Definition
Difference in safety event rates between drivers using AI-recommended split-sleeper berth strategies and drivers using consolidated 10-hour off-duty rest periods on comparable routes and duty schedules.
The Measurement Gap
Most fleets track Hours-of-Service compliance as a binary: legal or violation. They monitor HOS violation rate to confirm drivers stay within regulatory limits, driver safety event frequency as a fleet-wide aggregate, and accident rate per million miles without segmenting by rest strategy. All three metrics treat rest as a pass/fail regulatory checkbox rather than a variable that affects post-rest performance.
This creates a dangerous blind spot. A fleet can achieve 100% HOS compliance while systematically recommending split-sleeper strategies that produce measurably worse safety outcomes than consolidated rest. The AI scheduling engine optimizes for duty-cycle utilization (maximizing available driving hours) without feedback on whether the rest pattern it recommended actually produced a rested, safe driver. Split-sleeper berth provisions (7/3 or 8/2 splits) are legal under federal HOS rules, but legality does not equal safety equivalence.
RSSD closes this gap by comparing post-rest safety performance between the two strategies on matched routes and conditions, transforming rest quality from a compliance artifact into a measurable safety input.
Decision Supported:
Are AI-recommended split-sleeper strategies producing equivalent safety outcomes to consolidated rest? Should the fleet restrict split-sleeper recommendations to specific route types, driver populations, or operating conditions, or eliminate them entirely when the safety differential exceeds acceptable thresholds?
Formula
Plain language: RSSD subtracts the safety event rate after consolidated 10-hour rest from the safety event rate after split-sleeper rest, measured during the first four hours of driving after restart on comparable routes. If split-sleeper drivers experience more hard braking events, lane departures, following distance alerts, drowsiness warnings, or preventable accidents per 100 miles than consolidated-rest drivers, RSSD is positive and quantifies the excess risk.
Notation:
RSSD = SafetyEventRate_split − SafetyEventRate_consolidated
Where:
SafetyEventRate_split = (Total safety events in first 4 hours post-split-sleeper restart) / (Total miles driven in first 4 hours post-split-sleeper restart) × 100
SafetyEventRate_consolidated = (Total safety events in first 4 hours post-consolidated restart) / (Total miles driven in first 4 hours post-consolidated restart) × 100
Variables:
| Variable | Description | Source |
|---|---|---|
| Safety events | Hard braking, lane departure, following distance alerts, drowsiness alerts, preventable accidents | Telematics / AI camera systems |
| Split-sleeper restart | Driving resumption after 7/3 or 8/2 split-sleeper berth period | ELD duty status records |
| Consolidated restart | Driving resumption after 10+ consecutive hours off-duty | ELD duty status records |
| First 4 hours | Measurement window capturing fatigue-sensitive driving period | ELD / telematics timestamp correlation |
| Miles driven | Distance traveled during measurement window | Telematics / ELD |
| Route matching | Comparable route type, terrain, traffic density for paired comparison | Routing engine / telematics |
A positive RSSD indicates split-sleeper rest produces worse safety outcomes. A negative RSSD indicates split-sleeper rest is associated with fewer events: possible when drivers self-select into splits that match their circadian preferences. Statistical significance testing (p < 0.05) is required before acting on results.
Proposed formula, pending validation pilots.