Shift to driverless cars in the works at Simon Fraser University

Future vehicles will take over from distracted and drunk drivers, professor says

    1 of 2 2 of 2

      Since Karl Benz invented the modern car in the 1880s, humans have steered its descendants on the world’s roads.

      Ahmad Rad, a professor in Simon Fraser University’s School of Mechatronic Systems Engineering, told the Georgia Straight that times are changing. He’s researching driverless cars and advanced driver-assistance systems at SFU’s Surrey campus.

      “So maybe automobile is a misnomer,” Rad said during an interview at the Autonomous and Intelligent Systems Laboratory on the campus at Central City. “What we’re seeing now is a paradigm shift, in a sense that the car that we know is gradually morphing into an autonomous robot. The task of a driver will change from driving to supervising the car.”

      Rad and PhD student Mehran Shirazi showed the Straight the driving simulator in one corner of the lab. It consists of a car seat with a steering wheel, a dashboard, and—in place of a windshield—three flat-screens showing a city street. Two infrared cameras sit on the dash, facing the driver.

      The professor explained that the cameras constitute an eye tracker, which discerns where the driver is looking. This system can detect whether a driver is sending a text message, intoxicated, or falling asleep.

      Rad’s research team uses Fatal Vision goggles to simulate the impairment associated with particular blood-alcohol concentrations. In the future, he predicted, an automated car will temporarily take over from a distracted driver, and pull over and park or head home if someone is unfit to drive.

      “The human is driving the car,” Rad said. “But once it sees that the human is talking [on a cellphone] or sending texts, it will take over. Or, in worse cases, if the car recognizes that this person is drunk, it will take control of the car.”

      According to Rad and Shirazi, cars of the future will pair the eye tracker with a front-facing camera. This camera will be used to detect jaywalkers and road signs.

      “In many cases, drivers might not look at the signs,” Shirazi said. “The eye tracker will detect if the driver has seen the sign or not—the speed limit, any warnings. If not, the car itself would detect the sign and inform the driver that there was this sign that you missed.”

      Alternatively, Rad added, the car might override the driver in order to obey a stop sign or stay under the speed limit.

      The professor said that eye-tracking systems could be introduced in luxury cars in five to 10 years. However, he asserted that, once the software has been proven to be reliable, these systems should be made mandatory for every car on the street because they will reduce the number of collisions.

      “At some point, the seat belt was something which was designed for safety,” Rad said. “A lot of people didn’t wear seat belts. But now it is essential for every car—from expensive to cheap.”

      As for fully driverless cars, Rad predicted they won’t become widespread for another 20 years. He remarked that people shouldn’t underestimate the challenge of developing the artificial intelligence required for self-driving cars.

      “Driving implies a lot of things—a multilayered decision-making process,” Rad said. “And the decision-making part is the challenge. Just putting sensors in the car and getting information is one thing. But how to interpret the data into decisions—that part is really the key and creates challenges in designing a completely autonomous car.”

      SFU PhD student Mehran Shirazi plans to turn a toy car into a driverless vehicle.
      Stephen Hui

      In March, the B.C. Freedom of Information and Privacy Association released a report on the privacy implications of “connected cars”. Written by lawyer Philippa Lawson, The Connected Car: Who Is in the Driver’s Seat? includes vehicles equipped with driver-monitoring systems and autonomous cars integrated with intelligent transportation systems in this category. According to the report, intelligent transportation systems rely on vehicle-to-vehicle and vehicle-to-infrastructure communications.

      “As a key component of the expanding Internet of Things, Connected Cars are now feeding the Big Data machine and in particular, the increasingly personalized and targeted commercial marketing industry,” the report states. “The same technologies that create this data also create new security risks. Besides exposing vehicle operations to malicious hacking through new wireless entry points, they are creating huge new databases of personal data that is vulnerable not only to unauthorized access but also to unexpected use by governments, law enforcement agencies, insurers and others seeking to identify, monitor or take action against individual drivers or car owners.”

      At the Autonomous and Intelligent Systems Laboratory, Shirazi showed the Straight an Emotiv EPOC neuroheadset. The EEG device has 14 electrodes and records electrical activity in the brain.

      Using the headset, Shirazi said, they’re able to detect when a driver is thinking about braking, turning left, or turning right. This research could eventually allow someone to drive a car without touching the steering wheel or pedals.

      Noting that a neuroheadset is “very invasive”, Rad asserted that such a system would likely be used by people who have disabilities that prevent them from driving a standard car.