A Systematic Mapping Study on the Debugging of Autonomous Driving Systems
Nathan Shaw, Sanjeetha Pennada, Robert M Hierons, Donghwan Shin
TL;DR
This work presents the first systematic mapping study of debugging in Autonomous Driving Systems (ADS), addressing a critical gap beyond established ADS testing surveys. It defines a structured taxonomy for ADS debugging, covering problem categories (simplification, localisation, explanation, repair) and techniques (manual analysis, causal analysis, fault injection, iterative minimisation, ML-based debugging), and maps 15 primary papers to these categories. The study highlights a strong current emphasis on simulation-based development and root-cause analysis, with fewer empirical post-deployment studies and limited automated code-level localisation. It also catalogs tools and frameworks for ADS debugging, discusses notable trends and threats to validity, and outlines future directions such as handling non-unique causality, advancing repair strategies, and proposing general ADS debugging frameworks to standardize terminology and practice. Overall, the SMS provides a foundational, structured view of ADS debugging research, informing researchers and practitioners about gaps, tools, and promising directions to advance safe and reliable ADS debugging in the path toward commercial deployment.
Abstract
As Autonomous Driving Systems (ADS) progress towards commercial deployment, there is an increasing focus on ensuring their safety and reliability. While considerable research has been conducted on testing methods for detecting faults in ADS, very little attention has been paid to debugging in ADS. Debugging is an essential process that follows test failures to localise and repair the faults in the systems to maintain their safety and reliability. This Systematic Mapping Study (SMS) aims to provide a detailed overview of the current landscape of ADS debugging, highlighting existing approaches and identifying gaps in research. The study also proposes directions for future work and standards for problem definition and terminology in the field. Our findings reveal various methods for ADS debugging and highlight the current fragmented yet promising landscape.