Ali driverless vehicle route coordination: smart car expensive or intelligent road expensive?


Ali driverless vehicle route coordination: smart car expensive or intelligent road expensive?

As the last entrant, Alibaba unveiled its first new energy smart vehicle, a white-blue smart logistics vehicle the size of China Bus, at the just-concluded "Artificial Intelligence Summit" of the Cloud Habitat Conference.

"First of all, it implements the L4 autopilot capability, and it has the function of vehicle-road coordination, we will redefine the vehicle and road, and cloud three-terminal coordination." Chen Lijuan, general manager of the A.I. Labs, announced the news on behalf of the two-year-old team.

The American Automobile Engineering Association (SAE) classifies automobile driving into six levels, L0 to L5. The L4 is highly automobile driving, which is still required for roads and the environment. The highest L5 is completely automobile, taken over by AI and is not restricted by the road environment. Ali's UAV chooses the logistics backbone line as the application scenario, and gives the solution is "vehicle-road coordination". Light snow says the plan is for safer and lower costs.

Safety and cost is a difficult problem in the commercialization of automatic driving. According to Wang Gang, chief scientist of Ali AI Laboratory, the average price per vehicle is about 200,000 US dollars. Even if the bike is very intelligent, the high price can not be used by many people.

"China's high-frequency highway is very good, we can use very little money to make the road intelligent, at the same time, because the intelligent road can reduce the configuration of the car, reduce the hardware costs of the car, so that the overall cost of automatic driving is reduced, so that the technology truly universal. Wang Gang said.

One of the core technologies of the vehicle-road collaboration scheme is the perception base station, which means that the road has a God's perspective.

The sensing base station is composed of a variety of sensors and efficient computing units, suspended on the road, similar to traffic cameras. The sensing base station can sense the surrounding environment in real time and send the sensing information to the automatic driving vehicle with a working radius of 200 meters.

The perception base station has made up for the shortage of single car intelligence. Wang Gang demonstrated a scenario that perceiving the base station to pass the object in front of the road to the vehicle to help an auto driving vehicle beyond the line of sight. "We also did a lot of tests. If the car suddenly rushed out of a pedestrian in the 30 km/h scenario just demonstrated, 100% of the pedestrians would be injured, but with the help of the perception base station, we achieved a 100% success rate. This example fully demonstrates the power of cooperative intelligence."

Ali defines the perception base station as "a sense of shared economy." With such service concepts, the cost of automated driving and intelligent logistics vehicles could be greatly reduced in the future, says Mr. Shallow. "We have been thinking about the driving and evolution of self driving cars from the perspective of business, application scenarios, perceived accuracy and how to reduce costs."

In the scheme of Ali car road cooperation, intelligent road can greatly reduce the cost of bicycle intelligence. But for the deployment and cost of perceptual base stations, Snow said, is currently in talks, specific information is not yet easy to disclose.

But in terms of the definition of "sharing," Ali wants to use road-to-vehicle coordination as a breakthrough tactic for entering the UAV arena, as well as an ecosystem for autonomous driving.

5 months ago, Ali side revealed that Ali AI laboratory is conducting driverless research. By contrast, the Baidu Research Institute started its pilotless program in 2013, and Tencent set up an autopilot lab to develop pilotless technology in 2016.

Ali believes that road-to-vehicle collaboration is the ultimate form and future direction of auto-driving research and development, and that the core component of road-to-road collaboration is the perceptual base station, which is pioneered globally, not only greatly reducing costs, but also improving the safety of unmanned driving, and even for ordinary motor vehicles in the future. It provides intelligent road information service.

At the same time, some analysts believe that vehicle road synergy needs to lay signal base on roads, not only to build the infrastructure of autopilot cars, but also to build these infrastructures.

In fact, as early as the late 1950s, General Motors built a conceptual highway embedded in a large number of communications equipment, vehicles in the road information interaction to achieve a fairly elementary "automatic driving". This is probably the earliest "vehicle road coordination" plan in history. It focuses on communication, and Ali focuses on perception.

In addition, there are several automatic driving R & D teams involved in vehicle road collaboration technology in China. In September this year, Baidu introduced its vehicle-road collaboration progress with Ali Xuan and the Highway Research Institute of the Ministry of Transportation. Baidu's Apollo has more than 100 partners and tens of thousands of developers.

Ali wants to be more pragmatic. "The perfect combination of technology and business, to move the entire market. "Scientists joining the industry does not mean that this thing has been formally established in business logic," said Xue. If automatic driving is a technology, it does not answer deep business problems. The business problem is still returning to the essence, that is, whoever uses it, where it can be used and whether it can afford it. She sees the perception base station's release as "guiding the whole industry".

Snow introduced, the release of intelligent logistics vehicles, the current number of less than 100, mainly for the experiment. At the cloud habitat conference, Ali obtained the first batch of Hangzhou pilotless road test license. Snow said that the next step will be to refine which sections and routes, as well as the time period to operate these logistics vehicles.

Ali said the focus of the next phase, on the one hand, is to define a new generation of smart logistics vehicles, on the other hand, through cooperation with the Ministry of Transport Highway Institute, to promote road digitization, so that the vehicle-road coordination program as soon as possible landing.

(Studio 8, Liu Yang)


Waonews is a news media from China, with hundreds of translations, rolling updates China News, hoping to get the likes of foreign netizens