The lion's share of the press and hype for self-driving cars goes to the artificial intelligence (AI) that decides how to drive the car. Dig under the surface though and there's a wide range of other technologies which are all making big steps forward to deliver self-driving cars.
One area that caught my eye recently is sensors. If AI is going to replace the human brain behind the wheel, then sensors will replace our eyes. You can always spot a picture of a self-driving car because they have massive sensors strapped all over them. The problem is these are power hungry, not very durable and, perhaps most importantly, ugly. Recent research from MIT about low cost LIDAR shows the potential for sensors to get dramatically cheaper and smaller - a key factor in making self-driving cars a reality.
Another important technology self-driving cars will require is top-notch simulation. AI driving models will need to be re-trained for new cars, sensors or conditions. It's prohibitively expensive to drive cars around in real life for millions of hours, and when the AI getting it wrong can result in serious damage it's a risky venture. High class simulation technology will provide safe environments for new AI programmes to learn and experiment in.
Whenever a new massive hype technology comes along it's always interesting to try and identify what other technologies it will depend on, and the opportunities for start-ups to benefit from the hype without competing with Google.
We believe that commercial lidar-on-a-chip solutions will be available in a few years. A low-cost, low-profile lidar system such as this has many applications in autonomous vehicles and robotics.