Philipp Clausen

A general approach to time-varying parameters in pose-graph optimization

Pose-graph optimization is becoming popular as a tool for solving position and attitude determination problems, especially in the context of Visual Simultaneous Localization and Mapping (V-SLAM). Recently proprioceptive information sources are appearing in this context, such as inertial measurement units and kinematic/dynamic models. These models require other quantities to be estimated along with camera poses and landmark 3D positions. Examples are IMU bias processes, friction coefficients and other process modeling parameters. In this work we propose a general approach to the estimation of time varying parameters in pose-graph optimization: we store parameter samples at arbitrary rate in auxiliary vertices and we employ interpolation schemes to recover their value at sensor readings timestamps. Prior knowledge or stochastic process models can be plugged in as additional edges incident in parameter nodes. Our approach is evaluated in the context of inertial navigation, where accelerometer and gyroscope bias processes need to be properly modeled and estimated.

Construction of dynamically-dependent stochastic error models, 2018 IEEE/ION Position, Location and Navigation Symposium (PLANS)

Stochastic behavior of an instrument is often analyzed by constructing the Allan (or wavelet) variance signatures from an error signal. For inertial sensors, such a signature is conveniently obtained by recording data at rest. The analysis of this signal will result in noise-parameters adequate to such situation. Nonetheless, the value of the noise parameters may change under dynamics or other kind of external influences like for instance the temperature. In this research we study first the influence of the rotational dynamics on the signal of MEMS gyroscopes and then we show how to link this property to the noise-parameter estimation in a rigorous way by a modified version of the Generalized Method of Wavelet Moments (GMWM) estimator. The results of such analysis can then for instance be used in a Kalman filter, where the noise parameters are adapted according to such predetermined functional relationship between sensor noise and the encountered dynamics of the platform/sensor.