Driving school future
AI and augmented reality could fundamentally change how tomorrow's learners will hit the road Image Credit: Shutterstock

In the future, technology is expected to make cars so smart that they’ll be able to drive on their own, while you relax in the back seat and watch your favourite TV shows. At the heart of these self-driving vehicles will be artificial intelligence (AI) capable of processing and reacting to data from on-board cameras and sensors faster and better than any human driver can. In fact, fully autonomous cars won’t have steering wheels, depending solely on their AI chops.

Though we are still some way off from this future, AI and other technologies are already impacting how we drive. Even how we learn to drive, as an increasing number of driving schools and instructors leverage them to turn you into a better driver.

Artificial intelligence (AI)

AI has emerged as the technology with the deepest impact on driving education. Walid Lebourre, Technical Director of Proven Consult, observes that as more industries implement AI, it makes immense sense that driving schools too take advantage of what it offers. AI-powered enhancements cover learning and driving assistance and environmental analysis. “The learning assistant lets the driving school focus personal tests on lesser known topics, increasing success rates for candidates and also providing feedback to improve learning materials.”

The learning assistant lets the driving school focus personal tests on lesser known topics, increasing success rates for candidates and also providing feedback to improve learning materials

- Walid Lebourre, Technical Director of Proven Consult

The driving assistant provides behavioural analysis by using cameras to monitor a student’s performance inside the vehicle, giving feedback on reducing stress or correcting posture. Similarly, eye tracking ensures a learner is making proper use of mirrors and dashboard indicators. Meanwhile, car behavioural analysis provides alerts about rough driving or excessive braking, and the gear assistant helps identify the right moment to change gears.

Finally, environmental analysis helps students with roadside signal identification, or alerts them when they are turning, or changing lanes, without using proper signals. At the end of each session, the data is crunched into an insightful driving report, showing instructor and student exactly where there is room for further improvement.

VR, AR and simulators

Imagine you are zipping by on a highway, right at the edge of the speed limit, while humming to a tune playing over the radio. Out of the blue, you see a truck hurtling towards you. You have mere seconds to react, and everything depends on what you do next. If you get it wrong, you may not have another chance — unless, of course, this scenario is unfolding in virtual reality (VR) or in a high-tech simulator that mimics a real car. Simulation allows students to keep replaying emergency situations, until optimal reactions get ingrained and their split-second decision-making skills are fine-tuned.

But even under normal circumstances, realistic simulations can help students become safer drivers. In 2015, Toyota launched the VR-powered TeenDrive365 simulator that throws various distractions at drivers — including phone call and text messages — and challenges them to stay focused on the road ahead. VR is also being used to train future instructors — a driving school in the UK claims its VR course delivers the results five times faster than regular courses for instructors.

AR HUD, robots and voice assistants

Augmented reality heads-up displays (AR HUD), when combined with AI, can enhance driving education by projecting important information right on the windshield — warnings, performance stats, hints or tips. Throw in wireless connectivity or 5G, and AR HUD can display real-time traffic data, along with navigation assistance. Infopulse, a consultancy, predicts that the global AR HUD will grow from $1.1 billion in 2019 (Dh4 billion) to $4.3 billion by 2025.

Or how about a friendly robot sitting right next to you that, unlike a short-tempered driving instructor, has nothing but sweet things to say about your driving skills? But if you are in China, you could try RoboCoach, an Alexa and Siri-style voice guidance system that use 40 sensors to monitor your driving performance and later have a word, or many, with you about it.

Besides, if you happen to ask this robo-instructor about your looks, it will tell you that you are the most pretty person it has ever met. That’s something you will never hear a human driving instructor say!

Autonomous cars: What’s taking them so long?

On March 2, news broke out that engineers at Google had created a robot that learnt how to walk. No big deal, right? Well, the headline-grabbing bit was that the robot had figured out how to walk on its own, without any human intervention.

This self-teaching ability, which goes way beyond doing what has been programmed, is a key requirement for vehicles to become truly autonomous. They will be regularly subjected to unpredictable, never-seen-before situations that will force them to literally think on their feet — or rather, wheels.

In their 2018 research paper, Teaching Autonomous Driving Using a Modular and Integrated Approach, the authors note that the main challenge of autonomous driving tech is making sure that vehicles travel safely in complex traffic environments. That is difficult, especially when the vehicle is required to use its full manoeuvring capabilities. “One approach is to search all possible paths and identify the best path. However, this requires enormous computational resources that may make it incapable of delivering real-time navigation plans.” One answer might be to offload the complex thinking to the cloud or remote server farms. But this will need high-speed 5G networks that are available everywhere and have zero downtime.

Overcoming various tech and legal issues is expected to take at least another 10-15 years. In short, driving schools and instructors have plenty to do till then.

— D.K.