kradar_026frame 60
SENSI · uncertainty in perception systems

How far can we trust
what a vehicle perceives?

A sensor resolves less with range. The reference it is judged against is drawn by people reading that same degraded evidence. A model is characterised only under the conditions it was evaluated on, and the road supplies others. A system assembled from all four is given control of the vehicle.

4datasets
1,128recordings
5uncertainties
Δ — mnearest matched pair

For perception, simulation and planning teams · K-Radar, nuScenes, Waymo, NVIDIA

enhanced referencedetection · class as claimedlidar return

Four sources, none of them bounded

Where does the uncertainty come from?

The sensor returns fewer points as range grows, fewer again in rain and occlusion. The reference is an instrument too: boxes a metre from the returns beneath them, one vehicle carrying four identities, a length inferred from a face no sensor saw. The model is characterised only where it was evaluated. The world supplies the rest.

These do not simply add. The annotator reads the same degraded returns the model does, so the reference is least certain exactly where the model is weakest. Measured against an unaudited reference, a residual understates the true error, and understates it most in the conditions that matter.

01

Enhance the reference

Test the human reference against the returns it was drawn from, correct what disagrees, and state what remains uncertain.

Hygiene · Posimetry
Morphometry · Motion · Size
02

Measure

Measure how far each model departs from that reference, across five channels that behave differently from one another.

Association
Detection · Uncertainties
03

Estimate

Fit those measurements into models of the error, conditioned on range, class, weather and occlusion.

Model lab
Features
04

Simulate

Drive microscopic simulation with that fitted error, so agents act on perception that fails the way real perception fails.

Traffic sim
05

Plan

Give motion planning the same model, so a decision can account for what perception does not promise.

Planning

What is measured

Five distributions, each conditioned on the situation.

A single score ranks models against each other and says nothing about the situation a vehicle is in. These are distributions over range, class, weather and occlusion, so how much room does this planner need beside a truck at sixty metres in rain has an answer with a number in it. All five derive from one association matched on position alone: a class-aware match would refuse a pedestrian↔cyclist pair and conceal the misidentification it exists to measure.

Localization
dx dy dz dr dθ dφ

Where the object is, against the enhanced reference.

Spatial
d_length d_width d_height

How large it is. Length is the worst of the three: a lidar sees one end and infers the other.

Detection probability
detected

Whether it is found at all, at this position.

Detection consistency
per reference track

Across a track, the fraction of frames in which it is found. A track is usually held throughout or lost early, so the mean hides two populations.

Identification
misidentified label_same

What it is called, on a pair matched by position alone.

The scene above is kradar_026 frames 6099, played at the sensor's own 10 Hz. Each sweep is decimated from about 56,511 os2-64 returns to 14,000. The white boxes are the finalized reference version; the dashed boxes are a detector's output on the same frames. The diagrams here define the five channels and carry no fitted values: those are produced per recording inside the application.

Why it has to reach the decision

An uncertainty that is never quantified is not absorbed. It is inherited by the tracker, the predictor, and the decision.

A tracker binds a detection to an identity, a predictor extrapolates it, a planner commits to the result, each treating an estimate as a fact. A simulator fed exact detections validates a vehicle that will never exist. Both need the same object: the error as a distribution over the conditions that produced it.