The desired output of the modeling process is a linear, discrete-time, state-space model of the IMU stochastic errors suitable for augmentation to the INS error-state model. Along with this tutorial, supplementary open source software is available. One software component does the following: 1) Given a continuous-time state-space IMU stochastic error model selected by the designer to match the Allan variance, the software computes a discrete-time equivalent state-space model. 2) Given that discrete-time model, it produces a stochastic error sequence suitable for Allan variance computations. 3) Given a sequence of stochastic errors, it computes and plots the Allan variance. Given Allan variance data and a specific continuous-time state-space IMU stochastic error model structure, a second software component implements an optimization-based approach for selecting the model parameters to match the Allan variance data.