The government aims to have driverless, or autonomous, cars on the road by 2025. With that being just over a year away, we have already seen technological developments that automate some tasks, such as partial automation in emergency braking and lane assistance installed in vehicles since 2021.
However, concerns remain over moving from partial and conditional automation to high and full automation. While society is ready to accept Artificial Intelligence technology, and excitement is building around being free to enjoy a ride in your car without being controlled entirely, is the technology ready to remove the human factor altogether?
Reducing Human Error
According to The Royal Society for the Prevention of Accidents (RoSPA), 90% of road traffic incidents are caused by human error [i]. It is firmly believed that this can be reduced with the introduction of automation as high and full automation has been proven to prevent crashes, limit injuries and, more importantly, reduce risky behaviour on the road. When I was learning to drive, I remember my driving instructor saying, "You are in control of a metal killing machine; treat it with respect". That still rings true to me today. I assume public concern stems from this. Are we ready to relinquish all control, and how much will we require human intuition?
Drivers in driverless cars
For the government's objective to have 'driverless vehicles' on the road by 2025, it is widely believed we still have a long way to go to improve public trust. Especially as automation is expected in more than just cars - the consultation also includes vans, lorries and even coaches. The term 'driverless' is also deceiving. It has been acknowledged that the motor industry will never be ready to have an untrained driver in a vehicle by 2025. As such, sensibly, the industry has pulled together the use of more accurate terminology around automation.
If the self-driving revolution is to be achieved by 2025, what will be the driver's role, and what are the risks?
It is a given that drivers will be required to hold a valid driving license and complete instructor courses. Whether that will require an upskill in automated driving is anyone's guess. The government is still in consultation on its safety ambition, having allocated £100 million to develop new laws and standards [ii].
The risks?
Allocated Tasks: As we have seen with automatic braking and lane adjustment technology, there is a temptation to automate more manageable tasks such as this whilst leaving the more challenging tasks to humans.
Driver Disengagement: Lack of driving practice will lead to less skills and awareness in drivers; this could add threats if needed to intervene in a dangerous or emergency situation.
Cognition: Drivers could become bored, distracted or, concerningly, more tired if not engaged during the journey.
Driver alerts
Once again, we turn to technology and artificial intelligence as the most feasible option for retaining driver engagement. The American Automation Association performed a study in early 2022 to test various existing technologies to monitor drivers in current vehicles. The study found that camera-based technology offered better results than steering control systems [iii].
Cameras fitted within the cabin can identify the driver's attentiveness when the car is on autopilot, used in some models of Tesla vehicles, for example. If the driver is deemed less attentive, an audible alarm will alert the driver with sound and bring their attention back to the car. Whilst significantly more successful than steering control systems, the technology has limitations and, as such, may not be an immediate answer to fully automated vehicle safety. The technology relies on eye movement, head position and various characteristics associated with alertness. But what should the driver be completely unable to take control of the vehicle?
Advanced driver monitoring
To fully ensure safety concerns are met within automated vehicles, significant advancements in driver monitoring technology will need to be made. The camera, AI and machine learning technology will require greater sophistication and be able to distinguish more than driver alertness. At present, existing systems cannot ascertain driver health, for example. What if the driver, alone in the vehicle, had suffered a stroke or heart attack? How would the AI technology manage the situation?
The answer to many of these concerns relies on the collaborative use of health tech. At electronRx, we have developed technology that can monitor respiratory (lung), cardiovascular (heart), and electroencephalography/EEG – measuring the electrical activity in the brain (head), which can identify health concerns in the driver through camera technology. Once issues are identified, connected AI and machine learning technology can consider appropriate responsive actions. The vehicle could pull over for the safety of other road users. It could alert the emergency services of the person's situation. It may even use GPS to drive the driver to the nearest emergency department. The options are endless, with creative thinking and scientific engagement.
Driver health
One of the ongoing concerns is the health of the driver and therefore the safety of passengers. Health concerns can range from something as simple as being tired, and possibly falling asleep, to something much more serious as a heart attack or stroke. The car will need to have features that support the driver and passengers’ safety, by automating tasks that will react for the driver. If the technology understands the serious nature of a patient's health – i.e. understands from simple health readings that the driver has lost consciousness, or it can understand they have had a heart attack, it could pull over for the safety of those in the car. As the machine learning progresses, we could see the car calling the emergency services and providing a GPRS location. Or even driving the vehicle to the nearest emergency department or GP surgery. It all seems like a pipedream, but the technology is already in development, and it is not in the realm of possibility it will be achieved before many automated cars are on the road.
A safer automated future
Combining advancements in health tech innovations with existing alertness skills, we can confidently see a safer and more progressive future for vehicle automation. The existence of no driver may be a long way off, but the technology to prevent serious incidents and retain driver engagement seems much closer. The issue is whether society trusts technology enough as artificial intelligence, machine learning, and automation continue to evolve our way of life. I remain excited by the prospect of a driverless car than can mitigate for the poor health of the driver and the safety of passengers, do you?
[i] https://www.rospa.com/media/documents/road-safety/factsheets/autonomous-vehicles.pdf
[ii] https://www.gov.uk/government/news/self-driving-revolution-to-boost-economy-and-improve-road-safety
[iii] https://www.reuters.com/business/autos-transportation/some-cars-systems-designed-keep-drivers-eyes-road-fail-aaa-study-2022-02-01/#:~:text=The%20AAA%20study%20said%20monitoring,while%20vehicle%20software%20is%20engaged.
