How to evaluate the launch of the Avita 11 updated version, OTA upgrade of Huawei HUAWEI ADS 20, HarmonyOS car system?

Avita 11: Huawei Intelligent Driving ADS 2.0

Many people are concerned about the intelligent cockpit of Avita 11 empowered by Huawei, Hongmeng Car System. It has to be said that for consumers, the intelligent cockpit is actually a more intuitive experience of the intelligence of new energy vehicles in the new era. However, the intelligence of intelligent driving is relatively less noticeable by consumers. Intelligent driving is actually hidden under the steering wheel, under the domain control chip, under the sensors, and even under the chassis of traditional vehicles. When you uncover the veil of Avita intelligent driving, what you can see is “HUAWEI Inside”, the newly upgraded ADS 2.0 that will be released this year.

Huawei Inside Mode (HI) is different from Huawei’s designed Smart Selection Mode. It is the smart car solution brand HI of Huawei’s smart car solution under the car manufacturer. HI includes more than 30 self-developed technologies for intelligent cockpit, intelligent driving, and lidar, etc. In other words, the product experience design and definition are mainly controlled and led by the car manufacturer.

Sensor hardware

The sensors of Avita 11 include 3 MEMS-style lidars, 6 millimeter-wave radars, 12 ultrasonic radars, and 13 high-definition cameras. The 3 lidars equipped in Avita are most likely Huawei’s self-developed 96-line MEMS-style lidars. I have explained MEMS-style lidars in detail in my previous article. In simple terms, they use a vibrating mirror to scan a two-dimensional plane. Some people may be concerned about whether the vibrating mirror will be restricted by amplitude. Don’t worry, MEMS-style lidars usually use multiple devices together, such as multiple sets of transceivers. The diagram below is a patent diagram of Huawei’s lidar.

Diagram:

In the diagram, there are multiple ranging modules. The light emitted by multiple sets of devices first passes through a mirror to adjust the spatial structure inside the lidar and reduce its size. Sharing a MEMS mirror also facilitates the control of multiple sets of lidars by means of mirror oscillation, enabling multiple beams to scan the plane simultaneously.

These 3 lidars equipped in Avita 11 have a detection range of up to 150 meters with a 10% reflectivity, a field of view of 120° × 25°, a frequency of 25Hz, and a resolution of 0.25° × 0.26°.

Assisted driving

Currently, the advanced functions of ADS for Avita include City NCA (City Navigation Cruise Assist) and AVP (Automated Valet Parking), covering the full range of point-to-point closed-loop intelligent driving scenarios, including urban areas, highways, and parking.

When evaluating the level of an intelligent driving system, in addition to the quantified data of the average takeover mileage (MPI), which indicates how many kilometers on average need to be taken over manually, there are also more intuitive functions for us. After all, the process of obtaining a driver’s license involves completing various stages, such as parking, S-turns, hill starts, parallel parking, and road driving.

On the road, experienced drivers can summarize some difficult and accident-prone scenarios, such as the scenario of making an unprotected left turn, which means making a left turn at an intersection without a left turn signal. This is a comprehensive test for the intelligent driving system, as it requires the ability to perceive oncoming vehicles in the oblique front lane. If the planning is too conservative, the system will keep waiting in the traffic flow.

When Avita 11 waits to turn at an intersection, not to mention the cameras, just the three lidars on the front of the vehicle can cover a sensing range of 300 degrees FOV (Field of View). The advantages of using lidars to complement cameras are self-evident; it can reduce many misjudgments made by cameras.

Moreover, the advantage of lidars is that when the camera encounters scenes or objects that have not been labeled, the active sensing lidar can perceive point cloud data and handle many irregular obstacles, adapting to more scenarios, including nighttime, backlight, and tunnel scenes with large light intensity changes, accurately perceiving point clouds of obstacles ahead. After all, the logic of human driving is largely based on avoiding collisions during the process of driving. Cameras, like human eyes, have a smaller field of view at night and are often unclear under backlight. It takes time to adapt to scenes with large light intensity changes in tunnels. These scenarios are minor cases for lidars.

Avita 11’s AVP operation for automated valet parking is also extremely user-friendly. Besides parking in reserved spaces, it can also park in any available space during the driving process. When the target parking space is occupied, it will roam and search for other available spaces. Doesn’t this sound like a human driver? When parking, it can see the traffic signal lights and think twice before proceeding. This demonstrates perception in complex scenarios, including recognizing large vehicles, intersection stop lines, and traffic lights. The planning and control algorithms need to be skillful enough to achieve smooth start and stop without generating abrupt jerks. After all, the riding experience in a passenger car is also very important.

With the support of ADS 2.0, Avita 11’s perception capability has been enhanced, bringing the NCA (Navigation Cruise Assist) feature in urban areas without the need for high-precision maps. It can drive with or without maps. In scenes without maps, it uses the combination of navigation map information and real-time online perception of road structures by the vehicle, while in scenes with maps, it uses high-precision maps to provide sufficient information for planning and control. The first batch of NCA functions for 15 cities has been launched with ADS 2.0, and the second batch will be expanded to 45 cities within this year.

Regarding the debate on whether high-precision maps are necessary for autonomous driving, I have elaborated on it in my previous article. Having maps is a good thing, but scenes without maps rely more on perception capability. Fortunately, the architecture based on multiple sensors can significantly improve the upper limit and redundancy of perception capability. This is why lidars are highly valued, and the reason why lidars have been extensively introduced here.

The automated parking function of ADS 2.0 not only can handle multi-story parking garages and narrow parking spaces but also supports mechanical parking, namely, multi-level parking garages. I really hate parking in multi-level parking garages. The parking spaces are narrow, and it’s easy to scratch the wheels if you’re not careful. And some multi-level parking garages even require you to be careful not to hit the rearview mirrors when backing in. One can imagine how precise ADS 2.0 needs to be in recognizing parking spaces (perception capability), driving cautiously (planning and control), and parking in multi-level parking garages (positioning capability). This scenario alone demonstrates the significance.

With the support of ADS 2.0, Avita 11 will undoubtedly further enhance the Huawei Inside mode. After all, Chang’an, a century-old car manufacturer, boasts military-grade quality, and Huawei has great advantages in software algorithms. Not to mention Hongmeng Cockpit, which is among the best in the industry. With the capabilities provided by ADS 2.0, Avita 11’s assisted driving ability stands out among many new energy vehicle manufacturers.

Avita 11 Experience and Pros and Cons

This is a review of the Avita 11, a pre-11 car model with 90kWh battery capacity, two-wheel drive, in matte white color, running on HarmonyOS, and equipped with ADS 2.0 intelligent driving system. After driving it for over a month and covering a distance of 3000 km, I would like to provide a summary of my experience so far.

If I were to rate it on a scale of 5, where 3 represents average, the higher the score, the higher the satisfaction. Based on my comprehensive evaluation system, my ratings are as follows:

First, let’s talk about the satisfying aspects:

Automatic Driving

This is the fundamental reason for choosing this car. The purpose of this car purchase was to experience urban navigation assistance, and the options were limited to Xiaopeng and Huawei. We conducted a comparison test in early 2023, using the Huawei HI version of Polar Fox. At that time, both brands had their advantages and disadvantages, but overall, Xiaopeng had a higher level of completion. However, based on my recent experience with ADS 2.0 on the Avita 11, Huawei is now leading the way.

I know many people may find it hard to accept and think that this advantage is due to Huawei’s brand influence. If I hadn’t personally driven thousands of kilometers, I might have thought the same. Unfortunately, money, manpower, and organization can bring you anything you want.

Here are a few impressive modules of the automatic driving system. I will write a separate article about the pros and cons of this ADS 2.0 system:

• Automatic parking: It has gone beyond the capabilities of most ordinary drivers. The efficiency and stability of parking are acceptable. To put it simply, I have been driving for over a decade, but until I got this car, I would let the vehicle complete the parking as my first choice.

This is the distance after the vehicle has parked, and I wouldn’t dare to park that close by myself.

• Precise identification and prediction of adjacent vehicles, making it excellent in handling overtaking situations
• Good system stability with few unexpected shortcomings
• Strong ability to handle small radius curves, and can even handle some mountain roads

Exterior

The reason for choosing the Avita 11 over Queshijie and G6 lies in its appearance.

My focus on the exterior is mainly on the vehicle’s posture and proportions. The Avita 11 gives a very upright impression. As for the design language of the exterior, I won’t go into too much detail, but it is unique enough, and the matte white appearance provides a good texture.

Range and Charging

The range of my Avita 11 is 630 km according to CLTC standards, with a daily round-trip commute of 100 km, 80% of which is on the highway. The energy consumption ranges from 16-17 kWh, so it can run for about 540 km. This is equivalent to an actual range discount of 15%, which is considered normal performance.

The vehicle has a 750V voltage platform and a maximum charging power of 240 kW. Since Avita does not have its own supercharging piles, I tried it on the Nio S4 charger, and the peak power reached 220 kW. However, the duration of continuous charging is not long, and the compatibility is average.

Currently, public charging stations generally have piles with power ratings of 120 kW, 160 kW, and 180 kW, and the Avita 11 can be fully charged at most of these stations. The efficiency is good, but due to the large battery capacity, it still takes more than 30 minutes to charge from 20% to 80%. Personally, I think it would be perfect if it could be compressed to around 15 minutes. In that case, you could do some activities, use the restroom, check social media, and then be ready to go, achieving the same level of convenience as a gasoline-powered car.

Intelligent Connectivity

After updating to HarmonyOS, the infotainment system of the Avita 11 is almost identical to that of Queshijie except for hardware-related functions. I have previously commented on the Meizu Flyme system, which is like a smartphone manufacturer entering the car infotainment market and competing at a lower level. I believe people will have similar impressions of the Avita 11’s system. However, for now, it is not as good as Queshijie in terms of compatibility. It is not as smooth and occasionally experiences lag. The recent OTA updates have made some improvements, and I hope there will be further optimization in the future.

---

Now, let’s talk about the aspects that are not satisfactory:

Driving and Handling

This is one of the most disappointing parts.

First of all, the brake pedal feel is unsatisfactory. One of the drivers in our team is responsible for performance evaluations in car companies, and he reminded me before buying this car that the brake calibration is problematic. At that time, I didn’t take it seriously and thought it was just a matter of personal preference. After driving 3000 kilometers, I still haven’t found a good brake pedal feel in this car. The initial step is too soft, and the brake force and pedal position are not linear, making it difficult to control the brakes as expected, thus affecting my driving confidence.

Next is the steering system. The steering response of this car is very sensitive. As soon as you turn the steering wheel, the front wheels immediately respond. Conventionally, such tuning would be considered more sporty. However, the problem with the Avita 11 is that despite its sensitivity, the steering system’s overall completion is not high. For example, the “harmony between the driver and the car” feeling, which means the feedback from the steering wheel to the driver and the feedback from the vehicle to the driver, matching the driver’s intentions. The Avita 11 hasn’t achieved this yet and has a strong “electronic taste” in its steering.

Next, let’s talk about the chassis system. Similar to the steering system, this car’s chassis feels relatively stiff and sporty when driving initially. However, after driving on different road conditions, I found that the chassis is too stiff and lacks flexibility. During high-speed driving, small vibrations cause the overall body to shake, affecting stability and peace of mind. On undulating and uneven roads, the body movements are also more noticeable.

In summary, the driving and handling of this car cannot match the “sportiness and individuality” of its appearance. It still has a lot of room for improvement in terms of refinement and optimization to bring a sense of luxury.

Comfort

When I talk about comfort, I mainly refer to the negative impacts of driving and handling dynamics. Due to the soft brakes, quick steering, stiff chassis, and the inherent characteristics of electric vehicles with fast acceleration, the lateral and longitudinal movements of the vehicle are quickly transmitted to the passengers. These movements may differ from the passengers' visual perception and self-awareness, resulting in motion sickness.

In the past month, I have taken my family on two or three trips with the Avita 11, and each time the rear passengers experienced motion sickness. This has caused me extra stress, so now I mostly use this car alone.

As for the front seat cushion being short, I personally don’t find it problematic (my height is 182cm). Because I’m used to adjusting the seat position lower, at this position, my left foot rests on the footrest, and my legs are stretched out, providing a larger contact area with the seat cushion. Thus, I haven’t experienced any insufficient support from the seat cushion.

Interior

The interior design of the Avita 11 is somewhat “overdone” or sacrificing human-machine interaction for the sake of styling. In my opinion, it is somewhat misguided.

The most typical example is the design of the additional screen. Whether it is the position or size, its usability is not strong. It could have been removed to save costs, as most users probably wouldn’t mind.

The bottom corners of the central control screen interfere with the legs of the driver and front passenger. They are right at the position of the knees and often hit them, causing pain.

To accommodate the two glass sections for the driver’s head, the sun visor had to be made foldable, but it is not easy to use.

Others

Audio System

Although I have mentioned some points of dissatisfaction earlier, if I were to evaluate it overall, the biggest drawback would be the audio system. Of course, this is based on the assumption that I am a relatively passionate music lover.

For a car priced at 300,000 yuan, the audio system of the Avita 11 is unacceptable. The high frequencies have roughness, the low frequencies lack elasticity, the reverb and soundstage are not well tuned, and the sound is neither immersive nor infectious. If compared to the Bose or Denon systems equipped by some independent brands of the same price range, the difference would be quite significant. Even when compared to the ordinary audio systems of some traditional luxury brands, there is still a noticeable gap.

This is something I complain about every time I play music, and it deepens my disappointment when I realize that “the smarter the driving, the quieter the cabin, the more my expectations are not met.” It is because the vehicle provides me with ample conditions to appreciate the details.

Therefore, I have also come to a realization that as intelligent driving continues to strengthen and alleviate the driver’s burden, these entertainment-related functions and performance will become even more important.

Maps

For the Hongmeng system, my biggest criticism is the Petal Maps. Compared to Gaode Maps, there is indeed a gap in the real-time accuracy of the data. So, after the in-vehicle navigation suggests a route, I still use my phone to compare.

Conclusion

In summary, this car meets my requirements for autonomous driving and unique appearance, so it is worthwhile for me. Now, I am mainly waiting for the release of map-less navigation assisted driving in cities. After all the talk from various companies, it’s time to see if they can deliver on their promises.

Avita 11 Facelift: The Age of Intelligent Driving

According to the data from Autohome, Avita 11 delivered a cumulative total of 10,755 units in the first half of the year. Sales also need to be considered in relation to price. The price range of Avita 11 is relatively high, with an official guide price ranging from 319,900 yuan to 600,000 yuan. The average transaction price of a single car is around 360,000 yuan, which is comparable to the BMW X3 and surpasses the FAW-Audi Q5L, second only to the Beijing-Benz GLC.

“New luxury” is on the rise, and intelligence is a battleground. The launch of the Avita 11 facelift is obviously aimed at consolidating this advantage.

The era of intelligent driving in the automotive industry started around 2015 and has now blossomed with various competing technologies. As someone who has tried many car models, I have always wanted to understand whether car companies genuinely care about the needs of the general consumers.

“What? Can car companies not understand without you?”

If it were a specific group of users, I might not have the right to speak, but if it’s about all consumers, as an automotive engineer and car enthusiast, I believe I can represent the thoughts of many users—intelligent driving should have strong “tentacles” like the human brain, without crashing when encountering unfamiliar situations.

The goal of the Avita 11 facelift is exactly that.

1. Expanding the “tentacles”—we are experiencing the “age of small ball” in intelligent driving

Many people in their 70s and 80s have experienced the era of poor mobile signal. At that time, you had to find the right angle when making a phone call in rural areas until base stations were massively deployed and hardware was improved to enhance mobile signal transmission and reception, solving the problem.

The reason why intelligent driving lacked sufficient “tentacles” in the past is similar. On the one hand, it heavily relies on high-precision maps, resulting in regional limitations. On the other hand, it requires high-performing perception hardware, and many vehicles lack the ability to support real-time decision-making. Therefore, they must rely on high-precision maps, making it difficult to break free from this limitation.

Of course, this is also related to algorithms and other factors. The industry is making steady progress in this area, and I believe it will only get better in the future.

How to let more people experience the fun? This kind of thinking has always existed in the basketball world as well. NBA used to be a game for tall and strong players, but how many players over 2 meters tall are there in reality? The NBA’s solution was to open the “age of small ball,” encouraging fast-paced offense and making agile three-point shooters the “regular weapons” rather than just icing on the cake. The emergence of small ball tactics subverted the traditional playstyle centered around centers and offense-defense, abandoning slow-moving centers and instead using fast breaks and three-point shooting. The defining characteristic of this playstyle is the rapid transition between offense and defense by the players, with all five players taking turns with the ball and being able to shoot three-pointers during the attack. Lei Zhenzi must envy the MVP title that Steph Curry won with over 400 three-point shots in a single season.

Regarding how to solve the “high-precision map problem,” we are also experiencing the “age of small ball” in the field of intelligent driving.

The age of small ball has three main characteristics—endless switching, fast pace, and three-pointers. The “age of small ball” in the field of intelligent driving is the same, and the typical example is the Huawei ADS 2.0 equipped on the Avita 11.

2. Huawei ADS 2.0 is the epitome of the “age of small ball” in intelligent driving

Let’s examine whether it fits the core characteristics of “endless switching, fast pace, and three-pointers.”

(1) Improving accuracy (the foundation of the small ball era is accurate shooting)

Without Steph Curry, small ball might not have become so popular globally. Whether it’s shooting threes with the ball or catching and shooting threes, it’s historically unprecedented. In the roadmap of “intelligent small ball,” some car companies have adopted a pure visual solution, while others have adopted a map + visual fusion solution. No matter which one, they all hope to maximize accuracy on their most suitable path.

Tesla’s other models use a pure visual solution based on BEV+label learning+occupying the grille+road topology, and various AI networks. Although this solution may have inherent weaknesses, its sales volume is even significant, making a strong case for a map-less solution.

Another development approach is to use solid-state lidar to improve accuracy. The Avita 11 is equipped with three solid-state lidars, which, compared to one, provide excellent performance in challenging urban traffic scenarios such as making unprotected left turns, merging on and off ramps, and avoiding irregular obstacles.

Because mass-produced cars currently use solid-state lidars that cannot rotate, there are blind spots in their field of view. The solution with three solid-state lidars ensures visibility from all angles.

(2) Closed-loop data (the small ball era emphasizes that every aspect must be flawless, and standing and waiting does not exist)

The joy of intelligent driving abruptly stops halfway, which is more uncomfortable than anything else. Only with a closed-loop data system can the user experience continue uninterrupted.

For example, Tesla’s Model 3 uses a “shadow mode” to collect data. Although controversial, it is still being promoted to achieve closed-loop data. In China, Tesla slightly modified it by adding the option to “Allow Autopilot to analyze and improve automatically,” seeking user consent to collect and analyze Autopilot data for improvement.

Currently, mainstream NOA functions require high-precision maps for global and local planning. The advantage is that it can make accurate predictions in advance and reduce the workload of real-time perception. However, high-precision maps face challenges such as low freshness (updated quarterly), high collection costs, and long approval processes. Currently, the research direction for NOA functionality is to use lightweight maps to reduce dependence on high-precision maps.

The HUAWEI ADS 2.0 used in the Avita 11 incorporates key road environment information on top of regular Amap navigation maps, providing the effect of a “simplified high-precision map.” This approach is similar to NBA teams watching opponents' game recordings before a match to formulate strategies. They already have preliminary solutions in mind and can focus on handling unexpected events during the game.

Using a high-precision map solution, not only is it difficult to cover suburban areas, but even within urban areas, it is challenging to maintain updated map coverage due to ongoing urban development. However, for models like the Avita 11, this is not a concern. ADS 2.0’s urban intelligent driving functions can cover 90% of urban driving scenarios, and it also adds features such as intersection U-turn, roundabout traversal, and lane change in construction scenarios, significantly reducing the pressure of driving in special road conditions.

(3) Fusion of multiple perception hardware (the small ball era emphasizes “five players taking turns with the ball, endless switching”)

Perception hardware is the car’s “eyes.” Different perception hardware has its strengths and weaknesses, so they need to be fused together.

Fusing multiple sensors such as lidars, millimeter-wave radars, and cameras can help improve vehicle recognition efficiency and enable the vehicle to take automatic emergency braking measures before collision risks arise, avoiding accidents or reducing accident losses.

Based on these, we can have some expectations for the new features added to the Avita 11, such as low-speed automatic emergency braking (LAEB), general obstacle detection AEB (GAEB) to deal with ghost obstacles, and emergency lane-keeping assistance (ELKA) to prevent unintended lane deviation. These advanced driver-assistance technologies that require the car to “keep an eye out” can indeed bring a safer driving experience.

Taking GAEB as an example, traditional AEB can only recognize obstacles defined as “whitelist” obstacles. However, the “whitelist” is not enough to cover the massive variety of obstacles that appear in actual traffic environments, making it difficult to fully solve the long-tailed scenes commonly encountered in the field of intelligent driving.

GAEB incorporates the application of the “GOD occupancy grid network algorithm” from the GOD network into AEB technology. By applying the “GOD occupancy grid network algorithm” and fusing multiple sensors such as lidars, millimeter-wave radars, and cameras, GAEB can help the vehicle recognize irregular obstacles outside the “whitelist,” such as unexpectedly appearing mounds, prone pedestrians, overturned vehicles, dropped large cardboard boxes, falling rocks, and fallen trees, enabling immediate identification.

In addition, the ADS 2.0 version introduces features such as lane change in construction scenarios and passing through ETC toll stations. The ETC toll station pass-through feature allows the vehicle to autonomously enter the ETC lane and control acceleration and deceleration without the user taking over. The fact that the version pays attention to specific road sections indicates that there is no downgraded perception capability.

3. Conclusion

Everyone is talking about intelligence now, but what is the difference between the intelligence of luxury brands and non-luxury brands?

The answer has become increasingly clear—seamless and extensive intelligent driving experience.

Car companies are trying their best to solve the problem of how to ensure faster and more accurate intelligent driving recognition. Currently, three approaches could work. One is a pure visual solution that relies on users to collect a large amount of data to illuminate blind spots. Another is to improve the perception hardware’s capabilities (such as adding lidars) while relying on user data collection. The last approach involves “lightweight maps combined with self-collection” and improving perception hardware (adding lidars).

All three approaches have the potential to work. In terms of the urgent needs of users, the third approach allows users to enjoy the intelligent driving experience as soon as possible. The Avita 11 facelift uses this third approach, ensuring that it has won time. In terms of technical features, it also aligns with the style of the “age of small ball” in intelligent driving.

Huawei’s involvement and price reduction in the Avita 11 facelift

I believe that the core of the Avita 11 facelift lies in two main points: Huawei’s involvement in deepening the collaboration and lowering the price threshold.

By “Huawei’s involvement in deepening the collaboration,” it mainly refers to the upgrade of the in-car system to Harmony OS and the upgrade of Advanced Driver Assistance System (ADS) 2.0. Among these, the former can be seen as a critical transformation.

As a product of the strong collaboration between CHN (Changan, CATL, Huawei), I personally had high expectations for the Avita 11.

Recently, the Avita 11 has gradually gained momentum, but overall sales are still not ideal.

It should be said that there are several reasons why the development of the Avita was hindered in the past, but I won’t go into detail here.

In terms of products, one thing is worth mentioning:

When the Avita 11 was first released, the in-car system with Harmony OS was already well-known, and the advantage of the in-car system was widely recognized as contributing to sales.

On the contrary, the Microservices-based Domain Control Unit (MDC) was limited by the development of advanced driver assistance systems, so it was difficult to form a strong user perception in the short term.

The Avita 11 adopted Huawei’s MDC, and the underlying in-car system also adopted Harmony OS, while the in-car interaction and ecosystem were developed by Changan.

As a result, as we all know, software technology, team building, and user understanding are not the strengths of traditional automakers.

The reason why the Avita went for such a choice can only be attributed to the competition for dominance in the ecosystem.

With the introduction of the Harmony OS version, it is clear that Changan has realized something.

It is not efficient for automakers to form software teams within the company, and the cost cannot be reduced, and the user experience is not good enough, so why bother.

In the short term, instead of focusing on the “elusive” software ecosystem benefits, it is more important to improve the in-car experience and sell cars.

In the long term, this move may have a significant impact, but at present, “let professionals do professional things” undoubtedly benefits all parties involved.

As for the upgrade of Huawei ADS 2.0, it is a natural progression.

After the Avita M5 Smart Driving Edition was equipped with ADS 2.0, there were numerous live videos and car owner testimonials, which created a buzz comparable to the Harmony OS smart cockpit.

Initially, the Avita 11 used a high-end advanced driver assistance solution from another manufacturer, so upgrading to ADS 2.0 is only natural.

The price reduction is also quite evident.

This year, there is intense competition in the market for intelligent electric vehicles, and price wars are frequent, so the Avita cannot be an exception.

The rear-wheel drive Harmony OS version with a 90-degree battery has a starting price of exactly 300,000 RMB.

And this price of 300,000 RMB is clearly aimed at maintaining the brand’s positioning at the 300,000 RMB mark, with the added benefit of discounted deposits, resulting in an actual starting price of 280,000 RMB.

The rear-wheel drive Harmony OS version with a 116-degree large battery has a starting price of only 335,000 RMB, and with the discounted deposit, the actual price is only 315,000 RMB.

It is quite competitive for the 116-degree large battery to achieve this price.

Overall, the Avita 11 Harmony OS version has significantly strengthened its competitive power. It can be said that it is a facelift that shows sincerity.

The Avita 11 already had a good foundation, with the only minor drawbacks being the in-car system and slightly higher power consumption per 100 kilometers.

The upgrades in the Harmony OS version are crucial, and the price is also quite sincere. I believe that the product has become stronger at the level of the car itself.

There is still a lot of work to be done in terms of Avita’s channels and brand development. I hope that the future collaboration between CHN can have better development.

HarmonyOS price cut, outstanding performance.

Why did Avita create the #Avita 11 HarmonyOS version#?

In essence, it is the “champion version”, with the focus not on HarmonyOS, but on lowering the threshold.

• 90-degree rear-wheel drive version: 300,000 yuan
• 116-degree rear-wheel drive version: 335,000 yuan
• 90-degree luxury version: 350,000 yuan
• 116-degree luxury version: 390,000 yuan

Considering the previously provided 20,000 yuan optional equipment fund, the essential update of this HarmonyOS version is to reduce the threshold by around 40,000 yuan. However, in order to maintain the high-end image, the starting price is still set at 300,000 yuan, with a 20,000 yuan deposit-induced discount.

When you click on the purchase page, the price is directly listed as starting from 280,000 yuan.

This price is quite sincere. I even think that the most worthwhile option is to choose the entry-level version and add a 2,000 yuan set of wheel hubs. You don’t need to select anything else.

This way, you will get a 90-degree battery, rear-wheel drive, five-seater SUV with a first-tier intelligent cockpit and first-tier advanced driving assistance. Except for a slightly smaller space, it can compete with the same level Model Y or Zhiyi LS6.

You can see the performance of HarmonyOS in our previous video.

If you like, you can also consider adding 25,000 yuan for a four-seater. In this era where more seats make the car look cheaper, the four-seater version will make it resemble a car worth 400,000 to 500,000 yuan.

Avita ADS2.0: Strong Sensors, Self-built Mapping, Powerful Perception Algorithms

Starting with the Conclusion

Huawei ADS2.0 can be considered as a top-notch product in the intelligent driving industry. Among the best cabin products of Hongmeng Cabin, Avita 11 is the first one that excels in both intelligent driving and cabin features. Its outstanding performance positions it at the forefront of the high-end market.

What Makes Avita ADS2.0 So Powerful?

1. Strong Sensors

Avita II is equipped with a total of 34 sensors. Among them, there are 3 lidars, one in the front, one on the left front, and one on the right front. There are 6 millimeter-wave radars, one in the front and one in the rear, as well as one on each of the four corners. There are also 4 cameras in the front, with one pair of binocular cameras, two narrow and wide cameras (commonly known as big eye and small eye), 4 side cameras, and 1 rear camera. In addition, there are also 4 surround-view cameras and 12 ultrasonic radars. The performance and quantity of the sensors are at the top level. Why does Avita 11 need to maximize the hardware performance?

First of all, it is related to the development of intelligent driving technology. The current development route for intelligent driving technology is not yet clear. For example, the perception aspect emphasizes heavy sensing and light artificial intelligence. Specifically, it relies heavily on sensor output, including various lidars, millimeter-wave radars, ultrasonic radars, and cameras in the front-end perception. The back-end focuses on decision-making, planning, and control.

On the other hand, the light perception and heavy artificial intelligence approach mainly relies on camera output. It believes that humans can drive solely based on visual perception, so why can’t machines? As the main sensor, after receiving raw data from the camera, the Huawei ADS2.0 system trains the machine through semantic segmentation, feature extraction, and a multi-layer neural network algorithm model on the NPU (Neural Processing Unit), allowing it to have a better understanding of the environment, which is close to human perception. Therefore, Avita II’s hardware configuration can meet the requirements of different development routes in the future of intelligent driving.

Secondly, I believe that Avita’s hardware configuration fully considers the future upgrade to Level 3 (L3). Although Huawei ADS2.0 is installed on different vehicle models, the decision to develop L3 functionality is not the same for all models. If a certain vehicle model wants to develop L3 functionality, it needs to fully consider the L3 requirements in the early stage, including the design of the vehicle’s electrical architecture, sensor design, vehicle power supply architecture, and communication architecture. These requirements correspond to additional vehicle costs. Vehicle models priced below 300,000 RMB cannot afford these additional costs, but Avita II, which is priced between 319,900 and 600,000 RMB, can bear the additional costs.

The implementation of L3 autonomous driving systems is not a distant ideal. The Chinese government has already formulated or implemented regulations in this regard. For example, in August 2022, Shenzhen Special Economic Zone issued the “Regulations on the Management of Intelligent Connected Vehicles” for L3-L5 autonomous driving systems. All autonomous driving vehicles sold in Shenzhen must obtain approval from the Ministry of Industry and Information Technology or the Shenzhen Municipal Government before they can be launched on the market. The China Automotive Technology and Research Center is drafting regulations on “Intelligent Connected Vehicle Automatic Driving Data Recording System,” and the Ministry of Industry and Information Technology has issued a draft of the “Guidelines for the Access Management of Intelligent Connected Vehicle Production Enterprise-level Products” and the “Opinions on Strengthening the Access Management of Intelligent Connected Vehicle Production Enterprise-level Products.”

At this year’s Avita Technology Conference, Richard Yu also stated that Avita will be the first car company to achieve L3-level intelligent driving. Based on the above, I believe that Avita II’s intelligent driving hardware is fully prepared for mass production of L3 autonomous driving systems.

2. Strong Perception Algorithms

Before BEV (Battery Electric Vehicle) was applied in the field of autonomous driving, traditional perception used cameras to detect 2D information, and millimeter-wave or lidar sensors to output 3D information. Rule-based algorithms merged the information from various sensors and used the fused result as input for decision-making, planning, and control in the backend. The fusion process involved prior filtering of non-whitelisted targets. If a target is blocked, perception based solely on single-frame camera data cannot associate information across multiple frames in the front and back, making it impossible to estimate or predict the obscured target’s information.

The camera arrangement on the vehicle is based on a front perspective view and projects the real-world in a perspective view (from the world coordinate system to the image coordinate system). This view is similar to human perception: objects appear larger when they are near and smaller when they are far away; parallel lines converge in the distance. It is difficult to accurately determine the distance between the vehicle and the target in front, as well as the three-dimensional shape and size of the target.

BEV directly processes the perspective views captured by all cameras and inputs them into a neural network model after calibration. The features extracted from these views are then correlated to project a bird’s-eye view, which reflects the surrounding environment. This approach avoids the scale and occlusion problems caused by the front perspective view in traditional perception. It provides a more comprehensive display of information.

However, BEV’s view is still in 2D. In the field of computer vision, detection models often use rectangular bounding boxes, resulting in the loss of important elements. It lacks height information to accurately recognize road signs, and when there are protruding objects on vehicles, computer vision tends to ignore these features. This is crucial for path planning in autonomous driving, as it determines drivable areas.

The Grid-based Occupancy Distinction (GOD) approach further enhances BEV. The “GOD Grid-based Occupancy Distinction Algorithm” voxelizes the 3D space, dividing the world into small cubes similar to Rubik’s Cubes. Based on the data from sensors such as lidars, the algorithm judges the state of each cube and determines whether it is occupied and whether the object occupying the cube is moving or stationary.

3. Strong Self-built Mapping Capability

Most of the current navigation-assisted driving (NOA) technologies on the market heavily rely on high-precision maps. High-precision map data presents challenges for the development and application of NOA technologies, especially in urban areas, due to limited coverage, high procurement costs, strict qualification requirements, lengthy approval processes, and high maintenance costs. The current industry focus is on how to make these technologies work with or without high-precision maps.

The BEV approach has the advantage of integrating rich semantic information captured by cameras and sending it to a neural network model for perception of static information (such as lane lines, curbs, and road signs) and dynamic information (such as vehicles, pedestrians, and bicycles). This process naturally generates a lane-level road network topology, which is then matched with road-level attributes and path direction information provided by navigation maps. This achieves real-time construction similar to a high-precision map with navigation information.

Why is HarmonyOS (Hongmeng) so User-friendly?

Recently, I read a book called “The Laws of Simplicity,” which summarizes how companies like McDonald’s, Ford, Apple, and Ikea use minimalistic design to simplify product development, reduce product costs, and establish an unbeatable position in the market. Simplification is divided into two aspects: simplification of price and simplification of proposition. How to simplify the proposition includes convenience, practicality, and aesthetics. Avita 11’s HarmonyOS Cabin serves as a good case study in demonstrating the simplification of proposition. Let me explain the three-step cabin proposition simplification.

1. Convenience - Easy to use, low learning cost

In our previous impressions, there was a significant difference between the central control screen of a vehicle-grade product and consumer-grade smartphones or tablets in terms of hardware and user experience. Vehicle-grade products have higher hardware requirements due to the complex usage environment in vehicles (e.g., high temperature, humidity, vibration, and dust). Additionally, considering factors such as functional safety and information security, the application ecosystem of central control screens is relatively closed. In contrast, the HarmonyOS system adopts a similar operation mode as smartphones and tablets, reducing the learning curve for users. It is easy for older people and children to operate. It also defines a personalized shortcut bar (Smart Dock) at the bottom, while ensuring driving safety and ease of use.

With a large visual interactive area on the 15.6-inch central control screen, most settings and information can be accessed in just one step, achieving zero-layered operation. This significantly reduces driver distractions and enhances driving safety, allowing users to access information at a glance.

2. Practicality - Diversified and Personalized Functions

Diversified functions are implemented without compromising convenience. Diversification does not simply involve copying applications from smartphones or tablets to the in-vehicle system. Although smartphones and in-vehicle systems are both based on the HarmonyOS operating system, the large screen size requires targeted adaptation and development, rather than simply stretching or directly using smartphone applications.

Avita 11 is equipped with a Huawei application market specifically designed for intelligent in-car scenarios. It includes apps such as Huawei Maps, Huawei Music, Huawei Video, Kugou Music, Ximalaya, Netease Cloud Music, Pocket Stories, Youku, Leishi, Smart Fueling, and UnionPay Cloud Quick Pay. Based on insights into how users use apps on their smartphones, Avita’s in-vehicle HarmonyOS system has personalized some apps, such as Tencent Video, QQ Music, and Leishi KTV, to meet specific customer needs.

The goal is to eliminate information silos between mobile phones, vehicles, and homes. Through the HarmonyOS system, various hardware devices can be interconnected, expanding the usage scenarios of each device. Through a smart living interface, users can remotely control air conditioning and lights, enabling them to turn on the air conditioning and lights before arriving home and automatically turning them off when leaving home. This expands the connection from the phone to the vehicle to a smart home and, in the future, to the Internet of Everything.

3. Aesthetics - Exterior Quality and Sensation

The exterior quality of a product creates an impression for users. How can we create a comprehensive and pleasant experience for users without any verbal description? A simple and high-quality design provides users with an experience that sets the tone for how they perceive a product.

In the design of Avita II’s HarmonyOS cabin system, great attention has been paid to the size of texts, icons, and virtual buttons. They possess the DNA characteristics of Avita II and are clear, understandable, and rich in design quality. By employing industry-leading 3D rendering and animation technology, the system replicates the vehicle’s condition in detail, including the size of air conditioning vents. Different driving modes are accompanied by corresponding light and shadow effects. For example, the energy-saving mode is presented with a green background and leaves swaying in the wind. In terms of ambient lighting design, the interior lights can present different colors and brightnesses in sync with the music, making the cabin more vibrant, creating an atmosphere, and enhancing the interior ambiance. The entire interaction design adds fun and playfulness through light and shadow rendering and animation, providing users with a delightful experience.

Conclusion

Huawei’s automotive business covers intelligent connected vehicles, chassis control, and electrification, among other areas. In the fields of intelligent driving and intelligent cabin systems, Huawei has achieved outstanding results and proven its product power in multiple mass-produced vehicle models. The Avita II upgrade combines Huawei’s intelligent driving and intelligent cabin systems. To borrow a phrase from “Romance of the Three Kingdoms”: “Having one of the Sleeping Dragon and the Young Phoenix will secure the world!” If Huawei ADS2.0 is the Sleeping Dragon, then the HarmonyOS system is the Young Phoenix. Having one of them allows Huawei to take the lead in the intelligent connected vehicle field. Equipped with the Sleeping Dragon and the Young Phoenix, Avita II can secure a place in the high-end market.

Advantages of the Avita 11 hardware upgrade and HarmonyOS Cabin

The price of the Avita 11 has been lowered, and the actual price, including benefits, is now within 300,000 yuan.

With the price reduction and the upgrade in capabilities, I believe it has good competitiveness. Huawei’s strength in intelligent driving is indeed very strong. Jokingly speaking, although it is a bit of a stretch to say that it is far ahead, it is indeed a leading player in the industry.

Huawei’s ADS2.0 advanced intelligent assisted driving system, I believe many friends who are interested in new energy vehicles have heard of it. In the first half of the year, the Wanjie M5 was the first to adopt this intelligent driving system. As an onlooker, I have also seen many road tests on highways/city roads, and the effects are indeed impressive.

With the outstanding performance of Wanjie M5, although the Avita 11 and Wanjie M5 both use the same system, they are different models from different brands. I compared the differences between them in terms of hardware.

Wanjie M5 already has a complete set of hardware sensors in terms of types and quantities, but the Avita 11 surpasses it.

In addition to being equipped with 12 ultrasonic radars, the Avita 11 has more hardware in other categories compared to Wanjie, especially in terms of lidar. The Avita 11 is equipped with three lidars, one in the front of the car and one on each side of the wing plate, covering a 300-degree area outside the rear of the car, almost achieving a full forward view.

With its comprehensive set of sensors, the Avita 11’s data collection and coverage of blind spots and weak perception areas are definitely stronger than solutions with fewer sensors. I believe that comprehensive data collection is the foundation for intelligent driving. The basic data must be detailed and complete enough to assist in subsequent software judgments and decisions. At the same time, a multiple hardware solution also provides sufficient redundancy, which is beneficial for improving driving safety.

Here I also need to mention the Avita 11’s high-precision maps and sub-meter high-precision positioning system. They are compatible with or without maps, which shows a proactive decision-making approach that does not overly rely on a single data source.

Based solely on the current hardware match of the Avita 11 with the upgraded Huawei ADS2.0 intelligent driving system, I have reasons to speculate that the actual effect will be at least on par with the Wanjie M5. Referring to the performance in the first half of the year, it is very likely to demonstrate the intelligent driving capabilities of leading models.

The HarmonyOS car system, which has also been upgraded, is equipped with the HarmonyOS system. This is a true version of the HarmonyOS cabin.

As for my understanding of the HarmonyOS system, I started out as an onlooker. There were a lot of discussions on the internet about things like Android-based shells. Then, by chance, the old Huawei phone that was retired and given to my child as a study device received a push notification for the HarmonyOS system. So, I upgraded it out of curiosity, and I have to say, wow, it’s much smoother than before.

The phone which used to lag for various reasons suddenly became silky smooth, and that’s something.

In my opinion, there are two aspects that reflect the advantages of this system. First, based on the HarmonyOS system’s characteristic of interconnecting everything, this central control large screen system serves as the core to achieve interconnection with smartphones and other devices, with functions such as screen mirroring and interactive features. I have watched some videos, where the interaction between the main and auxiliary screens and the digital devices seems quite simple. In the future, we can look forward to more interconnections between devices to create an expanded cabin functionality.

Secondly, the home entertainment integration of the car system. I believe this is closely related to the fact that large-sized vehicles are increasingly focusing on the home usage aspect. The Avita 11 offers a “Scenario World” feature, providing six modes: rest, movie watching, karaoke, pet, car wash, and temporary leaving.

This setting is quite interesting, quantifying common scenarios, somewhat similar to the common “one-click” function, and also reflects the evolution of the car system towards simplicity and user-friendliness.

I truly believe that the ease of use of the car system is very important. Now, new energy vehicles have powerful and diverse functions, and if the convenience of use is not sufficient, the learning curve for users will greatly increase, which is not conducive to user vehicle usage.

I have previously test-driven the Xiaopeng G6, and I truly felt that the practicality of advanced driving assistance systems, such as Xiaopeng and Wanjie, is quite good. The Avita 11 should not disappoint in this regard. I just hope that the prices of advanced driving configuration can be lowered in the future to promote wider adoption and popularity.

Hongmeng Cockpit + Huawei Advanced Intelligent Driving System: The Core Competitiveness of Avita 11

The deep customization of the Hongmeng Cockpit + Huawei Advanced Intelligent Driving System ADS 2.0 empowers the soul of Avita 11, which is a new answer to the market based on the strengths of various Chinese companies.

Avita 11 has sold more than ten thousand units in the first seven months of this year, competing for its own place in the 300,000 to 400,000 price range amid the fierce competition of new forces in the market.

[Sales Performance of Avita 11 in 2023]

At the opening, the thunderous fencing performance by Avita CEO Tan Benhong, who achieved a breakthrough of winning a gold medal in fencing at the London Olympics, actually signifies that Avita has reached a crucial moment when it must unsheathe its sword. In the increasingly competitive market of 300,000 to 400,000, it has revealed its core trump cards. The two most important trump cards in the redesign are the Hongmeng Cockpit + Huawei Advanced Intelligent Driving System ADS 2.0.

[Fencing Performance, Moment of Unsheathing]

1. Thanks to Huawei’s abundant technological accumulation and integration capabilities of software and hardware, the Hongmeng Cockpit enables a seamless integration with the vehicle owners and a more diversified in-car ecosystem.

The Hongmeng Cockpit has various functions. I would like to highlight two that I think are excellent experiences. One is the most important core of intelligent travel, which is map navigation. Petal Maps can combine with the Avita 11 vehicle to track the remaining battery power in real time. If the current battery power cannot reach the destination, it can remind the user of the reachable mileage range and search for and recommend charging stations along the navigation route. This ensures a clear understanding of the remaining range and eliminates the anxiety of charging. Petal Maps can also automatically transfer the navigation route from the car’s display to the user’s smartphone or smartwatch after getting out of the car, achieving seamless transition of navigation information. This is an excellent navigation experience for those who are directionally challenged.

Another important aspect of intelligence is the diversified satisfaction of in-car entertainment needs. Similar to smartphones, the Huawei AppGallery in the Hongmeng Cockpit supports the download of over a hundred popular applications and will expand to provide more high-quality applications in the future, bringing users a constantly fresh smart experience and fulfilling a wide range of app needs.

2. Thanks to Huawei’s deep cultivation in the field of advanced intelligent driving systems, the Huawei ADS 2.0 brings a more practical and grounded assisted driving experience.

The Avita 11 Hongmeng edition is equipped with the Huawei Advanced Intelligent Driving System ADS 2.0, which will unlock multiple industry-leading features through OTA updates, covering highway, urban, and parking scenarios. Simply put, compared to previous versions, ADS 2.0 has more accurate recognition, stronger algorithms, safer and more reliable use, and a richer range of scenarios.

For example, it can precisely identify obstacles to enhance safety; it can realize intelligent driving with or without high-precision maps; it can further expand the use of advanced driving (AD) in cities and on highways to enhance users' peace of mind through precise recognition; the intelligent parking assistant can park in extremely tight spaces with a margin of only 0.4 meters, and supports parking in and out of mechanical parking spaces; the usage scenarios of Automated Valet Parking (AVP) are further expanded, achieving cross-ground parking capability and offering richer usage scenarios.

Of course, what is even more commendable is that Avita does not betray its existing customers. For all existing customers who placed orders and took delivery before August 24th, Avita will give them 40,000 E-value gifts (equivalent to 4,000 yuan), which can be redeemed in the online store. At the same time, by the end of October, they can also enjoy the free upgrade to experience the two major features of the Hongmeng Intelligent Cockpit and ADS 2.0.

In 2023, the automotive market will determine victory or defeat in an instant, and it will also separate life and death. We look forward to more car companies revealing more trump cards and offering consumers more diversified choices at this critical moment of life and death.

Electric vehicle price competition | Brand differentiation and delisting

[This battle, determines both victory and defeat]

In June of this year, Zhiji launched the LS7 Urban Edition by using a strategy of reducing configuration and using smaller batteries.

Bringing the price of medium to large-sized pure electric SUVs to below 300,000 yuan

Just a few days ago, the mid-to-large-sized extended-range SUV LanTu Free received an upgrade, with air suspension and dual motors for extended range, priced at 266,900 yuan, with an optional Apollo intelligent driving system for 10,000 yuan.

Pioneering the concept of “increased configuration and reduced price” in product upgrades

Yesterday, Avita received an upgrade, with all models now equipped with 90-degree battery placement, the Huawei Harmony cockpit, Huawei ADS advanced intelligent driving assistance system, 3 laser radars, 34 intelligent driving sensors, 750V high-voltage platform, Nappa leather… Starting at 280,000 yuan

Increased configuration and reduced price is just the basic operation, full-series flagship intelligent configuration is the real deal

The competition is heating up, with rapid iterations

—————————Dividing Line—————————

Why is this situation happening?

1⃣️ Market size means everything. The fewer people at the table, the more each person can get. Everyone wants to stay at the table and wants others to get off the table.

2⃣️ The current level of loss compared to future profits is insignificant. Don’t be reluctant to let go of the child just because you can’t secure the wolf.

3⃣️ The consequences of brands selling cars directly, the market is becoming more transparent. Without the constraint and buffer of a dealership system, brands are directly involved in retail. This is the necessary condition for this intense battle. Previously, sales numbers could be inflated at will, but now they are announced weekly.

Some observations and thoughts

1⃣️ Brands have already started to differentiate themselves, and the effect of the brand’s influence is now showing. Even with the news of price reduction, the impact varies. Some reductions make waves, while others only attract those with ulterior motives.

2⃣️ After the wave of price reductions comes the wave of brand delisting. Only when a stable brand pattern is formed in different price segments can the transition from chaos to order be completed. Thousands of troops cross the single-plank bridge. Only those who survive will be able to enjoy the spoils, regardless of how many stabs they receive or how much blood they shed.

3⃣️ Previously, we were still discussing the issue of soul and body. Now we are only discussing the issue of life and death.

Avita will not be the last one. Who will be the next one?

WM Motor will not be the first one. Who will be the next one?