There’s some real excitement around recent news about driverless cars. People are now imagining a future where they just sit back and relax as their car ferries them from one destination to the next. While it is plausible that this future is not far off, there are still some issues.
Among the most significant issues is the cost of the systems that pilot these vehicles. At the current time, the autonomous systems cost about $100,000 per vehicle, before you even pay for the standard parts that you find in any car.
As a result, you have dozens of startups, tech giants like Google, and companies in the traditional auto industry all working on ways to bring these costs down.
Where Attention is Being Directed
Much of the focus for these cost-cutting efforts is on the LIDAR systems that map the physical world around the vehicle. LIDAR systems come with an array of lasers. The system emits the beams and measures the amount of time it takes for them to return to the sensors. By combining the information from the array of lasers, the system can tell the vehicle where different objects are, what the object is, and the speed and direction of an object within the environment. Considering that a driverless car can pose an incredible risk in terms of auto accident injuries, collisions, or even deaths, this level of quality technology is essential.
A top of the line LIDAR system like Velodyne’s HDL-64E currently costs about $75,000. It has 64 laser channels and it can map 2.2 million points per second. This makes it one of the most reliable LIDAR systems available, but the high cost will prohibit the market. Realistically, it is not something that could be used to manufacture affordable, autonomous commercial vehicles.
As a lower cost competitor, you have the OS1 LIDAR system from Ouster. The San Francisco area startup announced the release of the OS1 in December 2017, and it offers near high performance for a significantly lower price. This system can map more than 1.3 million points per second, and it comes in at about $12,000.
It’s impressive that a new system can provide this level of performance for a price that is this low. However, $12,000 is still much higher than the price would feasibly need to be to make these cars available to the masses. To make driverless cars affordable for the average consumer, the price of LIDAR would have to fall well below a few thousand dollars and come in at just a few hundred dollars.
The Future of Driverless Tech
This may seem like a tall order, but hundreds of millions of dollars are being invested into LIDAR research, and experts are optimistic that the price can come down to affordable levels. In 2016 Velodyne announced an order from Ford, estimating a cost of $500 per unit at automotive mass production quantities. However, this unit is not the HDL-64E. Instead, Ford ordered the HDL-32E solid-state radar puck. This system operates on 32 channels and can map about 700,000 points per second.
If autonomous vehicles are to become as ubiquitous as many are predicting, the cost of high performing LIDAR is going to have to come down considerably. But there is reason to be optimistic. New systems are in development and the costs do seem to be trending downward as research continues. It is more a question of when the price will reach an acceptable level rather than one of whether or not LIDAR will ever be cheap enough to use in consumer automobiles.