As regulators define Data Storage Systems for Automated Driving (DSSAD) to support post-incident investigation and continuous compliance monitoring, one key technical question remains unresolved within international discussions: what camera image resolution is sufficient for investigatory purposes? Within international ADS safety working groups, visual image resolution for DSSAD remains under active debate. Unlike ADS perception systems—which rely on high-resolution, multi-sensor fusion—DSSAD visual recording is expected to be limited primarily to camera imagery due to data volume and accessibility constraints. The purpose of DSSAD imagery is not to replicate ADS perception, but to enable investigators to reconstruct context and assess compliance, responsibility, and potential negligence. The talk will explore: • What different camera resolutions actually look like in driving scenarios • What safety-relevant information can realistically be extracted from different recording resolutions • At what point resolution degrades evidentiary reliability in incident analysis • How lighting, weather, motion blur, and vehicle speed influence evidentiary clarity • What field of view is necessary for fair and accurate investigation • What “at least as safe as a human driver” implies for visual recording standards The session will also examine: • How resolution affects the ability to assess traffic rule compliance, signs, lane markings, and hazard recognition • The comparison between camera recording and human visual perception • Field-of-view considerations, including future ADS manoeuvres such as lateral repositioning and bi-directional operation.

• What resolution is sufficient for DSSAD’s investigatory purpose
• The trade-offs between data burden and evidentiary robustness
• What constitutes “fit for purpose” visual recording under the principle that ADS should be at least as safe as a competent human driver