An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving

TL;DR

This paper offers an application-driven framework for understanding and detecting corner cases in highly automated driving, integrating multiple sensor modalities and defining a new method layer for corner cases.

Contribution

It extends existing corner case definitions to include RADAR and LiDAR sensors and introduces the concept of method layer corner cases based on data and methodology uncertainty.

Findings

Extended corner case systematization to include RADAR and LiDAR sensors.

Described a toolchain for data acquisition and processing for corner case detection.

Introduced the concept of method layer corner cases related to uncertainty.

Abstract

Systems and functions that rely on machine learning (ML) are the basis of highly automated driving. An essential task of such ML models is to reliably detect and interpret unusual, new, and potentially dangerous situations. The detection of those situations, which we refer to as corner cases, is highly relevant for successfully developing, applying, and validating automotive perception functions in future vehicles where multiple sensor modalities will be used. A complication for the development of corner case detectors is the lack of consistent definitions, terms, and corner case descriptions, especially when taking into account various automotive sensors. In this work, we provide an application-driven view of corner cases in highly automated driving. To achieve this goal, we first consider existing definitions from the general outlier, novelty, anomaly, and out-of-distribution detection to show relations and differences to corner cases. Moreover, we extend an existing camera-focused systematization of corner cases by adding RADAR (radio detection and ranging) and LiDAR (light detection and ranging) sensors. For this, we describe an exemplary toolchain for data acquisition and processing, highlighting the interfaces of the corner case detection. We also define a novel level of corner cases, the method layer corner cases, which appear due to uncertainty inherent in the methodology or the data distribution.