Due to regulatory problems and technical difficulties, the development of self-driving cars has slowed down in recent years.
A few days ago, according to foreign media reports, Samsung Electronics has stopped the research and development of self-driving.
The decision is not alone, as several global technology giants, including Apple, have abandoned research and development in the field of self-driving.
Specifically, the Samsung Institute of Advanced Technology (SAIT), which is responsible for the medium-and long-term technology development of Samsung Electronics, removed autopilot from its scheduled research project this month and shifted its focus on robotics and other technologies.
This adjustment reflects the current change in the priority of technological development.
In fact, many multinational companies have begun to slow their investment in self-driving, and some have even withdrawn from this field completely.
This is mainly because the commercialization of fully self-driving cars has taken much longer than expected, and a number of self-driving-related accidents have heightened concerns about the safety of the technology, leading to increased scrutiny of self-driving technology in the United States.
Earlier this year Apple also announced that it had suspended the development of the Apple Car and disbanded a special project team focused on self-driving electric vehicles.
GM also cut its investment budget for the Cruise autopilot program, while Ford abandoned the level 4 autopilot program altogether.
At the same time, Motional, Hyundai Motor Group’s joint venture in self-driving technology in the United States, also decided to postpone the commercialization of self-driving cooperation projects with ride-hailing companies such as Uber Technologies and Lyft, and laid off staff to cut costs in the restructuring process.
It is worth noting that Samsung Electronics has made remarkable progress in the field of self-driving cars, including the development of advanced semiconductors, monitors and sensors, and even last year its self-driving system successfully completed a 200-kilometer stretch of road without driver intervention.
However, considering the overall trend of the industry and the adjustment of internal R & D strategy, Samsung Electronics finally chose to stop self-driving research and development.
Since the mid-1990s, with the introduction of adaptive cruise control (ACC) technology, the embryonic form of auto-driving has emerged.
This technology can intelligently regulate the acceleration and deceleration of the vehicle according to the motion state of the vehicle ahead, thus realizing the automatic driving of the vehicle to a certain extent.
After entering the 21st century, the development of autopilot technology has advanced by leaps and bounds and has already surpassed the initial “level 1” autopilot stage.
It is exciting that Mercedes-Benz is expected to be the first to launch a batch of “level 3” self-driving cars by 2023, marking a new milestone in self-driving technology.
The new generation of self-driving cars are equipped with advanced optical sensors and radar systems, which give vehicles the ability to run independently within a certain period of time.
This means that under certain circumstances, the driver can take his eyes off the road and the vehicle can still drive safely.
The driver does not need to intervene again until the vehicle gives instructions to regain control of the vehicle.
We can foresee that higher-level self-driving cars will gradually become a reality, in which “level 4” self-driving cars can achieve completely independent driving under specific circumstances, while “level 5” self-driving cars are more advanced.
it can operate on its own anywhere without supervision.
However, despite the broad prospects for the development of self-driving technology, the popularization schedule of these high-tech cars has been delayed due to technical and regulatory difficulties.
According to GlobalData’s research, since most self-driving car prototypes are still in the experimental stage, it is estimated that global sales of “level 4” and “level 5” self-driving cars will be only 250000 by 2030.
But that number is expected to soar to 4 million over the next decade.
However several high-profile accidents one after another have raised concerns about the safety of self-driving cars.
If autopilot programs are to be promoted more widely in the coming decades, they must be fully tested in the right environment and promoted gradually and cautiously.
At present, the biggest challenges facing self-driving car development are the severe regulatory environment and the lack of public trust in the technology.
Part of the reason for this is that some glitches in autonomous systems have caused fatal accidents, which has undoubtedly heightened public concerns about autopilot technology.
Sammy Chan, car sales forecasting manager at GlobalData, said: “although we have accumulated millions of miles of self-driving experience, the pace of technological progress is still relatively slow.
And any setback can be very serious and may even undermine the enterprise’s pursuit of full autopilot.
” In order to prevent further accidents, we need to develop a sound regulatory framework to conduct autopilot tests under more types of driving conditions.
Take the “L3” Mercedes-Benz S-Class launched last year as an example, which is currently only allowed to use its self-driving function on certain roads, such as some highways in Germany.
Regulators around the world from the United States to the European Union to Singapore are now cautious.
Nevertheless, they have approved the testing of self-driving vehicles at specific locations.
This shows that the development of autopilot technology is still in steady progress, and it is expected to be more widely used around the world in the future.
Recent discussions among some stock investors abroad have shown that growth in the mobile industry has slowed as electric vehicles fall into a so-called “stalemate”, a period of stagnation before new products or services are accepted by the public.
However, experts suggest that it is time to pay attention to artificial intelligence and related products such as radar and camera sensors related to self-driving.
Musk’s bet that electric cars and self-driving stocks were popular with investors a few years ago.
However, demand for electric vehicles has failed to keep pace with supply, and several companies, including Fisker, Lucid Motors and Polestar, have cut their production guidance for 2023.
Traditional carmakers such as Ford and General Motors have also taken a cautious stance on their investment in electric vehicles, and even Tesla has been slow to build a new plant in Mexico.
As a result, autopilot has also received close attention.
Last year, Cruise, backed by General Motors, suspended operations after the crash.
Elon Musk, CEO of Tesla, also thought that itsMuch of the valuation is related to its self-driving business, and the company also faces a crash investigation.
But Musk has made big bets on self-driving cars such as plans to launch a robotaxi in August.
Previously, Tesla’s FSD only used cameras to work, which almost runs counter to the entire self-driving industry, which generally believes that lidar is crucial to any fully self-driving car.
So there is a big question: whether Tesla can build a safe self-driving car in the current way is deeply worrying given that Musk is betting the future of the entire company on the technology of unlocking.
But Musk seems to have realized that what he was doing was wrong because Tesla soon bought a $2 million lidar from Luminar.
This transformation may seem trivial, but it can save Tesla.
Why is Musk betting entirely on autopilot? Tesla has abandoned the much-anticipated Model 2 in favor of robotaxi on the same platform.
It has also stopped the development of integrated die-casting, which could make all Tesla models cheaper to manufacture.
Musk not long ago fired the battery development director, vehicle development director, operations director and other core talent, and even dismissed the team involved in these projects! This seems to be a cost-saving measure aimed at using as much money as possible to develop Tesla’s FSD AI into a legitimate and fully autonomous system.
In other words Musk is sacrificing all Tesla’s advantages in the field of electric cars to achieve the results of self-driving cars.
Someone said, “the autopilot system consists of hundreds of CPU and software modules, including sensors, actuators, complex algorithms, machine learning systems, radar, lidar, cameras, etc.
” Even if the accuracy of each subsystem or system is 99%, more than 100 of them will reduce the accuracy of the whole system to less than ….
37% “.
While the security of FSD based on vision scheme is greatly reduced, which is confirmed by the data recently released by Tesla.
This is why almost all other autopilot programs combine lidar with cameras, radar and ultrasonic sensors to provide enough redundancy for the system to manage complex conditions.
Now all eyes are on Tesla and investors are hesitant about the shift to focus the company’s entire future on self-driving technology.
Even the slightest sign of robotaxi danger can have a disastrous effect.
Musk can’t risk it at all, so the lidar order is for the first robotaxi, the first sign that FSD will adopt lidar technology in the next few years.
According to related information Tesla is expected to launch robotaxi based on Model 2 later this year.
It is reported that the last time he came to China, Musk even proposed to China to test the robotaxi on the road, which the Chinese side was very happy to accept.
So, in brief, Tesla was granted permission to test his fully autopilot skills in China last month.
Tesla’s rapid adoption of lidar makes sense, as it makes Musk’s big bet on autopilot more likely to persist.
However, it is far from certain that it will pay off.
We don’t even know whether current artificial intelligence technology can achieve the consistency and safety required for self-driving cars, because the returns on artificial intelligence training and development are beginning to decline.
There are other problems, such as how to convert these lidars to Tesla’s car equipped with FSD, Tesla promised to eventually achieve autopilot.
Tesla also doesn’t have a lot of real-world lidar driving data, and they need a lot of data to train FSD AI to use this new type of data.
Most importantly Tesla has not yet confirmed any lidar plans.
, return to the first electric network home page >.