On 17 November 1895, Thomas Edison (he of incandescent light bulb fame) stated in an interview with the New York World:
“Talking of the horseless carriage suggests to my mind that the horse is doomed …… Ten years from now you will be able to buy a horseless vehicle for what you would pay today for a wagon and pair of horses”.
We saw in the last post that just thirteen years after Edison’s interview (1908), the number of cars in New York exceeded the number of horses.
Looking at the national legislation of most developed countries in the world today, the prediction that “within the next decade the internal combustion engine is doomed” may well be one that we can feel confident will also prove to be true with most cars being battery powered. Looking further ahead, what might the future hold for personal, motorised transport?
Well, for the last decade or so, we have heard a lot about autonomous cars: cars capable of detecting and classifying objects around them thus enabling navigation to a destination with little or no human intervention. Satellite navigation receivers, radar transponders, scanning lidar (laser ranging), cameras operating in the visible part of the spectrum, digital map databases and inertial sensors along with significant computational capability to merge data from these sensors, form parts of most such autonomous systems. Some systems will undoubtedly use a level of system learning, frequently referred to as Artificial Intelligence (AI), to improve performance based on real-life, on-the-road experience.
The principles and limited applications of AI have been around since at least the 1980s. Neural networks were used for facial recognition having been ‘taught’ using images of the target face in profile, full-face, with a hat, without a hat, with a beard, etc. The limitations of the early work centred on the camera and processing hardware available at that time.
Today, when we think of autonomous driving we probably think of the Tesla and Elon Musk’s Cybercab or maybe we have read about Waymo, the system developed by Alphabet, the company that owns Google.
A novel alternative approach is being developed by a company called Wayve who have a special dispensation to test and develop its AI on UK roads. In many respects the Wayve system mimics the way that a human learns from observation of their surroundings. The company’s test vehicles are equipped with cameras and the continuous images from these cameras are fed to the AI processor. One consequence of this system is that the car is free to travel in unmapped areas and is less dependent on either a space-based or terrestrial infrastructure.
The Wayve system differs from that being developed and deployed in the USA by Google, Tesla, General Motors and others. It learns continuously from its surroundings, just as the human driver gains in experience. It is clearly a significant rival to existing players and has attracted investment from major global companies including Microsoft.
Whilst it could be the best part of a decade before autonomous cars have fully developed the nuanced intelligence of an experienced driver and issues of liability have been resolved, there seems little doubt that we are heading for a future in which we have no need for a driver. This could have vast implications for the pattern of car ownership, congestion and, perhaps most importantly, the number of casualties on our roads.
This post was inspired by an article on the Wayve development in the Sunday Times by Matt Rudd on 26 May 2024.