Error Decomposition for Hybrid Localization Systems

TL;DR

The Kappa-Phi method is proposed which allows for the decomposition of localization errors into individual components, i.e., into a sum of parameterized functions of the measured state, and can be leveraged to improve localization predictions, correct map data or calibrate sensor setups.

Abstract

Future advanced driver assistance systems and autonomous vehicles rely on accurate localization, which can be divided into three classes: a) viewpoint localization about local references (e.g., via vision-based localization), b) absolute localization about a global reference system (e.g., via satellite navigation), and c) hybrid localization, which presents a combination of the former two. Hybrid localization shares characteristics and strengths of both absolute and viewpoint localization. However, new sources of error, such as inaccurate sensor-setup calibration, complement the potential errors of the respective sub-systems. Therefore, this paper introduces a general approach to analyzing error sources in hybrid localization systems. More specifically, we propose the Kappa-Phi method, which allows for the decomposition of localization errors into individual components, i.e., into a sum of parameterized functions of the measured state (e.g., agent kinematics). The error components can then be leveraged to, e.g., improve localization predictions, correct map data, or calibrate sensor setups. Theoretical derivations and evaluations show that the algorithm presents a promising approach to improve hybrid localization and counter the weaknesses of the system's individual components.

Figures (6)

• Figure 1: Comparison of viewpoint and absolute localization. Absolute localization describes the ego-position with regard to a global reference, while viewpoint localization relies on matching local observations of the environment with a knowledge base (e.g., map).
• Figure 2: Hybrid Localization System consisting of two localizers and an error decomposition component. Potential external sources of the localizers include ephemeris correction data for GNSS receivers or map geometry data for viewpoint localization. The inputs of the decomposition filter are the estimates of the localizer as well as an input vector $\boldsymbol{u}$ that describes measured states (e.g., heading angle or time), according to the chosen error model.
• Figure 3: Example of two erroneous position estimates $\boldsymbol{p}_\upalpha$ and $\boldsymbol{p}_\upbeta$ (obtained from two localizers $\upalpha$ and $\upbeta$). The difference vector $\boldsymbol{d}$ is decomposed in a state-independent translatory component $\boldsymbol{\kappa}$ and a heading-dependent component $\boldsymbol{\phi}$.
• Figure 4: Comparison of a vehicle-inherent error, e.g., calibration (left), and an environment-inherent error, e.g., due to a map offset (right).
• Figure 5: GNSS path with two segments highlighted: a mostly straight segment as well as a segment with five turning situations.