Today, the “arms race” of various car companies focuses on the release of L3 autonomous driving models, and the field of autonomous driving is surging for a while. Whether it is the increasing maturity of technical systems and solutions, the evolutionary breakthrough in the performance of sensors such as lidar, or the gradual completion of the upstream and downstream supply chains of the industry, they are telling people: we are getting closer and closer to true unmanned driving – L5 autonomous driving. closer.

The huge development potential of autonomous driving makes players on and off the track all eager to move. Not to mention the fields of vehicle manufacturing and related equipment applications that are highly related to autonomous driving, even the travel service market is eager to see the future of unmanned driving. Someone in the travel service industry said: “If the existing online car-hailing car is completely replaced, Robotaxi will grow into a huge track of 6 trillion US dollars.” Robotaxi refers to the self-driving car-hailing car.

Five years have passed, and the autonomous driving track is crowded with more and more players, some of whom have already run ahead, and some are trying to widen the gap.

But the meal needs to be eaten one bite at a time. Now that it has just entered the L3 mass production stage, the autonomous driving industry still needs to “walk a few steps” to achieve real unmanned driving. The existence of LiDAR allows car companies to travel more leisurely in the near or far journey of L3-L5.

At present, the development of self-driving cars has ushered in the best era, and it has received unprecedented optimism from the market. IHS predicts that by 2035, more than 50% of the vehicles on the road will achieve autonomous driving, and the total revenue of autonomous vehicles and related equipment applications will exceed 500 billion US dollars.

According to a report by the McKinsey Future Mobility Research Center, around autonomous driving, a lot of capital, expectations and publicity are emerging in the domestic market.

A large number of Chinese companies such as Beike Tianhui, SenseTime, Horizon, Jingchi, NavInfo, etc. are developing the core components of the autonomous driving technology architecture, including lidar, cameras, processors, software and maps/location-based services, etc. . Between 2012 and 2017, about $7 billion in venture capital was invested in Chinese self-driving technology companies, an amount comparable to the amount invested in U.S. startups during the same period.

The report also paints a picture of the future in which “autonomous driving will subvert the experience and use of cars”:

The popularity of unmanned driving will reduce the incidence of traffic accidents by more than 90%, and the driving environment will be safer and smoother. The unbundling between the vehicle and the driver enables people with disabilities and other people who cannot drive or are inconvenient to learn to drive to travel autonomously. In addition, not only will the concept of cars be rewritten with the popularization of driverless driving, but the travel service market will also undergo more earth-shaking changes.

In the entire transportation industry, the travel service market dominated by online car-hailing has never stopped exploring autonomous driving. Since 2015, Robotaxi (self-driving taxis) have been intensively tested on the road, and even e-commerce giants such as Amazon have entered this track. Someone in the travel service industry said: “If the existing online car-hailing car is completely replaced, Robotaxi will grow into a huge track of 6 trillion US dollars.”

From a numerical point of view, the expected market space for future travel services is enough to make players on the track go crazy. But for now, self-driving still needs many efforts to evolve into a real driverless car and become mature and popular, so that this $6 trillion “cake” can be seen and eaten.

There are two thresholds that must be crossed on the road to the realization of unmanned driving: cost and safety and reliability.

The high cost of autonomous driving systems has long been the biggest bottleneck affecting its development. DARPA began research on autonomous driving technology as early as 1984, but exploratory research and development regardless of cost is destined to fail to enter the commercial field. It is not until today, 37 years later, that L3 autonomous vehicles have reached mass production. As for the L5-level “true driverless car”, according to McKinsey’s survey: only 27% of the respondents believe that the cost problem will be solved before 2025, and mass production will be achieved.

Perceived performance, reliability and safety are another major bottleneck in the promotion of autonomous driving technology. Unmanned driving will completely replace manual intervention, which naturally places extremely high requirements on its reliability and safety. In order to achieve true unmanned driving, each component of the autonomous driving system must continue to iteratively evolve to achieve a performance that completely replaces humans and far exceeds that of humans.

In terms of hardware, among the various sensors required for autonomous driving systems, lidar has the best perception performance. By emitting and receiving high-frequency lasers, lidar can quickly obtain a large amount of position point information within the detection range, and conduct 3D modeling through the point cloud composed of position points to obtain information such as object distance, shape outline, and moving speed in the target area.

There are differences in the reflectivity of laser signals between different substances. Lidar can accurately distinguish obstacles with different materials within the detection range without requiring very complex support algorithms. Compared with the logic of visual perception to judge objects by color, the perception accuracy is greatly improved. .

Based on the above advantages, lidar is widely recognized by the industry as an essential core sensor for high-level autonomous driving. Before that, lidars have been widely used in addition to passenger cars, ranging from scenic shuttles, unmanned logistics vehicles, to sweeping robots, guidance and reception robots, and lasers can be seen on them. Radar works.

At present, LiDAR has become the standard configuration for most of the perception systems of various autonomous vehicles and driverless vehicles that have been mass-produced, will be mass-produced, are being tested, or are still in PPT. The lidar industry also started explosive growth at a time when autonomous driving was in the ascendant.

With the evolutionary breakthroughs in recent years, domestic and foreign lidar manufacturers have successively come up with LiDAR products that are expected to open the “door of unmanned driving”. For example, the image-level ultra-high-resolution solid-state lidar C-Fans-256 released by Beike Tianhui not long ago is the world’s first 256-line vehicle-grade solid-state lidar.

Lidar, known as the “eye of autonomous driving”, has also become a “Mr. Key” that can play an irreplaceable role in solving problems such as cost, performance, safety and reliability among various sensors required for autonomous driving systems. .

With a resolution of 0.1°x0.1°, C-Fans-256 can accurately capture pedestrians and vehicles at a distance of 200m, and can easily achieve large-scale long-distance regional monitoring. Compared with human eyesight, and even environmental monitoring sensors such as millimeter-wave radar, it is more “observing”. With its omnidirectional scanning field of view, the vehicle can keep abreast of the surrounding road conditions while driving. For difficult driving scenarios such as unprotected left turns, it can take care of vehicles or pedestrians on the right.

With the powerful ability to capture abnormal events given by the ultra-high frame rate of the action camera, C-Fans-256 can calmly deal with emergencies such as high-speed emergency braking, ghost probes, etc. which have extremely short disposal time, and even exceed the limit of human physiological function response. , at your fingertips.

The 8-bit grayscale resolution enables the C-Fans-256 to achieve clear identification of lane lines around the clock, whether in the middle of the night without street lights or in the smoggy dusk.

C-Fans-256 also achieves lower power consumption than household ceiling lamps, and has passed ISO16750 a total of 9 categories and 31 tests of high vehicle compliance. Can work stably and normally. In addition, C-Fans-32, which is the same series as C-Fans-256, is currently the only LiDAR product that has passed the “Intrinsically Safe” certification, not to mention conventional road driving, that is, in high-risk environments such as coal mines that are flammable and explosive , can also operate safely.

Due to the evolution of performance parameters, the market demand has increased, effectively sharing the research and development costs of precision optical components in LiDAR, and the large-scale quotations of LiDAR products on the market have dropped again and again. Successfully independently developed 24 types of chips in 5 categories, and achieved Beike Tianhui with completely independent and controllable core devices. Naturally, it also gave a surprise discount to the bulk purchase unit price of C-Fans-256, a vehicle-grade LiDAR product with excellent performance. , so that lidar is no longer an expensive “toy for a few people”.

From 2007 to 2017, the Chinese passenger vehicle market grew at an annual rate of 16%, and its share of the global passenger vehicle market also increased from 9% in 2007 to 30% in 2017. China has now become the world’s largest vehicle and travel service market, and is likely to become the world’s largest autonomous driving market in the future. Whether it is a car company or a travel service player, they want to eat in a broad future after the popularization of driverless driving. For more cakes, lidar products such as Beike Tianhui C-Fans-256 are indispensable.

Only such products with outstanding performance, conforming to vehicle regulations and no longer high cost are the key to realizing true unmanned driving. And the significance of the popularization of unmanned driving is far more than creating one or more trillion-level tracks, but also that it can liberate each of us from the position of “driver”, so that the car is no longer just a car, But a veritable living space.

In recent years, unmanned driving has always been the key research and development target of smart cars. Major auto brands have invested a lot of energy in this field, and related supply chain companies have risen to the challenge. Recently, the unmanned driving project has ushered in the assistance of the big players in the major technology fields. It is worth mentioning that Baidu Apollo has signed a cooperation agreement with Hesai Technology to develop hybrid solid-state lidars to help the fifth generation of fully driverless shared unmanned vehicles.

It is understood that this lidar will improve the accuracy of proximity measurement, optimize the actual effective perception distance, and will simultaneously support OTA to meet the needs of batch intelligent management and remote upgrade iteration, and achieve the dual goals of performance upgrade and cost reduction. At the same time, it is mainly used for unmanned commercial operation, supports multi-city usage scenarios, and large-scale normal operation, which will undoubtedly greatly promote the popularity of unmanned driving.

In addition to the performance improvement and cost reduction, the biggest highlight of this lidar is the guarantee of safety. It is said that the security system of this customized lidar is very complete, which can provide nearly 100 fault diagnosis types and radar surface contamination detection functions, and can accurately detect sewage, mud, serious dust, scratches, etc. in real time, aiming at “physical damage”. , the protection is full to ensure the normal operation of the radar. Not only that, it also pioneered an active network protection mechanism for lidar, which can effectively prevent cloud data from being tampered with and the internal storage, communication, and logic units of the radar from being counterfeited, and ensure the safety of operating vehicles.

With the advancement of technology, it is believed that there will be more and more such “black technologies”.