Driverless is a cutting-edge technology, which will create a huge market and have a great impact on society. From the analysis of the specific attributes of driverless, we have reason to believe that the future driverless market will form a dominant pattern, just as Google dominates today’s search market.
If you want to win the driverless market, enterprises must have some key capabilities. Tesla may be the ultimate winner. Why? Let’s analyze it.
Redundant camera data
If you want to win the driverless war, the driverless cars developed by enterprises must know more about driving than humans. Having redundant camera data means that the camera covers 360 degrees and obtains sufficient data. Combining these two points, the driving ability of robot cars may surpass that of humans.
Many driverless car companies believe that only “redundant camera data” can not achieve real driverless, and lidar is the right choice. However, the research report points out that the current in-depth learning technology can reconstruct the data provided by lidar, that is, lidar is basically outdated, and it is not a good choice for unmanned driving. As long as the camera covers 360 degrees and has excellent neural network training, the car can surpass human driving ability.
What are the major auto manufacturers doing?
GM added cameras to Cadillac, pointing forward. Human vision can also point to the front and side, and even observe the rear with the rear view, so GM’s scheme can not achieve real driverless.
Google waymo’s cameras are all over the car, but the cost is too high. The camera cost of each waymo car is as high as $7500 and the whole vehicle cost is as high as $200000. If it’s just a demo car, $200000 is acceptable, but it can’t be sold on a large scale. Without scale, it is impossible to provide sufficient data for the machine learning system, and the ability of the system to explain the world and drive safely can not surpass human beings.
If you have 360 degree cameras and a large number of cars, you can obtain enough data to make machine learning technology play its potential. Tesla car lures users to pay, then unlocks driverless technology and obtains data. This model is much ahead of others. Other automakers only pay to hire test drivers and get very few data sets, so these automakers advocate lidar.
The maintenance cost of electric vehicles is very low
If you want to win the driverless war, the cost of cars must be low enough to have an advantage in the competition.
We can certainly turn a gasoline car into a driverless car, but soon it will lose its competitiveness because the cost per mile is too high. According to some data, the operating cost of a gasoline car per mile is $1, because its maintenance cost is higher and the cost of gasoline is not low. For safety reasons, consumers can’t refuel themselves. The operating cost of electric vehicles per mile is only $0.25, mainly because the maintenance cost is much lower, it can be charged by electromagnetic induction, and there is no need to pay others for charging. When driverless really comes, the price of electric driverless cars will be much lower than that of gasoline cars, and the gasoline car company will go bankrupt.
In terms of efficiency and cost, Tesla has great advantages, and it will still be the leader in the future. As of January 2021, no production car has a longer range (265 miles) than the 2012 Tesla Model s. The range of the new model s is more than 400 miles.
If Tesla only blindly increases the battery capacity to improve the mileage, it is of little significance, but Tesla also uses other methods, such as optimizing the manufacturing process and reducing resistance. The drag coefficient of model 3 is only 0.23, which is lower than that of Toyota Prius. In this way, Tesla can extend the range and reduce the cost of driverless cars.
It’s also important to note that Tesla cars can travel 1 million miles in their whole life, as can batteries. If waymo can only travel 200000 miles and Tesla can travel 1 million miles, and the cost is apportioned to each mile, Tesla will take a lot of lead. Tesla attaches great importance to battery life, but other carmakers don’t care so much.
Continuous progress is Tesla’s advantage, and it will certainly find other ways to save costs in the future. For example, Tesla will die cast 70 independent components into one, which can reduce costs.
Customized neural network chip
Waymo driverless car has picked up passengers in Arizona. The car is equipped with a computer server, which swallows a lot of electric energy. Because of this, waymo can only drive for 30 minutes at a time, and it can’t go anywhere.
The analysis points out that Tesla chip is six years ahead of Toyota and Volkswagen, and at least three years ahead of NVIDIA Orin chip. Tesla’s chip is designed by itself, which is a strong competitive advantage.
From the perspective of energy consumption, Tesla chip also has advantages over waymo. If the energy consumption of cars is too high, there is no way to set up an electric vehicle fleet, because driverless cars need to be charged for 1 hour after driving for 5 hours.
Now NVIDIA’s chip is Xavier. The energy consumption of the next generation Orin chip will be much lower, but it still can’t compare with Tesla chip. When Orin is launched, Tesla will use 5-nanometer chips, with much lower energy consumption than current chips.
A progressive culture
If the enterprise does not have an excellent culture, it will not be able to maintain its advantage in the unmanned war.
For example, if a company launches the industry’s first driverless car, but it is a gasoline car, the company is likely to be destroyed by electric cars. If a company takes the lead in launching electric driverless vehicles, but it can’t reduce costs, it may also be out.
In the electric vehicle war, only by implementing the paid subscription mode, reducing costs, improving electric vehicle energy efficiency and enhancing safety, can we maintain our advantages. In addition, enterprises must have a correct corporate culture.
Tesla has all these advantages. It uses innovative solutions to improve energy efficiency, such as absorbing heat from computer chips and then heating the interior of the car. Press 70 parts into 1. All these measures can reduce the cost per mile, and the ultimate goal is to reduce the price of driverless cars.