Race safety
Pilot conceptUnder evaluation2027 Live Race-Safety Operations Pilot
Cadence Labs is proposing a limited live road-stage pilot that gives race control earlier awareness of selected hazards and convoy risks, while preserving the relevant evidence for human-led incident review.
This page describes a proposed pilot concept. It does not represent an existing UCI deployment, endorsement, or regulatory decision.
- The need
- Every safety measure depends on knowing quickly what has happened on the road, and on a reliable record afterwards.
- The solution
- Camera- and sensor-based units on selected official vehicles, designed to identify and flag selected safety-relevant events for human review, with officials in control throughout.
- The ask
- A 30-minute call to explore one selected road stage of an established professional event in 2027, with no commitment to adopt.
Why now
The UCI and its partners have built a serious safety framework: SafeR, the incident database, yellow cards for riders and convoy drivers, video review, and the move to real-time tracking. All of it relies on knowing quickly what has happened, and having a reliable record afterwards. Both are still hard to guarantee:
- tracking shows where a rider or vehicle is, not what is happening around them;
- much of what race control hears still comes by radio;
- reviewing an incident often depends on where a television camera happened to be.
The proposed solution
Camera- and sensor-based units on selected official cars and motorbikes. The proposed design would combine forward- and rear-facing cameras, positioning and motion sensors with on-board software designed to identify and flag selected safety-relevant events for human review. Local recording and buffering would be investigated so relevant evidence can be retained during temporary connectivity loss and synchronised when service returns.
Candidate use cases for pilot evaluation
| Use case | Alerted in real time | What it enables |
|---|---|---|
| Crash-like event or rider down | Under an agreed race-control protocol | Earlier awareness with location and context |
| Unexpected or potentially mobile vehicle | Authorised safety personnel | Assessment before riders reach the location |
| Suspected road contamination: oil, diesel, spills | Race control for human confirmation | Location flagged for response and review |
| Obstacles: debris, barriers, people, animals | Race control and relevant personnel | Earlier assessment, clearance, or warning |
| Road-surface anomalies, explored through motion data | Organiser for review | Structured route and safety learning |
| Potentially unsafe proximity or overtaking | Authorised officials | Objective information for official review |
| Rider-to-vehicle distance, where confidence is sufficient | Authorised officials | More consistent distance information, subject to validation |
| Near misses | Relevant safety stakeholders after the event | Measures targeted and effect assessed |
What the pilot could give the UCI
Earlier warning
of hazards
Faster response
to incidents
Reviewable evidence
for human-led incident assessment
Proximity information
for authorised official review
Near misses on record
not only crashes
Alert routing agreed with race control
- Step 1Flaggedby a unit on an official vehicle, with confidence and data-quality information.
- Step 2Sent livethrough an agreed race-control and safety protocol, with location and relevant video evidence.
- Step 3Reviewed or escalatedby the authorised personnel responsible for the relevant response.
- Step 4Acted on by officialswho decide what to pass on and how the event is recorded for review.
A live alerting system, not only a post-event log. Vehicles driving ahead of the race could flag selected hazards before riders reach them. The proposed design includes local buffering for temporary connectivity loss, with relevant alert metadata and associated video retained under agreed access, retention, and data-governance arrangements.
Officials stay in control
Human in the loop by design. Officials set the level of oversight for each alert type, and can change it during the race:
Alerts are reviewed centrally before escalation under the agreed operating procedure.
Predefined high-priority alerts may be routed directly to authorised safety personnel, subject to organiser and official approval.
At every level, a person decides every action. The system issues no penalties and never acts on its own.
Also: no rule change for the pilot · nothing fitted to riders' bikes or team cars · works alongside existing tracking and timing · shaped with teams, riders and organisers, including data ownership, retention and privacy.
The 2027 live road-stage pilot
One selected road stage of an established professional event with a willing organiser, preceded by closed-road testing and controlled validation to reduce deployment risk. Vehicle-to-rider distance measurement would be investigated only in selected scenarios where accuracy and data-quality thresholds are met.
Does race control learn of hazards and incidents sooner?
Can officials review every flagged incident with clear footage?
Does it lighten the load on race control?
- Outcome
- A written report suitable for review by the UCI and relevant safety stakeholders, recommending whether to continue, change, expand or stop. No commitment to adopt, and no SafeR involvement is assumed at this stage.
- Timeline
- Shape now to early 2027 · Validate before the live-stage deployment · Test one selected road stage during the 2027 season · Decide after the independent review
Next step
A 30-minute call with the right UCI, SafeR, race-operations or technology contact, to decide whether this is a useful problem to define and test together for 2027.
About Cadence Labs. We build cycling data systems in which every figure can be traced to its source. Founder Edward Mellor PhD previously led the delivery of decision systems for critical infrastructure.
This page describes a proposed pilot concept. It does not represent an existing UCI deployment, endorsement, or regulatory decision.