I just came across a short article by Michael Wayland mlive.com that summarises the efforts of Google, Continental, Ford and GM in the field of producing a driverless car.

In short the article doesn’t offer much new, mostly each company emphasising the need for more testing of the technology. Ford do seem to be indicating they are aiming for 2017 for some degree of “autopilot” technology in their vehicles. Continental on the other hand seem to be hinting at a 2020 goal for having their sensor technology available.

On one hand I’m a little sceptical of dates and deadlines, on the other with so many competitors trying to commercialise driverless car technology the pressure really is on to be the first on the market.

J.D. Morgan has produced an interesting graphic of people’s willingness to buy driverless technology:

Taken from www.jdpower.com

The original press release can be found here. Good to see they used a large sample size of 17,400 car owners. In summary, the majority of people aren’t willing to buy autonomous driving technology (at the prices J.D. Power suggested) at this point in time. However, the press release makes an interesting point:

The study also finds a high degree of interest (41%) in fully autonomous driving among vehicle owners who expressed interest in the automatic parallel parking feature, further illustrating similar interest in both semi-autonomous and fully autonomous driving modes.

So there is a market out there for autonomous assistance technology, and perhaps people will be more likely to embrace it once current driverless technologies become widespread.

In terms of the demographic break down; to me it makes a lot of sense that young people are willing to adopt new technology faster, which is reflected in these results. I’m curious as to the reason why men would be willing to adopt driverless technologies across the board. It could be that men are more willing to take risks on new, unproven technology, on the other hand it could simply be because we love are willing to spend a lot of money on our cars.

All in all, great to see J.D. Power generating this data.

(HT: Gabriel for this story)

Some weekend reading for you:

A press release from Panasonic over at engadget.com has announced that Panasonic has developed new millimeter-wave radar technology that could be quite useful in driverless cars. (HT: Gabriel)

Panasonic’s new automotive radar technology has overcome the difficulties with conventional millimeter-wave radar technologies; the advanced radar technology is capable of detecting humans and cars simultaneously in spite of the fact that human body reflects extremely weak radar signals compared with car body…..
….unlike optical and infrared cameras and laser sensors, whose detection performance can be significantly affected by visibility conditions, this new radar technology will not be subject to such conditions as nighttime, rain, snow or dense fog.”

No idea if or when this will come to market or how much it will cost but we will keep an eye on it.

On another note, this is our 100th post! We hope you are enjoying the blog and finding it informative. Thanks to everyone who has got in touch privately with your kind comments, if you have any comments or suggestions feel free to drop us a line.

An article over at technologyreview.com is reporting that an MIT lab has built a driverless electric car that uses Xbox Kinect to sense pedestrians. (HT: Next Big Future, nice to see one of my favourite blogs finally putting up another driverless car post!)

The researchers have outfitted a prototype electric vehicle (it’s about the size of a desk) with lights that look like eyes and the sensors from an X-Box 360 Kinect. The lights swivel to look at you when the sensors detect you, and blue LEDs flash to indicate the car has seen you. Directional speakers swivel toward you, too, and the car tells you it’s safe to cross. The system can also flash bright white LEDs to get your attention.

Sonar sensors can detect if a pedestrian is too close to the side of the car. If they do, LEDs in the wheels to turn from green to orange and red—getting redder as you get closer—to warn you, and let you know the car knows you are there.

Sounds simple and effective.

Taken from www.technologyreview.com/blog/energy/27797/

Its almost certain that Google is going to commercialise its driverless car technology. In a talk to engineers at the SAE World Congress yesterday Anthony Levandowski – the guy running Google’s driverless car project – confirmed that they want to see their driverless cars on the road sooner rather than later (HT: David Shepardson @ detroitnews.com).

Levandowski also shed more light on Google’s strategy:

“We don’t want to make cars. That’s not our interest.”
“All options are open. From giving the technology away to licensing it to working with (suppliers), working with (automakers) building a car with them — everything is open — and we’re trying to figure out which paths make the most sense.”

Good to see that Google is not going down Tesla’s route of making cars from scratch. As a scientist (and I’m sure a few people from Tesla will attest to) I know first hand the difficulty of solving an engineering problem from scratch. If Google can licence the technology or collaborate with a car maker they will shave years of R&D off their project.

The question is, if they do side with one car company or another who will they go with?

• They could stick with big US car makers like Ford or GM, that will give Google a big boost in with policy makers Washington DC for being patriotic. Ford especially is gaining a reputation for driverless technology already in its vehicles so it will have a willing customer base. On the other hand Ford and GM are developing their own driverless tech and may think they can beat Google.
• They could go with a smaller car company in the US which will also give them political points in Washington. Also a giant like Google could push around a small adventurous company like Tesla more than they could GM or Ford and get a better deal. The downside is the smaller company will lack the market penetration of a big automaker.
•  They could good with arguably the world’s largest car maker: Toyota. Google is already using Toyota’s Prius as its test bed technology and the company is absolutely ruthless in improving its production line. With driverless car technology costing $70,000 a vehicle Toyota would seem a tempting partner to help drive down costs*.
• A Chinese car company – As in the above point, driving down costs will probably be a big part of Google’s business plan and this is something China is very good at. The downside could be concerns about quality and China’s reputation for reverse engineering and copying high tech products.

*Google of course has the financial muscle to buy up high tech parts manufacturers instead and drive down costs on that front also.

My feeling is that Toyota would be the likely choice (or broadly licencing the technology). My only hesitation in suggesting this is that Levandowski indicated that they want a partner that is excited about driverless cars. We know Ford, Continental, BMW, Volkswagen maybe even Nissan are excited, but we’re not sure about Toyota just at the moment.

While not technically a driverless car Motorauthority.com is reporting that Chinese car maker BYD has developed a remote controlled car.

The car is called the BYD F3 Plus and is on display at the Beijing auto show. It will be interesting to see if this gets to market and what the uptake will be.

Thanks to Gabriel for posting the link on our Facebook page.

We loved the GM EN-V. It was small but oh so purty.

Now for something even better looking – the GM EN-V 2.0! It’s a concept car, sure, which is basically code for “don’t get your hopes up” BUT we like to think of it as having great potential. No longer balancing on two wheels, it feels a bit more like a normal car. While it’s not driverless – yet – we believe this is a model of the future.

GM is focusing this vehicle on China, especially some interesting concept cities which are on the horizon. With 2.5 million sales last year in China they are putting their first foot forward.

Is this what Driverless Cars will look like? We hope so.

(Oh and @GM – start sending us media releases like the other companies do. Tx)

There’s a quick piece over at the Technology Review today talking about the question : will Autonomous Vehicles save fuel?

Of course they will:

Simulations suggest that only a few cars need to have the technology to make a significant impact. If 1 percent to 5 percent of cars can send real-time data about their speed to a central hub, traffic jams can be spotted within five minutes, and cars equipped with GPS systems can be offered alternate routes to avoid them.

Adaptive cruise control could also prevent many traffic jams from forming in the first place. Drivers tend to change speed on hills or when approaching tunnels, which can cause traffic to bunch up. They also brake too much in response to vehicles ahead of them. After a few drivers in a lane do this, traffic can come to a halt. Cars with adaptive cruise control keep a steady speed on hills and brake more gradually if cars slow down.

Thanks to Gabriel for sending the link over.

Here’s a thought exercise for today.

A big problem for driverless car engineers is not common easy-to-imagine circumstance such as people suddenly braking on highways, but the random situations we occasionally find ourselves which are intensely human or simply difficult to predict.

It’s the unusual situations where manual override is needed which are going to keep the elderly, the blind, children and the drunk away from Driverless Cars in the short term.

Some situations that are atypical may include:

1. Drive-Through Take-Out – many companies have different set-ups and to program all of them into a car would take a certain level of genius.
2. Buying a parking ticket for an impromptu parking lot (such as for a festival or sporting event) and then parking.
3. Road-works or a traffic accident creating a detour.
4. Police flagging down a car.

Here’s two ideas for over-arching solutions to solve some or all of these problems above:

1. Colored Lighting – Light configurations could be set up that Driverless Cars will recognize and use. These could be installed into McDonald’s at low cost or added into a policeman’s baton. While there’s potential for people to create their own lighting configurations and steer the cars all over the road (for a funny Youtube video, for example) this issue could potentially be overcome by using highly specific light frequencies which are banned from common sale. Alternatively, simple hand signals could be developed to be used in these ad-hoc situations.

2. Telemetry – Taco Bell could easily use a system like a next-generation Bluetooth or Wifi to communicate with cars that have come into their commercial spaces and then direct them through. This could also be employed in all predictable ad-hoc situations like road-works or sporting events but not so much in the case of unexpected events!

The big issue I foresee with either of the ideas above is that of consent. Would an external system need to ask for the driver’s consent in order to then direct its flow? You would say the answer to this question would have to be “yes.” Yet how reliable is a drunk person going to be in giving consent? How about a blind person – how would they know what is going on? Or even someone who has a mental disability?

Due to the issues above it could be a while before the full transition takes place.

One of the big obstacles facing Driverless Car technology has been the inability of GPS systems to provide location with pin-point accuracy.

This looks like it has been solved, by Broadcom. Here’s a link to the press release.

The chip accesses a multitude of data points including multiple satellite constellations and even in some cases, WIFI, to create pinpoint accuracy.

From Singularity Hub:

Disregarding Intel’s proclivity to dub their chips with ethereal names, the latest chip offering from Broadcom Corp. is simply called the BCM4752. Broadcom reports that the chip communicates with four different satellite constellations, which means access to the 59 satellites hovering over our heads, in addition to cellphone towers. It establishes indoor positioning by pulling from a bunch of different wireless sources, like Wi-Fi, near field communication (NFC), and inertial sensors plus its built to capitalize on emergent technologies, like Bluetooth beacons and 5G Wi-Fi access points (using the new 802.11ac standard even though everyone is still catching up with 802.11n). The more precise vertical and horizontal positioning is accomplished by communicating with a host of various sensors, like accelerometers, step counters, gyroscopes, altimeters, and magnetometers.