With MARVEL R’s automatic parking, the average driver per capita

In the trend of the construction of the new four modernizations (intelligence, electrification, sharing, and networking) of the automobile, a large number of new things have emerged in the automobile industry, which makes people feel like they are doing what smart phones have done to functional machines . Even, in some extreme cases, a smart car is sloppily imagined as a smart phone with four wheels.

Seeing the SAIC "MARVEL R" launched on February 7th, you may have an illusion: it is not just a car, but an oversized 5G mobile phone.

In fact, for automobiles, the most important change is obviously not to become another digital life terminal for users so that they can grab the business of mobile phones, but to reshape the essential ability of automobiles to do things that mobile phones can’t do and do more and more. Great.

In other words, the most anticipated smart car revolution depends not only on who has the highest technology, but also on who is actually solving the pain points in the process of driving, and on who is making driving and riding a car become a reality. It's safer, more worry-free and enjoyable.

The world's first 5G smart pure electric car

If we must use one sentence to summarize our expectations for smart cars, it is: make driving more enjoyable.

As far as the current technology level and driving environment are concerned, this expectation can be roughly decomposed into several aspects: driving safety, assisted driving, intelligent reminders, natural interaction, and higher-quality in-car entertainment.

There are more subdivisions further down. Take the world's first 5G smart pure electric vehicle-MARVEL R as an example, and sort out the application scenarios of 5G V2X into 17 applications, including traffic light information push, parking start guidance, vehicle speed guidance, and traffic guidance. Fork in the road, conflict avoidance, etc. In the future, this number will expand to 50. In this regard, the more detailed the manufacturer, the better we will be as users.

Take the traffic jams that we often encounter as an example.

▲ Traffic jam on the road. Picture from: unsplash.com

The straight road is okay, overhead congestion can be a problem, and MARVEL R can achieve assisted driving with a 150m patrol radius, which can basically cover most of the domestic viaducts. In addition, in congestion, the vehicle can start automatically after braking within 30 seconds, without the user's frequent manual restart. In addition to congestion, MARVEL R also performs well in terms of automatic lane change, visual recognition reminders of traffic lights, and smart summoning. Overall, the smart experience of MARVEL R is very delicate and the scene coverage is large, which is more suitable for Chinese car owners.

To achieve this, in addition to 5G's inherent advantages of large bandwidth, low latency, and multiple connections, it is more important for car companies to understand the domestic driving environment and car owners.

MARVEL R's Anxiety Relief

Among the most terrifying things for novice drivers, parking is definitely one. And people have many expectations for smart cars, and the automatic parking display can be ranked. It is no exaggeration to say that if people can completely get rid of the pain of parking, people will have a much stronger driving desire.

In the past two years, semi-automatic and automatic parking technologies have been deployed on many mass-produced vehicles, but the complaints may be more than surprises. Among them, the most important slot points are focused on the identification of parking spaces and the time-consuming parking. Someone once did a parking test on a foreign electric vehicle brand in an underground parking lot, but he never expected that the link of parking space identification would get stuck. At that time, there were obviously many vacant spaces in the parking lot, and after the test car went around for a long time, the display screen was still blank and no parking space was found.

▲ Picture from: unsplash.com

This example is not to say that the brand's cars are not good, but to reveal an important fact: the domestic parking environment, especially the complexity of parking spaces, has more personalized algorithms and data for automatic parking. Claim.

Many automatic parking systems tend to be rigorous and slightly conservative. In contrast, MARVEL R's performance on automatic parking is more like a skilled and confident "veteran driver".

For example, its ability to recognize parking spaces is more compatible. There are bricks and grass on the ground, and parking is not delayed. It can also be parked smoothly in open parking spaces on rainy days and at night.

▲ Water in the road area on rainy days will affect parking space recognition, picture from: unsplash.com

Another point is more intuitive. It will plan the curvature of the reversing trajectory in advance, so that the steering wheel can rotate smoothly, and even if it can stop, it will never hit more. In addition, its parking posture is straight, and you can tell at a glance that it is the tactics of old drivers.

The reason why MARVEL R has this "confidence" style mainly comes from the "vision + radar" integrated automatic parking scheme it adopts. To this end, its hardware is well equipped:

2 front view cameras

4 Zhoushi cameras

4 surround view cameras

1 real camera

Millimeter wave radar 5

Ultrasonic radar 12

They allow MARVEL R to have a triple perception system to better model the surrounding environment. This is not only a basis for automatic parking, but also a starting point for future MARVEL R to continue to evolve in autonomous driving. This will not be shown for the time being.

In terms of software, SAIC has developed fully independent development of key algorithm software such as machine vision, lidar, and Slam.

The aforementioned situation where the parking space cannot be found is the algorithm's pot, but another important reason may depend on what kind of data is fed to the algorithm. In this regard, AI must talk about national conditions.

▲ Schematic of SAIC AI training platform, picture from: Fan Yi Shang Xing

In terms of data, SAIC has established a full-scenario test database based on the domestic complex road traffic environment, and has also cooperated with Tongji University to build an evaluation base for intelligent driving road testing. At the same time, SAIC independently built a dedicated AI training platform, forming a closed data loop. In addition, SAIC has established a distributed vehicle cloud crowdsourcing data collection platform. Not only smart cars are contributing data to SAIC, but also road environment data and driving behavior data can be obtained from traditional fuel vehicles.

Finally, I would like to mention a small but very practical function of MARVEL R-using a mobile phone to remotely control the vehicle forward and backward. I believe everyone has encountered a situation: you return to the parking space after finishing your work and find that the car parked behind is close to your body. At this time, let alone driving out, you may have difficulty driving the door. With this feature, you don’t have to panic, you can drive the car out on your phone.

Therefore, MARVEL R dared to say "the most understanding of the parking system in Chinese parking spaces", there are still two brushes.

There are still many "pain points + just needs" for driving such as automatic parking. I believe that only by having a deep understanding of the actual situation of domestic travel, so that the automatic driving system can truly be built on the "national conditions" and solve these problems one after another. A truly effective way to alleviate travel anxiety.

#Welcome to follow Aifaner's official WeChat account: Aifaner (WeChat ID: ifanr), more exciting content will be provided to you as soon as possible.

Ai Faner | Original link · View comments · Sina Weibo