Volvo: AstaZero & Beyond: Putting 100 Autonomous Cars on the Road by 2017

.

AstaZero is located in Sanhult, Sweden, which is in the western part of the country.  And while its name may smack of cyberpunk or anime, it is actually the name of what is said to be the “world’s first full-scale proving ground for future traffic safety solutions.”

They take automotive safety rather seriously in Sweden. AstaZero is owned by SP Technical Research Institute of Sweden and Chalmers University of Technology (yes, this is a Swedish university); its partners are AB Volvo, Volvo Car Corporation, Autoliv, Scania, The Swedish Transport Administration, Region Västra Götaland, Borås Stad, Vinnova, Tillväxtverket, and Test Site Sweden.

AstaZero has a 3.5-mile highway on the periphery, a four-block city area with a Potemkin village setup, a high-speed loop, and other common traffic features.

The site opened in August 2014.

One of the participants in AstaZero is going to be, in effect, taking it to the streets.  As in the real streets of Sweden, some 43 miles west of the test complex.

Volvo Cars has announced that it is going to be putting 100 cars on roads around Gothenburg by 2017.  These are cars that are not going to be piloted by test engineers or safety experts.  The “Drive Me” program is intended, in the words of Dr. Peter Mertens, senior vp, R&D, Volvo Cars, “to enable ordinary people to sit behind the wheel in normal traffic on public roads,” with the wheel in question being that of an XC90 that is equipped with an array of sensors.

According to Erik Coelingh, technical specialist, Active Safety, Volvo, “Ideally, these 100 customers represent the customer base, i.e., younger and older people, experienced an unexperienced drivers, early adopters and people that are skeptical.”

Yes, this is a research project on autonomous driving with real people involved.   

The roads that are going to be part of the Drive Me program are public roads, but not all roads will allow autonomous driving.  For example, Jonas Nilsson, technical expert, Dependability & Verification team leader, Autonomous Drive, says that the cars will “only drive on roads without opposing traffic.”  He explains, “The reason for this is that we believe we can offer self-driving technology under these circumstances to real customers in the near future.”  Nilsson adds, “The next step is to look into more challenging environments.”

At program start, there will be 31 miles of road upon which the vehicles will travel. Volvo is having the roads that will be run in the Drive Me program mapped.  Nilsson explains, “The map will consist of all the features in the environment that the vehicle can see—for instance, road markings, traffic signs and road barriers.  When the car drives on the road, it can compare what it currently sees with the information in the map.  By aligning these two images, the car can estimate its position on the map with high accuracy.”

Additionally, Karl-Johan Runnberg, director, Government Affairs, says that the test fleet will be monitored (“The Drive Me vehicles will only use 3G or 4G communication, as this technology is widely available.  We do not want to depend on dedicated V2v/V2I [vehicle-to-vehicle/vehicle-to-infrastructure] communication as this is not yet available.  Similarly, we do not require infrastructure adaptations, as this would limit the further deployment of the technology.”—Coelingh) by a test center.  If the traffic control center determines, Runnberg says, “there are situations not suitable for autonomous driving, the AD mode will not be available.”

In other words, the person sitting behind the wheel will actually have to use it and the pedals.

The Drive Me cars are equipped with radar, cameras, laser scanners, and ultrasonic sensors.

In all there are seven radar systems, seven cameras, one laser scanner, 12 ultrasonic sensors, GPS, and a Cloud connection.  “Additionally,” Coelingh says, “there are many sensors in the vehicle such as accelerometers, gyros, and wheel-speed sensors.”

Henrik Lind, technical expert, Drive Me, Sensing System, Autonomous Drive, explains, “In the forward view, besides three radars [one in the rearview mirror mount and two in the front bumper] the Drive Me cars additionally have a trifocal camera and a laser scanner allowing detection of objects both on long and short range.  Supplementary are four surround-view cameras mounted on each side of the vehicle monitoring the close range around the car for pedestrians, lane markings, and vehicles.  For short range at low speed, we also use the 12 ultrasonic sensors.”

Speaking of the sensor array, Lind says, “It is a very interesting and demanding project where the sensors have the monitoring task of the road normally assigned to the driver.  Current sensors typically detect vehicles and pedestrians.  In Drive Me, the sensors additionally will detect any object on the road that may pose a hazard so the vehicle control can take the appropriate action in time.”  He adds, “The robustness requirements of the vehicle require a redundant sensing set.” Which explains the large number of sensors. What’s more, “There will be at least two redundant 12-V systems” in the Drive Me XC90s, according to Nilsson.

In the event of sensor failure, the system is setup so that the car will locate a safe place to stop.  Should the driver become incapacitated, again, the car will locate a safe place to park.  (How will they know of the driver’s status?  Nilsson: “We are designing a driver interface which requires multiple inputs from the driver during a certain time interval.  If the driver, for whatever reason, does not take over, the vehicle can always bring itself to a safe stop.”)

One question that arises for the Drive Me system is that of cost.  The Volvo team is sanguine about that.  For example, Coelingh points out that a fully equipped new XC90 has three radar systems and four camera systems, which are being used for Drive Me, along with the algorithms used for those systems.  So part of the costs are being borne right now.

What’s more, Lind says, “Volvo believes that the technology developed for the active-safety mass market will bring down the cost for individual sensors in the coming years.  Additionally, a number of sensors will be standard on vehicles due to NCAP tests.  Volvo therefore believes it is possible a few years from now to provide a high-automation package that is affordable in relation to the additional functionality provided.”

How much is Volvo investing in Drive Me?  According to Coelingh, “The overall investment in the Drive Me project is 500 million SEK.”  Or for those of you not spending krona, that’s some $59.7-million.

And when do they anticipate that there might be autonomous Volvos for the rest of us?  Responds Nilsson, “It is at this stage difficult to say an exact launch date for full production vehicles, but 2020 is a realistic target.”