A Model for the Design of Automated Vehicle Event Data Recorders

We propose a model for the design of a new generation of event data recorder (EDR) aimed for automated driving systems (ADS) to support crash analysis, risk assessment and accidents reconstruction. Current EDRs, or described in standards and regulations do not address detailed design requirements that are necessary by all SAE 3016 driving automation levels 1 through 5. These standards and regulations include IEEE 1616.1, SAE J1698, U.S. CFR, Title 49 Part 563, FMVSS No 405, UN regulation 160, and others. Traditional onboard EDRs are a devices that capture and store critical vehicle data for several seconds before, during and after any crash where an airbag is triggered or there is an excessive rate of vehicle deceleration. There are efforts to design new kinds of EDRs to capture and store data during and after crashes that involve automated vehicles. IEEE 1616.1 defines requirements and metrics for data storage in a data storage system for automated driving (DSSAD), as well as related functions and common technical requirements. IEEE defines a DSSAD as a device or function that records and stores a data set (“timestamped flags”) during the automated driving sequences of any vehicle equipped with Level 3, Level 4, or Level 5 automated driving systems (ADS). However, the IEEE standard does not characterize the ADS driving modes. The model characterizes the driving modes of an ADS Intended to mitigate the so-called driver or mode confusion about the actual status of the ADS automation. The generic model can be instantiated for each automation level L1 through L5 and includes requirements for the recording and storage of data in the following categories: a) Vehicle controls, b) Monitoring of the driving environment, c) Fallback when automation fails, and d) ODD (operational design domain). The paper will include a model instantiation for an ADAS system.