Autonomous Vehicles: Lidar and GPS Alone aren’t Robust Enough

Safe autonomous driving, much like humans and our five senses, will likely only be fully realized with multiple types of sensors.

Lidar has been the standard for most research and development with autonomous vehicles. Lidar is a technique that uses shoots lasers at objects and measures how long it takes for the light to return. Lidar is similar to radar, but instead of radio waves, uses laser lights. Lidar is good for identifying nearby objects.

GPS navigation is used for defining the route that the vehicle will take to get from the starting point to the destination. GPS provides coordinates of input and geographic terrain information.

First-generation autonomous vehicles using Lidar have been impressive, but there are a number of problems with it, that include cost, limited range, difficulty in resolving dynamics information of surrounding objects that are moving, interference with other Lidar signals when multiple autonomous vehicles are in close proximity, and the fact that lidar detection can be easily spoofed.

In the past, Elon Musk said “in my view, LIDAR is a crutch that will drive companies to a local maximum that they will find very hard to get out of. Perhaps I am wrong, and I will look like a fool. But I am quite certain that I am not.”
He later followed up and said that lidar is “a fool’s errand. anyone relying on LIDAR is doomed. Doomed. Expensive sensors that are unnecessary. It’s like having a whole bunch of expensive appendices… you’ll see… We believe that a vision-only system is ultimately all that is needed for full autonomy.”

Elon’s preferred solution for autonomous vehicles is to deploy an extensive number of low-cost cameras around the vehicle. Vehicles equipped with cameras and access to satellite imagery are also able to navigate without the need of GPS. But there can be problems there too.

Anthony Fragoso, research scientist at Caltech, said that “the rule of thumb is that both images—the one from the satellite and the one from the autonomous vehicle—have to have identical content for current techniques to work. In real systems, however, things change drastically based on season because the images no longer contain the same objects and cannot be directly compared.”

It’s likely that ultimately autonomous vehicles will need to combine multiple types of sensory data to be able to drive robustly.