Learning-Based Error Detection System for Advanced Vehicle Instrument Cluster Rendering

TL;DR

This work tackles the challenge of detecting rendering errors in advanced vehicle instrument clusters where overlays and non-linear composition undermine traditional CRC-based checks. It introduces a telltale monitor that leverages an anomaly-detection pipeline on CNN-derived features, quantified via a feature reconstruction error (FRE) after PCA-based dimensionality reduction, with per-convolution PCAs to balance accuracy and compute. The system crops ROIs around telltales, computes an anomaly score, and uses a state-dependent threshold to decide OK vs NOK, while enhancements include masked scoring, per-element capping, and temporal smoothing. Experimental evaluation across six telltales and multiple error types demonstrates robust detection of perceptually significant errors with virtually no false alarms, and a safety analysis discusses undetectable-error risks and mitigation through multi-PCA configurations. Overall, the approach offers a scalable, perceptually aligned monitoring solution for modern instrument clusters that can tolerate overlays and background variation while maintaining safety-critical detection capabilities.

Abstract

The automotive industry is currently expanding digital display options with every new model that comes onto the market. This entails not just an expansion in dimensions, resolution, and customization choices, but also the capability to employ novel display effects like overlays while assembling the content of the display cluster. Unfortunately, this raises the need for appropriate monitoring systems that can detect rendering errors and apply appropriate countermeasures when required. Classical solutions such as Cyclic Redundancy Checks (CRC) will soon be no longer viable as any sort of alpha blending, warping of scaling of content can cause unwanted CRC violations. Therefore, we propose a novel monitoring approach to verify correctness of displayed content using telltales (e.g. warning signs) as example. It uses a learning-based approach to separate "good" telltales, i.e. those that a human driver will understand correctly, and "corrupted" telltales, i.e. those that will not be visible or perceived correctly. As a result, it possesses inherent resilience against individual pixel errors and implicitly supports changing backgrounds, overlay or scaling effects. This is underlined by our experimental study where all "corrupted" test patterns were correctly classified, while no false alarms were triggered.

Figures (20)

• Figure 1: Example display using an overlay of a telltale on a camera image lexusMirror. Classical error detection methods are hardly applicable for these emerging rendering effects. This paper proposes a novel and very effective solution.
• Figure 2: Illustration of display composition including some example telltales, navigation and speedometer. Flash indicates possible error source.
• Figure 3: Twice the same telltale, but on the right laid over a grey background result in different CRC check values
• Figure 4: Components involved to compose, verify and visualize data on a display. In this work we address errors within the video composition step.
• Figure 5: Components of our proposed telltale monitor.