Cars That Think iconCars That Think

Meet Budii, Rinspeed's Cuddly, Collaborative Robocar Concept

Autonomous cars can conjure anxious thoughts of HAL, the runamok computer in “2001: A Space Odyssey” – all-seeing and all-knowing, but neither cuddly nor collaborative.

Rinspeed, Switzerland’s utopian automotive think tank, has a more playful, humanist ideal in mind: the Budii concept. This sexed-up, self-driving take on the electric BMW i3 wants to be your friend on wheels. The Budii is designed not only to drive itself but to learn about itself, its driver and its surroundings and change its behavior accordingly.

A periscope-like laser scanner rises 70 centimeters from the rooftop to map terrain and deliver a 3D view of road surfaces or obstacles, adjusting the adaptive air suspension and ride height to match. Body panels incorporate fiber optics and semi-transparent effects to help the car communicate with pedestrians and other vehicles.

The cabin is just as future-funky. Doors open and close electronically. Folding fan blinds offer privacy for your mobile office. Stowaway plexiglas tables provide work surfaces. Storage compartments feature inductive charging for electronic devices. An elegant new operating and display unit, developed with Harman, learns a driver’s daily habits and preferences and records surroundings and interactions with other cars. It thus eases the job of the human being behind the wheel.

But the coup de grace is the ultra-sensitive, seven-axis robotic arm devised by Germany’s Kuka: It smoothly offers the by-wire steering wheel to either the left- or right-hand occupant. When they’d rather work or send texts, the steering wheel discreetly tucks into a central position to free up space. The robot arm can also become an “attentive personal valet,” a tabletop or a whimsical watch-winder: A high-res cabin camera monitors a watch by Lucerne’s Carl F. Bucherer, with the robotic arm automatically winding the timepiece when it runs low.

“It’s the world’s most intelligent watch-winder, and also the most expensive,” says Rinspeed founder Frank Rinderknecht, with a laugh.

Rinderknecht believes that autonomous freeway driving will become reality by about 2020. Yet as cars steadily become self-reliant, Rinderknecht says, people will need to change as much as their cars, relearning everything they know about the man/machine interface. It’s about developing healthy trust in the machine, he says, but not blind trust.

Although people now tend to quickly excuse a fender bender—to err is human, after all—they may well be less forgiving when the offender is an autonomous car, he says. Accidents will surely be reduced, but some are inevitable, even with machines in charge.

“Where will you put your anger?” Rinderknecht asks. “Will you kick the car? And when your anger cools down, no one will say, well, bits and bytes don’t always correspond, and things happen. There will be a huge outcry over the first autonomous accident, because people will expect 100 percent reliability.” 

Today Big Brother Watches Truckers, Tomorrow He'll Watch You

Operators of trucking fleets have been clamping down ever harder on the oft-sung liberty of drivers, most notably with in-cab surveillance cameras. The camera lets remote observers watch for errors and, just important, makes the driver feel that he’s being watching, even when he isn’t.

Big Brother-like cameras make sense for the sake of safety and legal liability. Then again, it also makes sense to get up at the crack of dawn for an invigorating run.  You’d rather not do it, and if you did, you’d rather it were your idea and not somebody else’s.

The way it works was laid out this month at a meeting of the Association of Equipment Management Professionals in Orlando, according to a report in Equipment World. First, the trucking company sets out clear policies for compliance. Next it installs systems to record drivers’ behavior, particularly when onboard systems notice anything unusual, like sudden braking. Finally, human supervisors review the records the next day to see whether to apply the carrot or the stick.

Read More

Tesla Model S: Summer Software Update Will Enable Autonomous Driving

For the past several years, we’ve been frustrated by the fact that most of the technology needed for autonomous driving is already present in many production vehicles. If your car has active cruise control and lane detection, there's no technological reason why it can’t operate autonomously on a highway at least some, if not most of the time. There are some very good non-technical reasons; namely, there’s no regulatory or legal framework in place yet for consumer autonomous cars.

Just like us, Elon Musk and Tesla Motors are getting impatient, but unlike us, Tesla is in a position to actually do something about it: Musk has announced that by this summer, a software update will enable the Tesla Model S to drive itself autonomously, hands free, on highways and other major roads.

Read More

Self-Driving Car Aims for First Cross-U.S. Road Trip

Cross-country road trips could be so much more fun without arguments about whose turn it is to get behind the wheel. A self-driving car aims to test that idea by making the 5600-km drive from San Francisco to New York City with a few human passengers tagging along for the ride.

Read More

Mobileye and Valeo Join Forces to Perfect the Poor Man's Robocar

Jerusalem-based Mobileye has long produced a pretty-good robocar system based solely on an optical camera.  Call it the poor man's robocar.

Paris-based Valeo is known for its comparatively affordable lidar—compared, that is, to the US $75,000 Velodyne unit you see rotating, Eye-of-Sauron-style, atop the Google Car. That’ the rich man’s robocar. 

Now the two companies are joining forces to produce—what else? The middle-class man’s robocar. Such a really-quite-good approach could easily be upgraded bit by bit, an incremental strategy that many big automakers appear to be following. This collaboration just shows off the strategy in a relatively pure form.

First, the two companies will take various nonlidar-powered sensors and integrate them into active safety systems, which for instance sense the slowing of the car in front and apply the brakes to avoid a collision. But the interesting part incorporates lidar for true self-driving applications.

By fusing data from four or more cheap lidar units with the output of a windshield-mounted camera, the companies hope to combine the sharp definition and precise range-finding power of lidar with the color sensitivity and other pattern-recognition powers of cameras. No doubt other sensors—like the radar that’s the mainstay of the big auto makers—will eventually go into the mix as well.

Mobileye has had a stratospheric market performance, and even now, after sobriety has set in, its market capitalization stands at $9 billion. This valuation is based more on prognostications of growth than on current income—prognostications supported by collaborations with auto companies, including the likes of Tesla Motors. 

Mobileye’s founders had one key insight: that a single, or “monocular,” camera on the windshield could support a self-driving system, at least for much of the time. When the system can’t cope, it hands the wheel back to the driver—just like the fancier systems that incorporate radar and other sensors. Lidar defines the upper end of fancy.

Valeo, a French auto supplier, and its partners are developing a lidar unit to be priced at several hundred dollars apiece. Lidar offers much of what radar does but throws in extras, like the ability to discern the edges of things. Four units could cover a car’s immediate surroundings about as well as those roof-mounted monsters, provided that the their data streams can be fused properly. Fusion would be the job of Mobileye’s system-on-a-chip.

Critics will cavil at such half-measures, saying it is not done well. Admirers will note that it is amazing that it is done at all—and for a price that the average buyer can stomach.

Tires Could Help Power Your Car

Goodyear is developing a concept tire that could someday allow vehicles to run on electricity generated when the rubber meets the road.

In many hybrid and electric vehicles, some energy is already provided from the wheels via regenerative breaking, which temporarily turns the electric motor used to drive the car into a generator: the force required to turn the generator helps slow the car down, friction brakes do the rest.  

But Goodyear is interested in producing energy directly from the tires, harvesting their heat energy with its BHO3 concept tire. According to Goodyear, while the car is sitting idle, the BHO3 tires would start to warm as sunlight hits the tires and the ground beneath. (The tire will be given an ultra black coat for maximum absorption of light and heat.) The heat would be transferred to the car through thermoelectric materials just under the tire’s surface that can generate voltage as the material flexes in response to temperature changes.

Read More

For China's Alibaba, the Magic Word Is "Connected Car"

Alibaba—China’s biggest tech companysays it’s entering a joint venture with SAIC Motor, China’s largest auto company, to build a connected car. The car, which could hit the roads as early as next year, would communicate with other cars via the cloud.

The news comes just days after Baidu—China’s biggest search company—said it would build a self-driving car, and just weeks after a Chinese government official suggested that tech companies should foster innovation in the auto industry. Maybe it was more than just a suggestion.

In any case, the news represents the confluence of two trends: the tech invasion of automaking and the attempt of governments to be seen leading the charge.

The techie invasion began in 2010, when Google unveiled its famous Google car project. Now Apple may be in the hunt, according to Valley gossip fueled by citings of sensor-festooned Apple cars and reports that the company is poaching engineering talent from Tesla Motors and battery-maker A123. Wags have already speculated how the experience of driving a future Apple car might turn out.

Government positioning also began its stately minuet some years ago in the United States, when states like California, Nevada and Michigan began competing with one another to be the first to welcome robocars to the roads. Joining them in recent months Britain, Germany, and the Netherlands have all announced robocar strategies, sometimes using language that betrayed a certain fear of being left in Google’s dust. One wonders how they will react to Alibaba and Baidu.

Will Self-Driving Cars Crash the Insurance Industry?

Insurance companies have been active participants in studies aimed at figuring out how self-driving cars will work in the real world. It makes sense, considering that before robocars take to the world’s roads, insurers must firmly establish who will be held liable in the event of a crash.

The Guardian reports that insurance companies are looking at the prospect of autonomous vehicles with trepidation. Assessing liability isn’t the problem. Instead some insurers fear that taking wheels out of human hands could make cars too safe, if that’s possible, and cut into their business. 

According to Cincinnati Financial, in its recent annual report to the U.S. Securities and Exchange Commission “technology innovations such as driverless cars could decrease consumer demand for insurance products.” Two other insurers, Mercury General and the Travelers Companies, also mentioned the deleterious effect that self-driving cars could have on the insurance business.

Read More

China's Search Giant Baidu Plans To Build a Robocar

Baidu, the Chinese search giant, is working on a self-driving car and may roll it out as early as this year, says chief executive Robin Li.

He spoke before attending a government conference in Beijing, according to Bloomberg. Just last week, the minister of technology had publicly encouraged Chinese companies outside the auto industry to work with it to spur innovation.

Another huge tech company thus joins the robocar bandwagon, the one Google started in 2010. Auto companies jumped on soon afterwards; just last year Mercedes fielded the first production car that could take over most highway driving. Now it’s the tech companies’s turn to climb aboard.

The past month saw a spate of rumors that Apple was developing an autonomous electric car, or at least working on the finely detailed mapping services such a car would need. In any case, the company has been on a hiring binge, snapping up engineers with experience in automotive batteries and guidance systems.

Baidu is no stranger to mapping, having already offered it as part of its suite of search services. Back in September it said it was collaborating with BMW on maps suitable for self-driving cars. 

How much of this is real and how much is for show? Who knows. Robotic cars are the newest big thing, and governments and companies alike are rushing to get in front of the convoy and thus appear to be leading it.

Mitsubishi Quiets Car Noise With Machine Learning

Mitsubishi Electric is claiming a breakthrough with its development of noise suppression technology to aid hands-free phone calls in the car and elsewhere. The technology improves the quality of the communication by filtering out almost all of the unwanted ambient sound that enters a far-field microphone while speaking.

Noises removed include rapidly changing sounds—which were, until now, difficult to deal with—such as passing cars, windshield wipers, and turn signals.

“Previously, only stationary noises such as road noise or the sound of the air conditioner were really dealt with, because the noise mixed with the speech could be easily predicted from past observations when the driver was not talking,” says Jonathan Le Roux, a principal researcher at Mitsubishi Electric Research Labs in Cambridge, Mass. “It is much harder to reduce noise when its characteristics are largely unpredictable.”

To better distinguish human speech from other sounds, the researchers are developing speech-enhancement systems that learn to exploit spectral and dynamic characteristics of human speech such as pitch and timber.

These systems employ machine-learning methods based on deep neural networks. (Facebook’s AI chief, Yann Le Cun, explained deep neural networks for us here.) These are trained to distinguish and suppress the noise and retain the clean speech using massive amounts of noise -contaminated speech data. The systems have millions of parameters that are optimized during training in order to reduce the difference between the output of the system and the original clean speech.

In order to reconstruct the clean speech, the neural networks construct special time-varying filters on the fly and apply them to the contaminated speech.

“The frequency contents of the speech and the noise can be intricately intermingled, and change abruptly,” says Le Roux. “Transient noises may last only tens of milliseconds, while speech changes from one phoneme to another every 100 to 200 milliseconds. So to effectively remove the noise, the filter needs to have a fine frequency resolution and be updated very rapidly.”

In tests, Le Roux says they were able to cancel out 96 percent of the ambient noise compared to just 78 percent achieved by conventional methods.

This technology fundamentally differs in approach and aim from active noise-cancellation methods such as those in anti-noise headphones, which try to physically remove ambient noise in a user’s environment. Examples of these methods applied in the car are Bose’s engine-noise cancellation and Harman’s road noise suppression.

Mitsubishi’s goal is to eliminate the noise picked up by the microphone while the user is speaking during telephone calls. Although active noise-cancellation methods could indirectly help with this problem by reducing noise in the cabin, Mitsubishi says they can only suppress low-frequency noise.  

“We want to make the driver’s speech more clear and intelligible to the person on the other end of the call by cancelling as much noise as possible, not just low-frequency noise,” says Le Roux. “Our technology will also be useful for hands-free command and control situations, such as when using Apple’s Siri or Google’s Voice Search in smart phones, as well as in call centers that use speech recognition to handle common requests.”

Mitsubishi plans to launch the technology in 2018 in its line of automotive navigation and communication devices.

Advertisement

Cars That Think

IEEE Spectrum’s blog about the sensors, software, and systems that are making cars smarter, more entertaining, and ultimately, autonomous.
Contact us:  p.ross@ieee.org

Senior Editor
Philip E. Ross
New York City
Assistant Editor
Willie D. Jones
New York City
 
Senior Writer
Evan Ackerman
Berkeley, Calif.
Contributor
Lucas Laursen
Madrid
 

Newsletter Sign Up

Sign up for the Cars That Think newsletter and get biweekly updates, all delivered directly to your inbox.

Advertisement
Load More