Tesla's Testing LiDAR Despite Elon Musk Calling LiDAR a ‘Crutch’
03-06-2021 | By Robin Mitchell
Recently, a consultant in the automotive industry spotted a Tesla with a LiDAR system and was later confirmed that Tesla is working with Luminar. So what has Elon Musk said about LiDAR, why will LiDAR be a critical technology, and how will solid-state LiDAR systems change electronics?
Elon Musk, Tesla, and LiDAR
Elon Musk, the CEO of Tesla, is often very vocal about technology, the economy, and humanity. Unfortunately, while he has been right in some areas such as his founding of SpaceX and reusable rocket technology, he has also been wrong many times. A good example is Hyperloop, a proposed technological marvel that will send people around countries at unreal speeds in a futuristic pod. However, the reality has been an embarrassing few student projects of wheeled carts going down rusty tubes at speeds no greater than the arrogant Range Rover drivers who insist on tailgating those doing the speed limit.
So when Elon Musk compared LiDAR technology in autonomous vehicles as a ‘crutch’, it immediately became clear that this statement would come back to haunt him. Elon believes that the future for autonomous vehicles is camera-only technology, and this makes sense when considering that humans can drive cars using only two eyes. If object recognition technology can be improved and reduce the time needed to recognize objects, cars could, in theory operate entirely on camera technology.
However, a self-driving car must not be as good as a person; it must be better. People make mistakes when the sun is in their eyes, there is heavy rain or fog, and black ice on roads makes it impossible to determine if a road is safe. LiDAR is a ranging technology that enables a vehicle to create a distance map of their surroundings, and LiDAR is not easily fooled. Of course, any autonomous vehicle should NOT be reliant on just one technology; they should contain an array of sensors that work together so that if anyone sensor fails, another can take its place.
Tesla Spotted Experimenting with LiDAR
Despite his reservations on LiDAR technology, an industrial consultant recently took a photo of a Tesla vehicle driving around with a LiDAR system on its roof. After further investigation, the car had a manufacturers plate (i.e. not a personal plate), which indicates that Tesla (the company) was testing the system out. After contacting Tesla, it was confirmed that Tesla is working with Luminar on a LiDAR system which may play an essential in future self-driving vehicles. This doesn’t mean that Tesla will use LiDAR, and that this experiment may just be verifying Tesla’s in-house technology. Still, the very act of using LiDAR as a “testing standard” demonstrates the power of LiDAR.
Of all sensor technologies, LiDAR is arguably the most powerful technology for self-driving systems for two reasons. The first is that LiDAR natively produces a distance map from a single laser source with no need for further processing (i.e. direct measurements from LiDAR are distances at position x, y). Secondly, data from LiDAR can easily create outlines of objects, and therefore provide a quicker and simpler method for object identification than objects in images. Thirdly, LiDAR can be combined with visual sensors to create a full 3D map of the surrounding environment using images instead of “heat-maps”.
However, any future self-driving vehicles will utilize various sensors all tasked with different jobs. For example, LiDAR will provide the workhorse for mapping, SONAR will provide accurate parking features, and RADAR will provide penetrative ranging in the most severe weather conditions. Future self-driving systems may even deploy Bluetooth or other short-range radio communications so that cars can detect the presence of other nearby cars and exchange information such as emergency braking and sudden acceleration.
How Solid-State LiDAR is the Future
While LiDAR technology is the most practical sensing technology, systems generally found on experimental cars do have their drawbacks. One of the biggest drawbacks of LiDAR is the large and bulky nature of its design. The need for a scanning laser usually sees the use of a spinning mirror that introduces size and weight constraints and potential mechanical failure.
This is where solid-state LiDAR comes in. Solid-state LiDAR systems are actively being developed, and these eliminate the challenges found with older LiDAR technology. For example, using a MEMS mirror or adjustable polarizer removes the need for motors and ruggedized enclosures while simultaneously removing the need for maintenance and the chance of mechanical failure. The reduction in size, and use of MEMS, also reduce the price of such systems thereby enabling them to be integrated into all kinds of devices including smartphones.
Future autonomous vehicles could utilise an array of LiDAR sensors around the vehicle installed on the bumpers, doors, and roof that are barely visible but provide 360° vision. Elon Musk may be right about using only cameras, but considering that vision alone can easily be hindered by bright lights, sudden changes in brightness, and bad weather, LiDAR will probably become the most important technology in autonomous vehicles of the future.
Read More
