STATE OF SUPPLIERS: Mighty Aisin tackles lowly parking lot

KARIYA, Japan — The natural corollary to a self-driving car is that it ought to be self-parking, too.

And if you've ever tried to thread a car into one of cramped Japan's ultranarrow parking slots, you'll know that doing so is hard enough for a human, let alone a computer.

Now Aisin Seiki Co., the massive Toyota Group supplier of transmissions and powertrain components, is trying to nail that parking challenge as its contribution to an autonomous driving future. The supplier is developing two advanced self-parking systems it hopes to start selling in the 2020s. The first is a remote-control parking function that lets the driver park the car with no one behind the wheel. The other is a "valet parking" system that will allow the car to patrol a parking lot by itself until it finds an empty spot, then park itself automatically.

It is admittedly a niche-market move for a big supplier that is eager to refashion itself from an old-school metal-bender into a forward-looking software player. That is the same worrisome mission that confronts auto parts suppliers of all types all over the world. Automakers and their suppliers are rushing to develop ambitious autonomous driving technologies for highways and city streets. But Aisin sees a more immediate opportunity by focusing on the lowly parking lot.

"Recently, we've seen a big focus on autonomous driving, and as a company, we don't want to be left behind in this," said Hiroshi Ishiguro, manager of parking assist development at Aisin.

It doesn't help matters that Aisin's biggest customer, Toyota Motor Corp., envisions all but eliminating traditional gasoline engines from its portfolio by 2050, putting a crimp on Aisin's signature product: old-school gearboxes. Aisin has responded to the shifting ground with a multi-pronged diversification strategy that is focused on a host of new products and segments. They include hybrid drivetrains, inverters for electrified cars, technologies for connected vehicles and, yes, vehicle control systems for self-driving cars.

But to start in that self-driving realm, Aisin is focusing on parking partly because rival mega-suppliers such as Robert Bosch and Valeo SA already have a head start in high-speed autonomous driving.

"For automated highway driving, there are many companies working on development right now," Ishiguro said. "But the technologies needed for highway autonomous driving and low-speed autonomous driving are different. We don't think achieving highway driving means you can then easily develop a slow-speed system. So we want to apply ourselves to this area."

Rudimentary self-parking systems have existed for years. Aisin introduced its first in 2003, a system that used a camera to detect the white parking lines. A second-generation version arrived in 2015 using cameras and ultrasonic sensors to park, even when there are no lines.

Aisin's current system takes over steering, but it still needs a human foot to work the pedals and a hand to shift back and forth from drive to reverse as the car eases into place.

The coming systems, however, will be fully no-hands, no-feet parking.

"We are starting from steering-wheel assist," Ishiguro said. "We aim to provide automation for all areas of parking, including the brake pedal, accelerator, shifting and safety checks."

Aisin is basing the systems on technology it already sells to Toyota and Honda.

Toyota's system, deployed in two non-U.S. vans, the Alphard and Vellfire, uses a single rear camera and ultrasound sensors to park the car. The Honda version, used in the Japan-market Odyssey and Stepwgn van, uses four cameras for a 360-degree surround view.

Both systems were born out of the challenge of managing Japan's tight public parking places. In fact, at least for now, Aisin is not even targeting the U.S. market for the developing technology because it believes America's parking spaces are so roomy by comparison that U.S. drivers don't need the help.

Aisin does see demand for it in urban settings in every world market. But when it comes to ultracrowded Japan, where space is at a premium, tight parking is the everyday norm.

It is not uncommon for a home garage to have just enough room to allow passengers to open the doors on only one side of a car.

Vehicles with sliding doors are popular in Japan for that very reason. Hence the need for a car that can roll itself into the garage with no one on board. Aisin's idea: The passengers get out before it enters the garage. Then the driver parks it with the touch of a button on a smartphone.

While far advanced from what most vehicles owners around the world can do today, the technology is considered only Level 2 autonomy because a human is still in control and must confirm the car's safety. But Aisin aims to progress to Level 4 with its valet parking plan.

Because self-parking cars would not, in theory, need room for their doors to swing open, the valet system would allow vehicles to basically pack like sardines into a public garage or lot.

That would save valuable real estate in a prohibitive market such as Tokyo or London by allowing even more spaces to be squeezed into a designated lot.

It also could alleviate congestion and even long waits at parking lots.

"You know how crowded it can get on the weekends at a shopping center parking lot," Ishiguro said. "For the shopping center, it's a lost business opportunity."

At least in the early stages of the rollout, valet parking will require dedicated parking lots. Aisin now is searching for real estate and parking enterprise partners, but that raises a chicken-and-egg problem. Few people want to invest in valet-only parking lots when so few cars currently exist with the self-park technology. And few consumers will want the technology if their parking opportunities are limited.

Ishiguro concedes that it's an r&d conundrum. But the issue is getting some support from the Japanese government, he says, which is helping to coordinate testing and opening the discussion of community infrastructure needs.

But still another ingredient will be required.

Although Aisin's hardware setup of around-view cameras and ultrasound sensors is enough to pull off remote-control parking, when it comes to valet parking, cars also will need vehicle-to-infrastructure connectivity.

That is because several cars may be cruising a parking lot at the same time. They will need to communicate with one another through a central hub. The hub will have to monitor a parking lot's open spots and basically act as a traffic cop, plotting the most efficient path for each car.

But even if those market hurdles can be cleared, initial rollout of Aisin's self-parking technologies likely will be limited to hybrid and electric vehicles, not traditional gasoline-burning ones.

Why? Because a car with a mechanical link between its shifter and its transmission still will require a human hand to change from forward to reverse.

For engineers to fully automate the shifting, it would require a dedicated actuator, and that would add too much cost.

Hybrids and EVs, with their joysticklike knobs, are shift-by-wire and don't face such constraints, Ishiguro explains.

For that reason, Aisin is targeting its technology at electrified cars. But to hedge its bet, it also intends to develop a low-cost actuator that will enable automated shifting in traditional vehicles.

Cost reduction remains a key piece of the r&d challenge as giant Aisin charts its road into the future. What is technologically possible is one thing, Ishiguro points out. But, he adds, the supplier can't make customers feel they are being asked to bear undue costs for the privilege of not​ parking their car.​