Smarter than humans? See how AI enables smart travel

1 Popular Science About AI

Speaking of AI (artificial intelligence), it is a very popular concept nowadays. Many people think that it is still in the realm of science fiction and too far away from our current lives - but in fact, AI is already changing our lives unknowingly. At the Baidu AI Developer Conference held on the 3rd of this month, Baidu released its latest achievements in the field of AI. Let's take a look at how these intelligent machines will change our travel in the future.

What is AI?

AI is a term in computer science. In short, it is the process of using algorithms and calculations to enable machines to have human-like perception capabilities, so that they can self-learn and proactively solve problems. From a technical perspective, AI mainly includes the machine's recognition, analysis, learning and decision-making capabilities, and behind this, algorithms containing hundreds of thousands or even millions of codes are the core of realizing these functions.

We know that one of the great advantages of computers over humans is that they can tirelessly perform large-scale, long-term repetitive calculations and analyses, and can also guarantee absolute reliability. However, if computers are to analyze information that cannot be simply quantified, such as images and sounds, they need to have deep learning capabilities, and through perceiving and analyzing information features, they can "learn" the methodology of analyzing problems, thereby realizing the function of automatically processing and analyzing information in large quantities.

However, if AI is to be widely used through deep learning, higher requirements will be placed on the computing power of computers. According to statistics, from 2012 to 2019, the computing speed of computer chips increased by 30 times, which seems amazing, but today's AI algorithms have increased the computing power requirements of computers by 300,000 times. How to balance the huge gap between the two requires optimizing algorithms and computing architecture.

Today, AI has been widely used in all walks of life, including medicine, education, finance, military, and video games. Even the smartphones we use every day have AI. Let's take a look at what magical effects will occur when AI is combined with transportation and automobiles.

2 How does AI change our travel life?

At the conference, Baidu first demonstrated how the traffic efficiency of Baoding City was greatly improved through the deployment of intelligent network optimization of traffic infrastructure. We know that although the traffic lights at the intersection can ensure traffic rules and safety, the traffic flow of two-way roads varies significantly during the day because the traffic light time is fixed, which causes congestion on one side of the road and empty traffic on the other side, affecting the traffic efficiency of the road.

After mastering the data, AI can infer the future traffic flow of major roads through analysis and calculation, and combine it with the big data of previous peers. By controlling the timing of traffic lights at intersections, the road flow can be optimized to improve peer efficiency.

Another major application area of ​​AI is autonomous driving. Obviously, when a car is equipped with various cameras and radars, it has a wider and clearer field of vision than human drivers. Combined with deep learning capabilities, today's autonomous driving functions can already achieve automatic vehicle control in real road environments. At the conference, Baidu announced that it had won the T4 (i.e., in an open road environment) autonomous driving road test license, which has the highest technical level, the highest standard, and the most difficult test scenario in China. At the same time, the first batch of passenger cars equipped with L4 autonomous driving technology jointly built by Baidu Apollo and FAW Hongqi officially rolled off the production line in a car workshop in Changchun, China. The first batch of vehicles will be shipped to Changsha for testing. By the end of 2019, ordinary citizens will be able to take driverless taxis in Changsha.

Of course, considering the current road environment, infrastructure construction, laws and regulations, it will take some time for L4 or even higher levels of autonomous driving to become popular. However, the Valet Parking technology released by Baidu, as a branch of autonomous driving technology, is about to start trial operation.

In today's driving environment in China, the "last mile" of most car owners often becomes their driving pain point: finding a parking lot near the destination has become a time-consuming and laborious task. According to statistics, there is currently a shortage of 50 million parking spaces in China, while the effective utilization rate of parking spaces is less than 50%. Car owners either have a shortage of parking spaces near their destinations, or the vacant parking spaces are too far from their destinations.

Baidu's Valet Parking technology can be used in home or office scenarios. Through simple learning and training, the system can enable the vehicle to park along a fixed route within sight distance after the driver gets off the vehicle. In popular locations such as shopping malls and hospitals, the vehicle can automatically drive into the parking lot and find a parking space on its own. In transportation hubs such as airports that are part of the smart city, the driver can get off directly at the entrance, and the vehicle will automatically drive along a fixed route to a nearby low-heat parking lot and park automatically.

In addition, another typical application of AI in automobiles is the smart cockpit.

In this context, Baidu has developed the Xiaodu Car OS underlying operating system for in-vehicle interconnection. Speaking of the biggest feature of Xiaodu Car OS, it is also AI. We have previously conducted a professional evaluation of the smart car connection system of Xingtu-TXL equipped with the Xiaodu Car OS system ( portal ), which also confirmed that the functions and user experience of this system are indeed convincing. At this conference, Baidu released Xiaodu Car OS 2019, and its functions have been further evolved.

Summary:

Unexpectedly, Baidu AI Developer Conference became popular on the Internet through "a bottle of water". But returning to the technology itself, we should see that the era of intelligent cars has arrived. There are many third-party suppliers in China that have provided mature products for the functions mentioned above, but there are few companies like Baidu that have come up with a complete solution based on its own AI technology. As more and more vehicle manufacturers join Baidu's "AI ecosystem", I believe that Baidu's AI label will become more and more popular.