General Motors CEO Mary Barra announced this week that the automaker will launch the Super Cruise advanced driver assistance system and vehicle-to-vehicle communications within the next two years. Both technologies will find their way into 2017 model year vehicles, automating key aspects of driving for Cadillac owners.
Barra made the announcements on Sunday during her keynote address at the Intelligent Transport System (ITS) World Congress in Detroit.
"GM will put the first V2V-enabled car on the road in about two years, " Barra said. "What's more, I'm announcing that we will bring in advanced, highly automated driving technology to the market in this same timeframe."
The highly automated technology is the Super Cruise system that Cadillac began demonstrating and refining back in 2012. It plans to launch the system on an "all-new" 2017 Cadillac model.
Super Cruise falls short of the fully autonomous commuting imagined on concepts like the Zoox Boz and Rinspeed XchangE, but it's a class above today's automated cruise control systems.
"With Super Cruise, when there's a congestion alert, you can let the car take over and drive hands-free and feet-free through the worst stop and go traffic," Barra explained. "And if the mood strikes you on wide open roads, you can take a break from wheels and pedals and let the car do the work."
The hands-free bit is what advances Super Cruise past today's automated driving suites. Mercedes-Benz's Distronic Plus with Steering Assist is about as close as you can currently get to autonomous driving from the factory. While Mercedes marketing materials gush about that system's following, braking and steering capabilities, the small print way down below includes a disclaimer that the driver still has to remain attentive and continue operating controls such as steering and braking.
Super Cruise will allow the driver to hand physical control over to the car "in certain highway driving conditions," assuming GM's legal team doesn't slide in a Mercedes-like footnote before its market release. Super Cruise will then handle lane centering, braking and speed maintenance. GM is careful to point out that the feature is designed for attentive drivers. You won't be able to kick your feet up on the dashboard and take a nap just yet.
Barra stopped short of revealing the Super Cruise launch model, but she did say that it will slot into a segment that Cadillac isn't competing in today. That admission lines up quite nicely with rumors that Cadillac is developing a flagship to go head to head with the likes of the Mercedes S-Class and BMW 7 Series. That Cadillac model is expected to (fingers crossed) bear some resemblance to the stylish Elmiraj concept, which debuted in Pebble Beach last year. If such a model is in the works, it would seem like a logical launch vehicle for a new automated technology suite.
GM did provide a specific model for the launch of vehicle-to-vehicle communications technology, though, stating those capabilities will be included in the 2017 CTS. V2V communications will be a large part of future autonomous driving systems, allowing cars to communicate with each other toward increasing safety and traffic efficiency.
"V2V communication technology could mitigate many traffic collisions and improve traffic congestion by sending and receiving basic safety information such as location, speed and direction of travel between vehicles that are approaching each other," GM explains. "It will warn drivers and can supplement active safety features, such as forward collision warning, already available on many production cars."
GM spruced its narrative up by providing a visual timeline of its intelligent tech, starting with the 1956 Firebird II, its first imagining of an autonomous car. The gas turbine-powered concept car featured theoretical push-button autonomous driving capabilities designed for the "electronic safety highway." GM imagined the car being navigated down such highways by way of electronic signals.
You can see more of the Firebird II, along with other concepts and research vehicles that helped GM pave the way toward modern automation tech, in the accompanying photo gallery. The short Firebird II promo clip below is an interesting "future tech of the past" piece that's worth a look.
What do you think; were concept cars cooler back in the 50s?
Source: General Motors
If GM strikes out to finally take the plunge, someone will pay dearly.
For example, if the airbag system loses communication with sensors, it merely deactivates the airbags & notifies you on the dash. It fails "elegantly".
Another thing, what if a deer or large animal suddenly lunges at the car from out of the ditch. Will the car attempt an evasive maneuver if it can't stop in time? If so, could it be put into an uncontrolled skid, or direct you toward a head-on collision with oncoming traffic? Or on the flip-side, would it fail to take evasive action even if there was no oncoming traffic?
Lastly, would super cruise put you in danger by driving too fast for the conditions, i.e. snow, rain, fog? Or how about freak circumstances such as the road washes away or a bridge suddenly collapses, or a lane ends due to construction or somebody with a flat tire!
Because at its core what we are talking about is getting sensors and cameras down to a reasonable price and combining it with fistfulls of mobile computing so the car can use all the data to figure out what's happening around it.
Check out some on board footage from MobileEye to get an idea of what the system is doing: http://www.youtube.com/watch?v=H_wMyUEeIzQ
They are one of the vendors several auto companies have been working with.
.
For those that rely on GPS, until a way is found to cope with the jamming of their signals - and good luck with that - they have to be rejected if loss of the GPS signal could cause the vehicle to be involved in an accident. On reflection, I suppose detectable guide-wires laid beneath the road surface might provide some safeguard, but at what cost?
A more insidious GPS problem, is that of the signal being hijacked, allowing those vehicles within range to think that they are somewhere different to where they actually are. If the vehicles involved were tractors or grass mowers, then it would be a nuisance, but only tempers would be likely to be lost. However, if it involves an autonomously driven vehicle going at high speed along the left-hand lane of a single carriage way main road but thinking it is in the right-hand one and about to meet a 'normal' vehicle, also being driven at high speed, whose driver knows damn well that they really are on the left-hand lane and are about to lose their life, if they have the time to think anything, of course other than "WTF!" which I believe is traditional in such circumstances. I suppose if the GPS signals could be encrypted, there might be a solution, but if it involves new fleets of satellites I rather think that the expense might prove too much in these straitened times. And we must not forget that all existing sat-navs would need to be replaced and compensation paid, I assume.
Over and above the GPS problems, there is the little question of sensor reliability. Any vehicle guidance system that relies on analysis of camera images to avoid j-walkers or silly cyclists and the like, has the not insignificant problems of snow blocking the lens or rain distorting the image beyond practical application. Another problem is that of knowing when a sensor has failed, or has gone significantly out of calibration that it is useless. What do we do in these circumstances? Plough on regardless is not an option. Imagine whole cities brought to a stand-still because of the wrong type of rain or snow!
Finally, as far as this comment is concerned, I recommend watching some of the dash-cam video compilations of road accidents which are available on YouTube. It is surprising - to me, anyway - just how many instances there are of accidents that begin on the opposite carriageway/lane some considerable distance away. A human driver is capable of spotting the way things might develop and reduce speed accordingly, plus take any other evasive action that might be possible. I simply do not know how one could construct an autonomous driving system that could be expected to sensibly determine what to do in such circumstances without misreading other, quite harmless incidents and thus slow down or stop unexpectedly, fooling the following drivers, many of whom will not have autonomous vehicles - in the early days of adoption of the technology, at least. A 40 tonne truck can do a lot of damage in those circumstances, as many videos demonstrate, sadly.
We can safely leave the problem of coping with flooding for another occasion.