Shaping the future of self-driving cars

Gorodenkoff/Shutterstock.com

For millions of people with disabilities, they promise newfound independence, removing barriers to mobility and making everyday transport easier. Then there is the environmental impact.

Are self-driving cars electric? Many are, and when combined with autonomous vehicle technology and ridesharing, they could lead to cleaner, greener and more efficient cities. By removing human error, autonomous vehicle technology has the potential to make roads much safer.

What is an autonomous vehicle and how do they work?

Using a mix of cameras, sensors, GPS and advanced AI, these vehicles can navigate streets, avoid obstacles, fully respect traffic laws, and even make split-second decisions, all without human input. According to ISO/SAE PAS 22736, there are six levels of driving automation, ranging from zero (no automation at all) to full autonomy where the car does everything, everywhere, on its own.

It all comes down to a powerful mix of sensors and smart AI working behind the scenes: radars, cameras and LiDAR. These high-tech sensors are constantly scanning the world around the vehicle, feeding a stream of real-time data to its onboard computer. Using advanced algorithms, this digital brain processes everything in milliseconds, analysing the environment, predicting movement and making decisions about steering, braking and acceleration.

AI is a key part of autonomous vehicle technology, but it does not “drive” the car or make spontaneous decisions on the road. Instead, AI is used to train and improve the systems that do. AI is used to process massive amounts of data during training, helping autonomous vehicle technology learn to recognise patterns – like pedestrians, traffic signs or lane markings.

International standards

There is no such thing yet as a true Level 5 autonomous vehicle. One that can handle any road, any weather, any situation without a human ever taking the wheel. To ensure self-driving car safety and long-term reliability, the industry needs clear standards, tougher regulations and real-world testing that leaves no blind spots.

That is where international standards come in. ISO 22737 is the first global benchmark designed specifically for autonomous vehicles, laying down the ground rules to keep development safe, consistent and on track.

Some of its key requirements include:

• Speed limits: Low-speed automated driving (LSAD) systems must be capped at 32 km/h (20 mph) to keep movements controlled and predictable, especially in pedestrian-heavy areas.
• Pedestrian and cyclist detection: Autonomous vehicles must be able to recognise people and cyclists, even when partially hidden, to reduce the risk of accidents in busy urban settings.
• Defined operating domains: Manufacturers must clearly outline the vehicle’s Operational Design Domain, specifying exactly where and when it can operate safely, whether it’s making left turns, handling nighttime driving or navigating in heavy rain.

Beyond ISO 22737, additional frameworks like ISO 34503 offer guidance on managing safety throughout the entire life cycle of autonomous vehicle technology, from development to deployment. Together, these standards lay the groundwork for a future where self-driving cars aren’t just efficient and innovative, but also reliable and worthy of public trust.

The future of self-driving cars is unfolding in real time. But there is still work to do. Stalled vehicles, broken traffic lights, unexpected construction, unpredictable human behaviour – these are the real-world variables AI continues to wrestle with.

Source: ISO.