By Luis Lopez, AI transportation consultant, CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast
Most conversations about AI driver fatigue tools start with in-cab cameras that watch a driver’s eyes. That is one approach, but there is another that many fleets can start with using data they already collect: schedule-based risk scoring. Instead of watching the driver, it looks at the work pattern and asks whether the schedule itself is setting the driver up to be tired.
This article covers how schedule-based fatigue scoring works, what signals it uses, how dispatchers can act on it, and where the limits are.
Why Schedules Matter More Than People Think
Fatigue risk is not just about how many hours someone drives. It depends on when they drive, how much sleep opportunity they had, and how their schedule has changed over recent days. A driver who works the same daytime hours every day is in a different position from one who switched from days to overnights this week, even if both are within the same legal limits.
Human body clocks have low points, especially in the early morning hours. Driving through those periods, after a long day on duty, or after a short break between shifts all raise risk. Those patterns show up in schedule and route data well before anything shows up on a camera.
Legal Is Not the Same as Rested
Hours of service rules under 49 CFR Part 395 set the outer boundaries for property-carrying drivers: up to 11 hours of driving within a 14-hour window after 10 consecutive hours off duty, a 30-minute break after 8 cumulative hours of driving, and a 60-hour or 70-hour limit over 7 or 8 days. Electronic logging devices record driving time against those limits.
But HOS compliance does not by itself mean a driver is rested. Ten hours off duty is an opportunity to sleep, not a guarantee that sleep happened. A driver can be fully legal and still be at elevated risk. Our guide to hours of service for shippers covers the rules in more detail.
Signals a Schedule-Based Model Can Use
A fatigue risk score built from schedules and routes typically draws on several inputs:
- ELD duty status patterns. How long the driver has been on duty, how the 14-hour window is being used, and how much off-duty time came before the shift.
- Start-time shifts. A driver who started at different times on consecutive days, especially moving earlier or flipping from day to night, is adjusting to a new sleep pattern.
- Night and early morning driving. Hours driven during circadian low points carry more weight than the same hours in mid-afternoon.
- Dock waits. Long detention at shippers and receivers eats into the 14-hour window, so the driver ends up driving later in their day than planned.
- Consecutive long days. Several near-maximum days in a row add up, even when each individual day looks fine.
- Planned route and appointments. Upcoming loads that will force driving through the night or a tight turnaround after a short break.
AI helps by weighing these together, learning which combinations matter most for a specific operation, and projecting risk forward onto the next planned load rather than only scoring what already happened.
How Dispatchers Use the Score
A fatigue score is only useful if it changes a decision. As an AI transportation consultant, the first thing I look at is whether the score shows up where dispatch decisions are made, at the time they are made. A report reviewed the following week does not help the driver on the road tonight.
Practical actions a dispatcher can take when a load or driver is flagged:
- Reassign the load to a driver with a better rest pattern or schedule fit.
- Plan rest into the trip, such as a longer break before a night segment.
- Move appointments to avoid forcing driving through early morning hours, working with the shipper or receiver.
- Check in with the driver and ask, rather than assume, how they are doing.
Long dock waits are one of the most controllable drivers of risk. Our post on how to reduce truck detention time covers ways shippers and carriers can shrink them.
Driver Trust and Privacy
Any fatigue program fails if drivers see it as a tool for punishment. Scoring should support safety managers and drivers, not discipline them. That means being transparent about what data is used, how the score is calculated in plain terms, and what happens when a driver is flagged.
Good practices include:
- Explain the program to drivers before it goes live.
- Let drivers see their own scores and the reasons behind them.
- Make it acceptable for a driver to say they are too tired, without penalty.
- Limit who can see individual scores and how long the data is kept.
- Use flags to trigger conversations and schedule changes, not automatic disciplinary action.
The same principles apply to broader safety programs. Our article on AI driver safety scoring and coaching covers how to keep scoring constructive.
Data Limits
A schedule-based model sees the work pattern, not the person. It does not know whether a driver slept well during their off-duty period, whether they have a new baby at home, or whether they spent their break on a second job. It also depends on accurate logs. Duty status errors, unassigned driving time and edits all degrade the signal.
Never treat a score as proof that a driver is or is not fatigued. A low score does not mean a driver is alert, and a high score does not mean they are impaired. It means the schedule pattern carries more or less risk.
Measuring Whether It Works
Before you roll out a fatigue scoring program, decide how you will judge it. Useful measures include:
- How often flagged loads were actually changed, and how.
- Whether the share of driving in early morning hours goes down over time.
- Driver feedback on whether the flags match how they felt.
- Trends in safety events and near-misses, reviewed carefully and over enough time to be meaningful.
- Whether dispatch planning starts building in rest before the model has to flag it.
Start small: one terminal or one group of drivers, a simple set of signals, and a clear review process. Add complexity only when the basics are being used.
For more conversations on freight and AI, subscribe to the Freight Guru Podcast.
About the author: Luis Lopez is a Miami-based AI transportation consultant and logistics entrepreneur, the CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast.