Passenger use is hard to separate from the driver’s
Aliases: lockout collateral · occupant ambiguity · center-stack lockout
What it is
The center stack is a screen whoever can reach. Speed locks and gear locks treat it as the driver’s hand, so a front passenger typing a destination, answering a message, or changing a track in motion is locked as if they were driving. That is collateral from lockout: the rule punishes “the car is moving” and cannot see that the person touching the glass is not the person driving. This entry is only that injury. Who each display is for, and whether the passenger should have an unrestricted surface of their own, is the next layer.
Why it happens
Lock conditions are usually speed, gear, and park—vehicle state, not person identity. A hand reaching from the passenger seat still meets “in motion,” and the grey buttons do not care. Seat occupancy, belts, and cabin cameras can hint that a passenger is present; presence is not “this touch is theirs.” The driver can still tap while a passenger sits there. The opposite error exists too: the system thinks a passenger is using it, the driver has leaned over, and the lock opens. One screen, one lock, one speed condition cannot structurally tell two hands apart. When a passenger is injured by the lock, the task does not vanish. It returns to the driver or flees to a phone—collateral feeds the other two failures.
Studying it
Run the same lock with a passenger present, with the seat empty, and with the passenger explicitly operating the stack, and ask whether on/off matches who is touching the glass. Cabin video codes which side the reach came from. Occupancy, belt, and gaze estimators can be treated as classifiers; report a confusion matrix, not only “is there passenger detection.”
Independent variables: action from driver seat or passenger seat, passenger seated or not, lock keyed only to speed or not. Dependent variables: false lock (collateral), false unlock, rate at which driver actions are labelled passenger.
Lab “please have the passenger do it” is instructed; who touches the glass on the road is messier. An empty-seat baseline must be reported on its own, or “unlock when occupied” will be inflated by false triggers in an empty car.
Where it stops holding
On a single-occupant trip there is no collateral, and a speed lock keyed to the driver is the right default. Right-hand drive or a biased screen changes who reaches more easily; the direction of collateral follows cabin geometry. When the rear seat cannot reach the stack, stack collateral is mostly a front-passenger problem; the rear is another display. A child seat can occupy the sensor without being an operator; occupancy-as-unlock then false-opens.
Applying it
- Do not let speed stand in for “who is using the stack.” At least when a passenger is present and the reach comes from that side, do not grey out input tasks as a bundle.
- When the system cannot tell whose hand it is, prefer a short passenger confirm (“passenger is typing”) over a silent speed cut. The confirm itself must be short and must not become a driving task.
- Verify by having the passenger enter a destination in motion. A keyboard greyed by the speed lock is collateral. A lock that opens when the driver leans over is a false unlock. Measure both.