Skip to content

CONCEPT Cited by 1 source

HOV Lane (High-Occupancy Vehicle Lane)

An HOV lane (high-occupancy vehicle lane; colloquially carpool lane) is a lane on a road network reserved for vehicles carrying multiple passengers (thresholds vary by jurisdiction: 2+, 3+, sometimes with EV / transit / taxi exemptions and time-of-day restrictions). The design intent is to move more people per lane-hour than a general-purpose lane during peak congestion by rewarding higher-occupancy trips with lower travel time.

From a distributed-systems / routing-product perspective HOV lanes are interesting because they create a heterogeneous road-network graph: two trips over the same geographic edge can have materially different travel times depending on per-trip eligibility (occupancy, time of day, vehicle class). Any system that predicts trip time (see concepts/estimated-time-of-arrival) or ranks routes over that graph has to encode the HOV/non-HOV distinction as a feature and serve different predictions for different users.

Canonical data point

  • Utah Salt Lake Valley (cited in Google Research's 2025-06-30 post): average speed 68.18 mph in HOV lanes vs. 58.60 mph in general lanes, ~16% faster, during the measured peak window. This is the numerical justification for Google Maps specialising its ETA surface by lane type rather than folding HOV travel into a single lane-agnostic ETA (Source: sources/2025-06-30-google-hov-specific-etas-google-maps).

Consequences for routing / ETA systems

  • Per-trip eligibility feature. The routing system has to know whether this trip on this road at this time is HOV-eligible. Google Maps' HOV ETA launch uses a trip classifier to make that determination.
  • Specialised ETA surface, not a single biased model. Folding HOV and non-HOV speeds into one ETA biases both directions (HOV over-predicted, general-lane under-predicted during peak). The launch splits the surface by classification (Source: sources/2025-06-30-google-hov-specific-etas-google-maps).
  • Jurisdiction-specific rules. Occupancy thresholds, time windows, and HOT-toll conversions vary by metro/state. The captured raw does not describe how Google localises the classifier or ETA surface; noted as a caveat.

Seen in

Last updated · 200 distilled / 1,178 read