IGA, Japan — JTEKT Corp., the top steering supplier to Toyota Motor Corp., finds itself in a predicament shared by many top global suppliers. It is master of a critical component that won't be so critical when the industry moves to autonomous vehicles.
JTEKT says it is responding with a series of new products for that emerging era. They include steer-by-wire systems, sensors that detect whether someone is gripping the wheel, new technologies to smooth the transfer of steering from machine to human and another system that sends automatic road feedback into the self-steering calculus.
The new technologies could make it to market as early as 2020, the Japanese supplier says. But executives are mum on what automakers they may have lined up as their first customers.
JTEKT knows self-driving vehicles pose a challenge to its mainstay lineup. It is the world's No. 2 maker of the old-school hardware that conveys the driver's hand motions into the car's front wheel movements. Having no need for human input changes that equation dramatically.
"That future is far away, but it is coming. That is why we are responding with these products," says Haruhiko Segawa, JTEKT's executive managing officer for engineering. "We want to be in a position to start offering these kinds of technologies to customers around 2020."
The company's answer is an array of products tailored to a scenario in which a computer steers the car or a human and computer switch control back and forth.
The supplier also is working on a patchwork of computerized redundancy systems to work as backups in case one of the self-driving computers fails.
JTEKT, which is about 22 percent owned by Toyota Motor and is the world's biggest maker of electric power steering systems, demonstrated its latest autonomous driving developments last month at its proving ground in the mountains between Nagoya and Osaka.
The first prototype was a steer-by-wire system that JTEKT wants to start selling around 2020.
It breaks the physical connection between the driver's hands and the front wheels. Instead, wires carry signals to actuators that steer the tires right and left.
This technology is seen as key to autonomous driving for a couple of reasons.
First, it increases steering response time. It could also free up space in the instrument panel area because it doesn't require a bulky steering column.
Designers covet that space to build in all the creature comforts that fire the imagination about a self-driving future. Think loungelike cockpits with foldaway steering wheels where the driver can plow through email on the ride to work or sit back and watch a video.
A second focus will be on new functions that aid steering when no human is at the helm. JTEKT is working on three with an eye on starting production around 2020.
One is a grip detector that can tell whether a human is holding the wheel. Knowing whether a human is on standby or totally disengaged is key to deciding when to switch a car back and forth into self-driving mode. Some semiautonomous vehicles require at least a light grip, even in autopilot mode, as a safety precaution. In some cases, the auto-drive mode won't activate without it.
Switching between autonomous and human control is further complicated by the sudden drop in torque that typically accompanies the hand-over. JTEKT wants to smooth the transfer through software tweaks.
Finally, JTEKT wants to make computer navigation more precise by having input from the road contribute minor adjustments to the steering.
JTEKT uses the car's steering system to read the feedback of the road conditions. It then sends the findings back into the steering computer, which makes the subtle steering adjustments.
Backing up all these systems will be a network of computer fail-safe systems, JTEKT says. It expects to be selling complex circuitry that will allow self-driving cars to keep steering, even if one computer system fails and the driver is fast asleep. Its vision calls for backups of everything from the motor controlling the wheel angles to the battery providing the electric power.