MDS 2.1 adds AV-specific incident reporting, real-time policy push, and enforcement data structures — moving the governance toolkit closer to autonomous-vehicle oversight

Headline Finding

MDS 2.1, in final review as of March 2026 after more than 1,000 days of development, introduces a dedicated incidents-and-crashes endpoint covering AV-specific events — remote takeovers, ADS disengagements, near misses, harsh braking, and collisions — alongside a real-time policy-push mechanism capable of sub-one-second geofence updates and new enforcement-and-violations data structures. Together these features constitute the first version of MDS explicitly architected for for-hire autonomous vehicle oversight, giving cities a permit-enforceable channel to receive safety telemetry from AV operators and to broadcast emergency operational rules in real time. The release is a material step in closing the gap between the governance clock and the deployment clock for AVs.

Detail

The Open Mobility Foundation’s MDS Working Group held a European-focused session on 26 March 2026 to review the forthcoming MDS 2.1 release, which has been in development for over 1,000 days — described by OMF Director of Open Source Operations Michael Schnuerle as the longest release cycle in MDS history. The session drew 63 attendees from cities including Amsterdam, Bergen, Oslo, Stockholm, Columbus (OH), Philadelphia, and New York City, as well as operators, vendors, and research institutions across Europe and North America.

The headline governance-relevant additions in MDS 2.1 are: (1) an incidents-and-crashes endpoint that can capture, for any mode, unplanned stops, remote takeovers, ADS disengagements or engagements, harsh braking and acceleration, tip-overs, vandalism, theft, near misses, and collisions — each with location, date, time, severity, medical-response flag, accelerometer data before and after the incident, and a report ID linking to an external system; (2) explicit for-hire autonomous vehicle support, clarifying that cities can broadcast MDS policy as situational-awareness signals to AV fleets in emergencies; (3) real-time policy push, upgrading from a pull-only model to a POST endpoint that can deliver geofence or policy changes to any number of operators in as little as one second; and (4) enforcement-and-violations data structures allowing operators to self-report citation IDs, actions taken, warning violations, and associated costs, with the same schema mirrored in the Curb Data Specification for cross-standard consistency.

The City of Amsterdam was cited as a live precursor: it is already receiving near-miss and harsh-braking data from Cowboy, a private bike operator, feeding a dashboard used to prioritise infrastructure improvements before crashes occur, and intends to adopt the MDS 2.1 incident endpoint once approved. Blue Systems was cited as having built emergency geofencing for for-hire AV fleets that translates dispatch-system alerts into MDS and pushes them to autonomous vehicles in the field, with a multi-jurisdictional implementation in London. The OMF noted that MDS is used by 200 commercial mobility service providers operating in over 1,200 cities and dozens of countries, with at least 2 billion trips tracked. MDS 2.1 is a non-breaking minor release: operators on 2.0 can declare 2.1 compliance without changes, and adopt new endpoints selectively. The approval process — working group, steering committee, technology council, board — was described as beginning after the European feedback period, with a final review working group meeting scheduled for 9 April 2026.

NYC DOT data scientist Matt McBride noted in the session chat that New York City legislation (NYC Council file referencing legistar ID 7861511) is currently in progress and that the city hopes to specify MDS standards in the rulemaking process, seeking OMF guidance on how to embed the standard in agency rulemaking or legislation. POLIS Secretary General Karen Vancluysen confirmed that earlier European efforts to create a domestic alternative to MDS were abandoned once MDS achieved full GDPR compliance, and that POLIS is running a tutorial series with OMF support to build city-side data skills across European member cities.

Governance Implication

1. The incidents-and-crashes endpoint moves the legitimacy lever: for the first time, cities have a standardised, permit-enforceable channel to receive AV safety telemetry (remote takeovers, ADS disengagements, near misses, collisions) rather than relying on operator self-disclosure in unstructured formats — directly addressing the accountability gap that makes AV permitting politically contested.
2. Real-time policy push (sub-one-second geofence updates) moves the spatial-allocation lever: cities can now dynamically re-allocate right-of-way in emergencies without relying on email or phone calls to AV operators, making the public space a more actively managed resource rather than a static permit zone.
3. The enforcement-and-violations schema, mirrored in CDS, creates the data infrastructure for performance-based permitting — cities can condition AV operating licences on self-reported compliance records, shifting from input-based to outcome-based regulation.
4. NYC DOT’s live legislative effort to embed MDS in city rulemaking, and POLIS’s abandonment of a European alternative in favour of GDPR-compliant MDS, together signal that the standard is crossing from voluntary adoption to regulatory instrument — a legitimacy lever move with implications for operators who have not yet upgraded from MDS 1.x or 2.0.
5. The two-clocks reading: AV deployment is already outpacing governance in both the US and Europe (Waymo, Zoox, robotaxi pilots), but MDS 2.1’s AV-specific features represent the governance clock catching up at the instrument layer — cities now have the technical vocabulary to write AV data obligations into permits; the remaining gap is political will and implementation capacity, not specification.

Sources

https://github.com/openmobilityfoundation/mobility-data-specification/wiki/Web-conference-notes,-2026.03.26-(MDS-Working-Group) https://github.com/openmobilityfoundation/mobility-data-specification/tree/dev https://www.openmobilityfoundation.org/about/