The minimal risk manoeuvre (MRM) is the manoeuvre which is applied in case an ADF can no longer perform the driving task or / and the driver does not respond to take over requests (TOR). The general objective of the vehicle’s manoeuvre is to reach the safest possible state in the given situation and minimising risks in traffic. It is not possible to define one single MRM for all types of ADF. While designing the ADF, it’s essential to establish an appropriate MRM concept, taking into account the potential for diverse reactions, depending on the specific driving situation and condition. It should outline the situations in which the MRM should be activated and when it should remain inactive. Furthermore, the concept must ensure the safe operation of the MRM (FuSa and SOTIF). The analysis should not be limited to the ego vehicle but should also consider surrounding traffic and other road users.
Depending on the TOR status (not issued, issued and noted, issued and not noted), automation level (L3 or L4), and driver reaction (no reaction, reaction), there could be different characteristics for initiating or not initiating an MRM. The TOR, a critical factor for an L3 or L4 ADF with dependencies on MRM, must be carefully considered and designed to reduce the likelihood of needing to initiate the MRM. For an L3 function, where a driver is assumed to be able to take over at any time, the MRM strategy could be relatively simple. (UN ECE ALKS 2021), mandates the initiation of MRM in the event of a failed TOR or severe ADF, vehicle failures. The MRM concept should also include aspects of notifying the driver and other road users, and the maneuver itself should be planned safely to avoid posing additional risks to the occupants and other road users. This aspect is also relevant when considering SOTIF.
Additional information with examples for MRM and minimal risk condition (MRC) please check:
- UNECE R157 (2021), “Uniform provisions concerning the approval of vehicles with regard to Automated Lane Keeping Systems”
- (Wood et al. 2019),“Safety first for automated driving”, and the related ISO technical report ISO/PRF TR 4804 (2020).
Ongoing standardisation acitvities:
- ISO/FDIS 23793-1 Intelligent transport systems — Minimal risk manoeuvre (MRM) for automated driving
- Under Proposal ISO/NP 23793-2 (Ed 2) Intelligent transport systems — Minimal risk manoeuvre (MRM) for automated driving — Part 2: Road shoulder stop — Minimum requirements and test procedures.
Main Question
Is an adequate and validated concept for MRM defined?
Sub-Questions
- Is a concept for the MRM in the ADF foreseen (e.g. degradation, take-over)?
- Is there an appropriate mechanism for a safety related fallback solution for the ADF considered?
- Does the concept consider to bring the vehcile to a safe stop if TOR leads to no appropriate reaction from the driver for an L3 ADF?
- Is the concept defined for different driving situations and conditions?
- Is the targeted / final MRC defined?
References
- ISO (2020) TR 4804: Road vehicles — Safety and cybersecurity for automated driving systems — Design, verification and validation. Available at: https://www.iso.org/standard/80363.html (Accessed: 18 October 2023)
- UN Regulation No. 157 – Automated Lane Keeping Systems (ALKS). Available at: https://unece.org/transport/documents/2021/03/standards/un-regulation-no-157-automated-lane-keeping-systems-alks (Accessed: 12 February 2024)
- ISO FDIS 23793-1: Intelligent transport systems — Minimal risk manoeuvre (MRM) for automated driving Part 1: Framework, straight-stop and in-lane stop.(Under development) Available at: https://www.iso.org/standard/81711.html (Accessed: 22 May 2024)