Time to Ask: What Are Robo-cars For?

Time to Ask: What Are Robo-cars For?

VERSAILLES, France — With Automated Vehicles (AV) at the top of the hype cycle, not a week goes by without encountering news about these robo-cars. This week alone, two separate robo-taxi plans emerged independently from Waymo and Navya, followed by an inconveniently timed accident.

Navya’s self-driving shuttle got into a fender-bender on its first day of service in Las Vegas.

The media and public, including the engineering community, are too often breathless about the simple notion of a robot taking over the driver’s seat. Many engineers are in awe of the increased computational power in AI (artificial intelligence), and the potential of a connected world enhanced by the power of the cloud.

The popular conclusion, which has become boilerplate for AV marketing, is that highly automated vehicles will make our roads safer and help us — riders — work more efficiently. The prospect of fewer vehicles, also, will make cities less congested and cleaner.  

In reality, we have yet to see much evidence or data analysis to back up most of these utopian claims. The tech industry has been too busy promoting bigger, faster and smarter technologies, with the media as its cheerleading squad.

There is ample opinion about one technology being — arguably — superior to another. But rarely has any commentator examined the basic premise, or even asked the first question someone should have asked long ago: Why AV? What’s it for?  

20 years of AV development
EE Times this week met with Carlos Holguin and Michel Parent, the CEO and president, respectively, of AutoKab. The startup, focused on developing “safety-assured automation for commercial vehicle fleets,” is located here on the campus of the French Institute for Research in Computer Science and Automation (INRIA). INRIA’s headquarters is the historic site of the former central command of NATO’s military force.

Only two years old, AutoKab consists of a team of founders with more than 20 years of real-life experience in public transport planning, and road vehicle automation and operation.

Michel ParentMichel Parent


AutoKab’s garage-like R&D space is filled with older autonomous shuttles, golf carts and passenger vehicles designed and used over the last two decades. It has the look of an autonomous vehicle museum.

Michel Parent, who has spent half of his professional life doing research at such institutes as Stanford University, MIT and INRIA, and the other half in the robotics industry, has given much thought to “autonomy” in transportation. Parent said, “Sure, 20 years ago, I thought about designing a robot that can drive a car like a human, but then, I said to myself, what’s the fun in doing that?” Instead, he said, “We need to ask ourselves what problems we’re trying to solve.”

The problems with today’s cities are congestion (traffic jams), space constraints (too much space spared for parking cars), inconvenience (taking too long to connect to mass transit) and energy inefficiency. “We want connected, shared — public — and electric vehicles” that can run faster, said Parent. AutoKab’s CEO Holguin summed it up: “The last-mile problem is what we’re trying to solve.”

Carlos HolguinCarlos Holguin

Noting that unless AVs are shared, “as an automated bus-like or taxi-like service,” Holguin predicted, “autonomous cars don't reduce traffic congestion, [and] in the worst case they multiply it.” In his opinion, the goal of AutoKab is “fighting against the solo driver.”

Key personnel working at AutoKab come from the two different worlds of “technology” and “transportation.” Holguin said, “We are using technology to fill in the missing link for commercial operation of automated vehicles.” More specifically, Holguin claimed that his company is offering what he calls “the world’s first autonomy service by-the-mile” for public transport operators.

Automation Kits for Autos and Buses
AutoKab does both technology development focused on safety and data analytics to collaborate with cities for more efficient mass transportation.

The name AutoKab stands for “Automation Kits for Autos and Buses.” The company also develops hardware. AutoKab “automation kits” are designed to plug in to any vehicle. It consists of a variety of sensors — radars, lidars and vision systems — and computing processors (from Intel, Nvidia and others) for safety, explained Cristian Sandu, chief technology officer of AutoKab.

From left, Carlos Holguin, Cristian Sandu and Michel ParentFrom left, Carlos Holguin, Cristian Sandu and Michel Parent


Placing the kit on a vehicle properly requires some work, including software calibration, acknowledged Sandu. But it’s plug-and-play in the sense that the kit “can be installed and un-installed,” he explained, “connected to the automobile’s body via CAN bus or FlexRay, depending on a vehicle.” AutoKab embarked on the automation kit project not because it’s interested in selling hardware, but because such a kit is critical for anyone who wants to reduce the cost of AVs.

AutoKab, however, isn't interested in selling hardware, Holguin stressed. Rather, its focus is in working with cities so that they can devise a plan that allows transportation operators to run efficient last-mile services. AutoKab, in turn, gets paid “by the mile” AVs run in a city.

Next page: Extremely costly

 


PreviousChip Incubator Warms New Startups
Next    MEMS Sensor Startup mCube Buys Xsens