LADRI: LeArning-based Dynamic Risk Indicator in Automated Driving System
Anil Ranjitbhai Patel, Peter Liggesmeyer
TL;DR
The work tackles the inadequacy of static risk analyses for ADS in dynamic traffic and proposes LADRI, a learning-based Dynamic Risk Indicator that uses real-time On-board Sensor data and Artificial Neural Networks to predict risk, outputting severity $S$ and controllability $C$. The framework accommodates multiple NN architectures and integrates data preprocessing, feature extraction, and cross-validated evaluation within a high-fidelity driving simulator, demonstrated on ACC scenarios with multi-sensor input. Results indicate LADRI can track rapid risk changes and move ADS toward higher safety levels, while noting the need to improve explainability for real-world deployment. Overall, LADRI offers a data-driven, adaptive approach to dynamic risk in ADS, bridging gaps left by traditional safety standards and enabling more timely, context-aware safety responses.
Abstract
As the horizon of intelligent transportation expands with the evolution of Automated Driving Systems (ADS), ensuring paramount safety becomes more imperative than ever. Traditional risk assessment methodologies, primarily crafted for human-driven vehicles, grapple to adequately adapt to the multifaceted, evolving environments of ADS. This paper introduces a framework for real-time Dynamic Risk Assessment (DRA) in ADS, harnessing the potency of Artificial Neural Networks (ANNs). Our proposed solution transcends these limitations, drawing upon ANNs, a cornerstone of deep learning, to meticulously analyze and categorize risk dimensions using real-time On-board Sensor (OBS) data. This learning-centric approach not only elevates the ADS's situational awareness but also enriches its understanding of immediate operational contexts. By dissecting OBS data, the system is empowered to pinpoint its current risk profile, thereby enhancing safety prospects for onboard passengers and the broader traffic ecosystem. Through this framework, we chart a direction in risk assessment, bridging the conventional voids and enhancing the proficiency of ADS. By utilizing ANNs, our methodology offers a perspective, allowing ADS to adeptly navigate and react to potential risk factors, ensuring safer and more informed autonomous journeys.