A University of Michigan prototype uses inflatable zones to announce turns and lets drivers squeeze to push back, though the same study reports some drivers trusted the system too much.
A University of Michigan prototype turns the steering wheel into a two-way channel between the car and the driver. Inflatable zones at 10 and 2 o'clock expand to announce a planned turn in that direction. The driver squeezes to disagree. A pulse through the rim signals an obstacle ahead (University of Michigan release).
Most driver-assistance interfaces work in one direction. The car beeps, lights up, or vibrates, and the driver either notices or does not. This prototype asks the driver to push back with their hands on a car that is already driving. The wheel becomes a negotiation surface rather than a notification display.
The study, published in Human Factors (DOI 10.1177/00187208261483656), tested the wheel against two baselines: no information, and a one-way display that announced upcoming turns without letting drivers respond. Participants then encountered either agreeing or disagreeing automation, and in some cases a sudden obstacle. The team reports that the two-way wheel produced better task performance, lower self-reported workload, and faster recovery of trust after the automation made an error.
The same paper surfaces a complication the press release softens. Some drivers trusted the system too much. The release calls it "excessive trust" in a subset of participants. A faster trust recovery is not necessarily a better-calibrated one. A wheel that feels reassuring can also feel authoritative, and a driver who stops questioning a steering wheel has stopped doing the job the steering wheel exists to enable.
Toyota Research Institute funded the work, and a U.S. patent (US12227190B2) has been granted on the mechanism. The team is now searching for commercialization partners (miragenews summary).
The wheel itself is a research artifact, not a product. The simulation did not test on-road crash reduction, real driver-assistance integration, or production-grade sensor latency. Cost, manufacturing tolerance, and the basic question of whether drivers will actually squeeze a steering wheel to disagree with their car in traffic, in the seconds they have, remain open.
The researchers are taking meetings to find out whether suppliers will follow. A two-way steering wheel does not have to vanish when the car is driving. It has to do something only a wheel can do: register, in physical terms, that a person is still in the loop.