Artificial intelligence: driving towards the future
Autonomous vehicles are gradually becoming a reality. Though they are not on the road yet, their high-tech features have rapidly evolved over time and are already installed in some vehicles. If the driver is distracted, the Artificial Intelligence (AI) system can bring the vehicle to an emergency stop
The astounding success of the “Knight Rider” series excited the imagination of those watching a self-driving car and there was a particular fascination with the voice of the car. The Eighties were also prolific in aesthetic excesses and futuristic imagery. However, few would say that after four decades we would be still discussing ideas such as AI applied to cars. It sounds like a Hollywood fiction movie, but it is only a matter of time until we see AI systems in all cars.
Several technologies have been tested by different companies. Not all of them belong to the automotive sector. Google, for instance, has built an autonomous vehicle whose computer system is AI based. Three years ago, in 2016, the prototype of the American conglomerate was the most advanced to date, sparking the interest of carmakers, manufacturers and technology companies.
Elon Musk’s Tesla is one of the leading enterprises in the sector. Tesla Autopilot is a vehicle with advanced driver assist features, such as lane detection, cruise control, self-parking, among many other technological innovations. Moreover, this is not the only example: Uber, as well as Volvo, mainly in the transportation sector, are two giants investing in technology in order to develop self-driving prototypes.
The investment of brands in R&D has been massive. From partnerships with different industries, the idea is to design prototypes that may be on our roads in a near future. How will that be possible? Through artificial intelligence. The rate of AI-based systems inside new vehicles is estimated to grow by 109% in 2025. In 2015, this rate had a much more modest estimate: 8%.
Does this mean we are close to see self-driving cars travelling on our roads? Pessimists say that it might happen by 2030; much more moderate projections point to a five-year timeframe. Tesla, General Motors, Ford, BMW, Toyota and Renault-Nissan are investing in robotic innovations that were mere onscreen fantasies not long ago. The real common denominator among them all is AI.
By gathering data from ultrasonic sensors, radars, cameras, and lidars (sensors that measure distance through light), information is centralised and processed faster than in the human brain. It is based on this information - which, once stored, allows the vehicle to create patterns and the ability to “learn” - that an autonomous vehicle makes “behind the wheel” decisions.
Up till now, the automotive sector has offered cars with level 3 automation, on a scale up to five. Zero is applied to a non-automated car, while level 5 applies to self-driving vehicles. Jaguar, for example, recently introduced a technology that works through thermal sensors: a steering wheel that helps the driver stay on the road using a quick warming system on the left or right side, according to road-traffic handedness standards.
In July, the manufacturer unveiled another one of its technologies that relies on artificial intelligence and is already implemented in the Land Rover: a driver’s mood can be detected and the vehicle can make adjustments according to various settings when in stressful situations. The method works in real time through a camcorder.
CO-DRIVER’S “BEST FRIEND”
A car can be called ‘smart’ when it is equipped with an average of 10 sensors, up to eight cameras and two radars. Some commercially available models use these sensors and artificial intelligence to protect the driver in the event of distraction, taking control when an imminent collision is detected: by emitting a visual and audible signal, and ultimately emergency braking.
Among level 2 models already on sale, some are capable of speeding and braking on their own, resetting routes and turning without the driver touching the steering wheel, while others can go further by detecting cars travelling around a curve, parked vehicles, pedestrians and bicycles.
The virtual assist may turn out to be the co-driver’s “best friend”, by detecting traffic lights, road signs, safety distances and adapting speed. Also making use of simple details, such as an audible and visual warning – displayed as an icon; a cup of coffee, for example - to advise drivers to take a break from driving. The question is how to share a cup of coffee with this new “best friend”...
