Scientists develop a low-cost silicon chip to help cars see the world more clearly
The breakthrough could reduce the cost and size of Lidar systems for autonomous vehicles
Scientists in the UK have developed a new low cost Lidar prototype aimed at improving the performance of the 3D mapping technology while reducing the size and cost for widespread commercial use.
Researchers at the University of Southampton Optoelectronic Research Centre (ORC) and US-based 3D technology company PointCloud Inc collaborated on the new integrated system, which uses a silicon chip to create high-performance 3D images for use in autonomous navigation systems.
Used by self-driving cars to take 3D images of its surroundings, Lidar technology acts like the ‘eyes’ of the vehicle, sending out thousands of laser light pulses per second to scan and detect objects around it, and calculating how far away objects are and if it's moving.
Automakers like General Motors and Toyota already make use of Lidar in self-driving systems, with Apple announcing it will also use the technology in its Apple-branded car expected in 2024.
But other car manufacturers have criticized the cost and accuracy of Lidar, with Tesla CEO Elon Musk calling it a “fool’s errand” in 2019.
For context, the industry-leading system by lidar manufacturer Velodyne which sees 360-degrees with a 300-metre range, costs around US$75,000 a piece. Buying in bulk would likely reduce the costs for car makers but still places a prohibitive price for the widespread rollout of lidar in consumer vehicles.
However, competitors are attempting to drive down the price, with lidar manufacturer Luminar announcing its plan to develop production-ready lidar systems in 2019, costing as little as US$500.
While the new prototype claims to reduce the costs of the technology, further details as to how much are yet to be confirmed.
What we do know is the latest tests of the new silicon chip show an accuracy of 3.1-millimetres at a distance of 75-metres, which project scientists say is a higher accuracy than previously available.
"The silicon photonics system we have developed provides much higher accuracy at distance compared to other chip-based LIDAR systems to date, and most mechanical versions, showing that the much sought-after integrated system for LIDAR is viable,” said ORC Professor Graham Reed.
"LIDAR has been promising a lot but has not always delivered on its potential in recent years because, although experts have recognised that integrated versions can scale down costs, the necessary performance has not been there. Until now.”
The density of pixels currently in integrated systems has also been addressed, with the prototype increasing the density to 512 pixels using a large-scale 2D detector, compared to 20 pixels in previous systems.
The research teams are now working to further improve the pixel density and beam technology make the system better suited to real-world applications and improve performance.