Why do you always misunderstand driverless?

BGA double-headed needle 031-JJ-5.7L features a round head and has a total length of 5.7mm.

Factory direct sales - solid capacitor - surface-mounted aluminum electrolytic capacitors, ensuring first-class factory goods!

At the end of March this year, Wang Jin, Senior Vice President of Baidu and former General Manager of the Autonomous Driving Department, publicly announced his departure from Baidu to start his own company, Jingchi Technology, officially entering the driverless car startup space. Soon after, the company received a road test license from the California DMV (California Vehicle Authority). Let’s explore this topic with our car electronics editor.

Wang Jin’s exit was not entirely unexpected, as the past few years have offered a great opportunity for technical talents to launch their own ventures. In fact, back in 2016, Baidu’s youngest engineer and top 10 talent, Tiancheng, left Baidu to found Pony.ai, which also secured a California DMV road test license.

Additionally, Yu Kai and Huang Chang, who were deputies to Li Yanhong and key architects of Baidu’s deep learning lab, also left the company. They founded Horizon Robotics and later showcased an ADAS (Advanced Driver Assistance System) prototype capable of detecting lanes, vehicles, and pedestrians on real vehicles. Later, Wu Enda also departed and recently launched a new project called Deeplearning.ai, with speculation that it may aim to develop key AI infrastructure.

As mentioned above, several high-profile departures have led to the creation of new companies. A common characteristic is that they are all involved in the driverless car revolution.

Did You Misunderstand Driverless?

Driverless and autonomous driving are two completely different concepts. To be precise, autonomous driving should be referred to as “assisted driving,” while full automation can be called driverless. Unmanned vehicles use cloud data and advanced algorithms to act as the “brain” of the vehicle, replacing the driver. In a driverless world, people can switch freely between being passengers or drivers, deciding where to go and how fast to drive, with the car acting like an invisible taxi driver following commands. Google even believes that driverless cars might not need steering wheels at all, as these manual controls represent outdated technology.

On the other hand, an autonomous driving system can also take control of the vehicle, but the system will return control to the driver when necessary. If an accident occurs during autonomous driving, the law may not necessarily blame the vehicle. This is similar to the unresolved Tesla incident in the U.S.

Compared to each other, driverless technology is one level higher than autonomous driving, representing the ultimate goal of automotive development. For example, Google’s driverless car has no steering wheel or pedals—just a start and stop button. Once inside, you set your destination, and the car handles the rest. However, the driverless car Baidu tested on the road was still primarily an automated vehicle.

Currently, there are two main technical paths for autonomous driving: ADAS and artificial intelligence. One approach involves gradually automating through ADAS, while the other uses AI to achieve full autonomy. Both paths aim to reach full self-driving, or driverless technology.

A Revolution of “Everyone’s Participation”

The development of the automobile industry has brought many challenges, including traffic congestion, air pollution, road accidents caused by fatigue, and safety risks due to human error. With the growing number of private cars, utilization rates remain low, and complex road conditions combined with unstable driver states create serious safety hazards, such as frequent truck accidents on highways. In China, according to recent government directives, traffic safety has surpassed coal mine safety and become a top priority for national security.

This is not just a problem in China—it’s a global challenge. Automated driving seems to be the solution people are seeking (which is one of the reasons why so many companies are rushing into the field). As data accumulates and technology advances, all parties must move forward quickly. Under fierce competition, the trend toward driverless vehicles is unstoppable.

Seizing this opportunity, we’ve compiled a list of domestic and international startups involved in the driverless sector for your reference.

The image above shows only a portion of the startups entering the driverless market. From the perspective of various players in the automotive industry, everyone wants to capture a key position in the next generation of vehicles. Although many think of tech companies’ projects as related to “driverless cars,” traditional automakers are also making significant efforts in this area.

At this year’s CESAsia, Honda demonstrated a concept car called “HondaNeuV,” equipped with an “emotional engine” that can assess the driver’s mental state and assist in safe driving based on facial expressions and tone. It can also learn the driver’s lifestyle and preferences, offering personalized suggestions.

At CES 2017 in Las Vegas, Hyundai’s IONIQ Electric completed day and night autopilot tests and received a “Level 4” rating from the American Society of Automotive Engineers.

Traditional automakers focus more on performance in the driverless field, leveraging their long-term manufacturing expertise and design strengths. Additionally, they tend to be less radical, keeping elements like steering wheels and brakes to preserve the driving experience and user enjoyment.

A market research agency recently released a list of 18 global driverless car manufacturers in its report, ranking them. The top five were Ford, General Motors, Renault-Nissan, Daimler, and Volkswagen—five traditional automakers.

Domestic manufacturers are also actively working in this space. In June 2015, BAIC and Didi signed a strategic cooperation agreement to collaborate on driverless technology. The National Defense University of Science and Technology and FAW Group developed the Red Flag HQ3 driverless car in 2011, though no further updates have been released. BYD and Singapore Labs reached an agreement to jointly develop driverless vehicles, while Geely relies on Volvo’s safety technology for testing. GAC collaborates with the Chinese Academy of Sciences on driverless vehicle trials, and SAIC partners with AVIC for relevant tests.

Different from traditional car companies upgrading through intelligent assistant systems, internet companies aim to eliminate the steering wheel entirely using cloud computing, big data, and AI—a different technical path.

This choice reflects the internet industry's technological DNA. Moreover, the business logic behind it drives differences between the two approaches. While car companies launch self-driving cars focusing on active safety and rely on car sales for profit, internet companies view cars as a platform to enhance driving experience and unlock value-added opportunities once drivers are freed from manual control.

In the chip-making industry, Intel spent $15.3 billion to acquire Mobileye, a leading supplier of self-driving hardware. Other chipmakers like Qualcomm and NVIDIA are also heavily investing in autonomous driving.

After Intel acquired Mobileye, NVIDIA quickly partnered with Bosch to jointly develop an AI-based autopilot system and prepare a super chip, Xavier, designed to handle various autonomous driving tasks. The resulting system will be sold to all car manufacturers.

Previously, Qualcomm acquired NXP, a major player in automotive semiconductors, while Freescale also has strong capabilities in automotive chip manufacturing.

Google and Driverless Cars

To better differentiate levels of autonomous driving, SAE International introduced a six-level classification system in 2014. The U.S. National Highway Traffic Safety Administration (NHTSA) originally had its own system, but in September 2016, it adopted the SAE standard.

The SAE classification includes six levels: 0, 1, 2, 3, 4, and 5. Level 0 means no automation, while Level 5 represents full autonomy without any human intervention.

Why Do You Always Misunderstand Driverless?

Most cars today fall between Levels 0 and 1. Features like adaptive cruise control and driver-assist systems are typically classified as Level 1 or 2. Tesla’s Autopilot is at Level 2.

Level 3 is somewhat controversial. According to the definition, the driver must be ready to take over in case the system fails. This limits the driver’s ability to multitask, making Level 3 less practical in real-world applications despite its technical necessity.

Many companies are working on Level 4, including Drive.ai, Baidu Autopilot, Uber, Tesla, and BMW. Level 4 is considered to have clear commercial value.

Google, which popularized the idea of driverless cars, aims for Level 5 fully autonomous driving. Unlike most automakers and tech companies that follow a gradual approach, Google focuses directly on full automation. However, this long-term strategy has delayed commercialization, as the current market requires quicker results.

According to Momenta founder Cao Xudong, “Google’s approach is too long-term. Despite strong funding and R&D capabilities, it hasn’t commercialized yet because of the current market limitations. If there are issues, it’s often blamed on not solving core problems.”

Alphabet, Google’s parent company, is also focused on business interests. When it sold Boston Dynamics to SoftBank, it was speculated that the company might not produce marketable products for years. Similarly, for Google’s driverless technology, it’s unlikely to enter such a field unless it adapts to existing rules and evolves step by step.

Easier to Enter the Game

Many companies are drawn to the driverless car market. No one wants to miss out on such a huge opportunity. People believe that drones will eventually become part of everyday life.