Bosch’s Achievement in Mass Production of Advanced Autonomous Driving

The automotive industry has undergone rapid changes in the past decade, with consumers increasingly accepting smart driving as one of the crucial skills valued by automakers.

However, for most automakers, mechanical manufacturing is their strength, while they are weaker in terms of intelligence. Currently, automakers are gradually strengthening their layout in this area, but due to varying levels of investment and time in the market, their product development progress does not meet their development requirements well.

As an industry with a very long supply chain, there are suppliers in every niche market. Previously, the smart driving configurations of automakers were provided entirely by international Tier 1 suppliers like Bosch.

In the era of L2, Bosch became the industry leader with its leading product strength, customer-centric service attitude, and over 300 mass production projects. However, in the L2++ era, with domestic suppliers rising strongly, giants like Bosch seem to be left behind.

This is merely a surface phenomenon; as an international Tier 1 giant, Bosch is very sensitive to changes in the automotive market and has a clear understanding of the development of smart driving, already making corresponding layouts.

Recently, Bosch collaborated with WeRide to develop advanced smart driving products, which have already been implemented in Chery’s first smart car, the Xing Jiyuan ES, and currently can achieve high-speed NOA functionality.

This makes Bosch the first player among international Tier 1 companies to mass-produce advanced smart driving and push high-speed NOA functionality.

So, how does the advanced smart driving function developed by Bosch perform? How did Bosch launch this product? What is Bosch’s perspective on the next phase of smart driving?

With these questions in mind, Che Dongxi recently experienced Bosch’s high-speed NOA function in Shanghai and had a conversation with Bosch’s Advanced Smart Driving Project Director, Shou Guangyuan, and Bosch’s Advanced Smart Driving R&D Director, Zhou Jian, to find the answers.

Recently, Chery’s new smart electric vehicle, the Xing Jiyuan ES, officially launched the high-speed NOA function after an OTA update, with Bosch as the global Tier 1 technology provider.

As Bosch’s first foray into the field of advanced smart driving, how does the high-speed NOA experience of the Chery Xing Jiyuan ES perform?

Recently, Che Dongxi also experienced the high-speed NOA function of the Chery Xing Jiyuan ES in Shanghai. According to on-site information, this solution utilizes one forward-facing LiDAR and adopts two NVIDIA Orin X chips for computing, while also requiring high-definition maps.

From a hardware configuration perspective, this solution is already among the top configurations in the industry, and Bosch did not disappoint in the real-world experience.

Using NOA on the highway, there are several scenarios that cannot be overlooked.

The first scenario is automatic lane changing. Before experiencing Bosch’s solution, Che Dongxi had also tested solutions from several other providers, and frankly, most players did not perform well in lane changing; some players’ solutions seemed hesitant during lane changes, which lowered user confidence.

However, Bosch’s solution performs exceptionally well in this regard, with lane changes occurring almost seamlessly. Once the vehicle confirms that there are no cars in the side front and that vehicles in the side rear are at a safe distance, it will quickly change lanes, making the entire process very close to experienced drivers’ operations. A media colleague accompanying Che Dongxi gave a high evaluation of this experience.

The second scenario is dealing with sudden lane cuts. While driving on the highway, one might encounter other vehicles’ uncivil driving behaviors, such as sudden lane cuts, which are unpredictable and can be challenging for many L2 systems.

Bosch’s solution, due to its use of LiDAR, has faster and more accurate environmental perception, allowing it to quickly judge the intentions of adjacent lanes and decelerate to avoid collisions.

Notably, Bosch’s solution performs exceptionally well in decelerating to avoid collisions, employing a linear deceleration method rather than abrupt braking like some players. During Che Dongxi’s experience, the vehicle’s recognition of other vehicles, deceleration to avoid, and subsequent acceleration were all very natural, without any mechanical feel.

The third scenario involves temporarily closed lanes, which occur quite frequently on the road, whether due to lane repairs, temporary occupation for road maintenance, or road closures due to traffic accidents.

In the early days of smart driving, the recognition rate for this scenario was very low, leading to several accidents across various models. Bosch’s solution, while primarily guided by high-definition maps, employs more sensors for precise road perception, allowing it to accurately recognize such scenarios and change lanes in advance to avoid obstacles.

Additionally, there are several other representative scenarios.

With high-speed NOA activated, the vehicle can automatically enter and exit ramps, which presents two challenges: first, most ramps have large curvature bends, which can affect the vehicle’s centering; second, the vehicle must find the right moment to merge into the main road amidst heavy traffic.

If these issues are not resolved well, the vehicle may repeatedly adjust its position in the bend, reducing the experience; and when merging into traffic, it may hesitate like a novice driver.

However, based on Che Dongxi’s experience, Bosch’s NOA solution consistently performed well without any issues in these scenarios.

Moreover, Bosch has also incorporated some user-friendly designs into this solution; for example, when passing large vehicles, the car will proactively make slight evasive maneuvers to reduce the driver’s psychological pressure.

After the experience, this solution left a deep impression on Che Dongxi. From the riding experience, this solution is very close to real human driving, providing strong confidence whether as a passenger or while driving, with excellent handling of various driving details.

Clearly, Bosch’s first submission in the field of advanced smart driving deserves a high score.

After experiencing Bosch’s high-speed NOA, Che Dongxi was left wondering why Bosch, as a new player in the field of advanced smart driving, could deliver such an excellent response.

In discussions with Bosch’s Advanced Smart Driving Project Director, Shou Guangyuan, and Bosch’s Advanced Smart Driving R&D Director, Zhou Jian, this doubt gradually dissipated.

First, it is important to clarify that although this solution officially received an OTA update this year, Bosch’s considerations for advanced smart driving began much earlier. In 2021, Bosch established the Intelligent Driving and Control Division, also known as the XC Division, focusing on the research and development of intelligent driving, smart cockpits, and intelligent connectivity, initiating the development of advanced smart driving.

Simultaneously, Bosch’s collaboration with Chery was almost a perfect match. Bosch needed to clarify the automaker’s requirements to define platform solutions, while Chery, as an outstanding OEM with international success, required a more global Tier 1 to provide technical support.

The communication between Bosch and Chery began around 2021, and during the planning phase of the advanced smart driving platform, discussions with Chery started, culminating in confirmation in July 2022. Throughout the R&D process, Chery was deeply involved in platform planning and provided technical route suggestions.

This is also why Bosch’s solution was the first to be implemented in Chery’s model.

Throughout the development process, WeRide was also a significant participant. On May 25, 2022, Bosch strategically invested in the domestic autonomous driving startup WeRide, and both parties collaborated on the development of intelligent driving systems.

In the collaboration between Bosch and WeRide, Bosch primarily provided a full-stack advanced smart driving solution, including sensors, computing platforms, algorithm applications, and cloud services, as a Tier 1 supplier to the OEMs.

WeRide’s strengths lie in the algorithm layer, and Bosch and WeRide primarily collaborate on algorithm applications and toolchains, such as perception, prediction, planning for driving parts, and some toolchains.

Through this collaborative development approach, Bosch and WeRide completed platform construction and initial mass production within 18 months.

However, completing initial mass production is just the first step; continuous product iteration is also key. Currently, Bosch has established data closed-loop capabilities and developed a “shadow mode” similar to Tesla’s on mass-produced products, enabling the collection of sufficient data. Bosch has also established a complete data compliance mechanism through its proprietary cloud.

Thanks to its data closed-loop capability, Bosch can achieve rapid product iteration. From a functional perspective, Bosch already possesses L3-level autonomous driving capabilities.

By collaborating with OEMs and domestic autonomous driving companies, Bosch has taken a crucial step from 0 to 1 in advanced smart driving ahead of other international Tier 1 suppliers.

As a global Tier 1 giant, Bosch’s inherent advantages have become its secret weapon in terms of user experience. Bosch provides a wide range of products to automakers, including chassis control, steering control, braking systems, and more, all of which are closely related to autonomous driving.

Bosch has in-depth knowledge of automotive actuators, allowing it to propose more detailed control parameters in the fields of perception and planning after control. Many engineers have experience in braking or steering tuning.

Furthermore, within Bosch, the various product development departments use essentially the same system of tools, minimizing communication barriers and costs, leading to a more unified understanding of the entire system.

Therefore, the full-stack that Bosch emphasizes is not just about intelligent driving but rather a “comprehensive stack” encompassing actuator, chassis, and the entire intelligent driving domain.

Leveraging these advantages, Bosch ultimately achieved a high degree of integration between the vehicle’s intelligent driving system and actuators, providing consumers with a more comfortable driving experience.

Additionally, Bosch has made targeted enhancements to the safety of the intelligent driving system.

Bosch’s solution is currently the only one globally that uses NVIDIA Orin X chips and is developed based on the QNX operating system. Compared to the commonly used Linux systems in the industry, QNX, as a real-time operating system, enables the intelligent driving solution to achieve a higher level of functional safety, better security, and facilitates various validations overseas.

Under the trend of the “Four Transformations” in the automotive industry, the automotive supply chain model has also been reshaped. Previously, suppliers provided complete solutions directly, but now the trend of suppliers collaborating with automakers is strengthening, requiring automakers to have more flexible business models.

Bosch has also taken this into account.

Reportedly, Bosch’s current business model for intelligent driving is quite flexible; they can provide clients with “hardware + underlying software,” offer purely software models separately, or directly sell middleware, and even sell toolchains directly to users.

These models are currently being explored by Bosch and are under discussion with clients.

In addition to the already mass-produced high-speed NOA, Bosch is also actively testing and validating urban NOA. Che Dongxi similarly experienced the urban NOA product currently being tested by Bosch.

The urban NOA function will also be applied to Chery’s Xing Jiyuan ES, and thus Che Dongxi drove this vehicle in Shanghai for the experience. Bosch’s urban NOA solution is currently primarily based on high-definition maps, with plans to gradually shift to light mapping technology by the second quarter of next year, moving away from high-definition maps.

From Che Dongxi’s experience, the vehicle currently performs lane changes, overtaking, and stopping at traffic lights quite proficiently on urban roads. However, since the product is still in testing, the experience of the final mass-produced product may be even better.

It is reported that Bosch’s urban NOA function will gradually be rolled out in the second quarter of this year, initially landing in Shanghai and Wuhu, with plans to cover 20 cities within the year, including Hefei, Guangzhou, Beijing, Shenzhen, Suzhou, Chengdu, Tianjin, Changchun, Wuhan, Dongguan, Hangzhou, Chongqing, Nanjing, Changsha, Jinan, Shenyang, Xiamen, and Qingdao.

As an international Tier 1, although Bosch’s solution is primarily developed by the Chinese team, its ultimate target market is not solely focused on the domestic market, but also prepares for overseas expansion.

Bosch’s intelligent driving platform achieves ASIL-B across the board, which can safeguard domestic automakers’ overseas expansion, and Bosch and Chery’s collaboration will undoubtedly be implemented in overseas projects.

Additionally, Bosch’s solution may also directly target overseas automakers for more comprehensive international expansion in the future.

There is a saying in the industry that electrification is the first half of the smart electric vehicle industry, while intelligence represents the second half. Currently, both domestic and international automakers have made arrangements for electrification, completing the first half of the transformation.

The second half will be one of the key competitive points in the next decade; achieving success in intelligence has become an important criterion for evaluating an automaker’s future competitiveness. This change will not only impact automakers but also the entire supply chain.

In the wave of intelligence, players in the supply chain must keep pace with the trend of intelligence to avoid falling behind in future competition. Bosch actively laying out advanced smart driving is a solid step forward.

In the competition of the automotive industry’s second half, Bosch has already secured its ticket.