Lidar (Waymo) Vs Computer Learning (Tesla): Making the future of transportation safer
The Society of Automotive Engineers (SAE) has defined six levels of autonomous driving for cars. These levels range from no automation to full automation.
Tesla vehicles are currently in level 3, which is defined as a more advanced driver assistance system than on most automobiles today. Tesla offers its customers two types of automation packages- Autopilot and Full Self-driving. Presently, the autopilot comes standard in all new Tesla vehicles. The autopilot work in conjunction with the adaptive cruise control and lane-centering system. Moreover, Tesla customers have the option of installing a Full Self Driving system (FSD) for an additional cost of $5000. This system includes features like “Smart Summon” and “Navigate on Autopilot,” a system that can navigate the car on highways while being able to make interchanges and lane changes.
Waymo, on the other hand, is on level 4 on the SAE’s automation scale for its operational fleet in Phoenix, Arizona. Level 4 automation does not require anybody to be sitting behind the wheel while the vehicle is in service. However, It is to be pointed out that such a fleet is currently operational only in Phoenix. Waymo also makes use of a completely different technology called Lidar (Light detection and ranging), a remote sensing method commonly used for scientific explorations such as space missions and to examine the surface of the earth. In autonomous cars, Lidar is installed on the top of the vehicle; high-speed laser beams are emitted in 360 degrees to collect information while it is converted into a 3D map for the vehicle. Lidar is considered to be a more accurate system than the computer vision technology that Tesla and other manufacturers are currently using. It is also a much more expensive technology. Waymo recently hit 10 million autonomous miles driven by its fleet of 600 vehicles.
Computer vision, the technology used by Tesla for their autonomous vehicles, is a technology that comes at a fraction of the cost of a Lidar system. This technology uses a form of machine learning, which keeps improving over time. Tesla has collected data for over a billion autonomous miles driven by all its more than half a million vehicles, on all types of road conditions, compared to Waymo, which has data only for the city of Phoenix. Tesla has its reasons for not implementing Lidar on their vehicles, one of the main factors being that the system is too expensive and adds an additional cost of $10,000 to each vehicle. Lidar also has a tendency to breakdown faster due to the constantly moving parts. It is highly possible that stereo cameras can create maps nearly as accurate as a Lidar system in the near future.
Currently, insurance costs for autonomous cars are higher than conventional cars, mainly due to the high cost of replacing their parts. Once these cars are mass-produced on a larger scale, it is expected that their parts will become much cheaper, translating to lower insurance costs. Advancements in the two technologies would also make these vehicles much safer than conventional cars on the road, which should also be a factor in lowering their insurance rates. It remains to be seen which of these technologies would prevail in the near future.