How is Baidu layout unmanned?

The automotive industry is undergoing a major transformation. There are three very big trends in this transformation that are driving this industry change: the arrival of new energy, shared economy and intelligence.

In July last year, I began to visit domestic and international large and small car factories. They told us very firmly that they would take the road of assisted driving and gradually develop into the future of unmanned driving. At that time, they thought that no one was driving at least ten. A few years will come.

In January of this year, many big and small car manufacturers, many colleagues in the automotive industry began to change their views. Today, in less than a month, as far as I know, there are already a lot of big car factories that have changed their attitudes. In addition to developing ADAS (Assisted Driving) technology, they began to set up independent teams to develop fully automated driverless.

From assisted driving to unmanned driving is a huge leap. It is a fundamental change from quantitative change to qualitative change. There is a huge gap here. Two reasons are: one is the sensor used, the second one is artificial intelligence. Technology is fundamentally different.

In the eyes of the Internet industry, the future car is a computer with wheels. At the beginning of this year, Forbes magazine published a research report based on Stanley. They also believe that 60% of the value of future cars comes from software. Its hardware, which is familiar to everyone, is only 40% valuable.

Of course, this is a matter of opinion.

The future car is defined by software. Its artificial intelligence is controlled by six large parts: visual, listening, and speaking to simulate human eyes, mouth and ears respectively; predictive ability, planning ability, and action control ability are determined together. The development of artificial intelligence.

Unmanned driving based on "deep learning"

In fact, the development of artificial intelligence has been more than 60 years, and the past 50 years have not been successful. Until a decade ago, in 2006, the computer industry had a breakthrough technology development called “deep learning”. Compared with the past, the fundamental part of "deep learning" is to give it a lot of data, and it can continuously improve its own capabilities.

What is the ability of artificial intelligence to develop, what is the most critical thing for deep learning?

The first is an excellent algorithm. The key factor is the level of excellence of the talent, which tests the level of the engineer.

The second key element is the scale of cloud computing. Three years ago, Baidu built a relatively small-scale cloud computing with 48 servers, which was equivalent to the computing power of two Tianhe No. 1 supercomputers in China.

The third key element is knowledge, that is, massive data. When the computing power is very strong and the algorithm is very good, it depends on the data, who has more data, and whose artificial intelligence will be more intelligent.

Today, driving a driverless car out, every new situation it encounters is learning new abilities, not learning in the car, uploading new situations to the cloud, to the data center. There are tons of information and scenes here, and these cars learn how to deal with new situations and new scenes in the data center.

As the cloud capabilities intensify, it learns to handle a wide variety of scenarios. It simplifies and shrinks this data and the resulting models and downloads them into every car. Every car knows how to deal with any scene it knows in the past. When it encounters a scene that you don't know, it will be uploaded to the cloud computing center. After such a cycle, the artificial intelligence capability will continue to increase and continue to increase with the scale.

Today is how people are teaching how to drive, and the terrible thing about artificial intelligence is that it is a car that teaches cars to drive. If 100,000 unmanned cars are laid out and 20,000 kilometers per car per year, the car will accumulate 2 billion kilometers per year. For artificial intelligence, any new experience learned by any car, new processing methods, in the cloud can form a comprehensive model, comprehensive ability, and then downloaded to all cars, all cars have more than 2 billion overnight Kilometers of driving experience and ability, this is a terrible place for artificial intelligence.

Only with such mileage, talent, and computing power can artificial intelligence unmanned cars cross this gap.

Baidu in unmanned layout

In the past, the big fish used to eat small fish and small fish to eat shrimp, but in artificial intelligence, fast fish eat slow fish. Whoever first lays out artificial intelligence cars and driverless cars will be in a decisive lead. So what does Baidu do in it?

Baidu has the world's largest "deep learning" network, and the most fundamental indicator supporting the order of magnitude is the trillion-level parameter;

Baidu is based on "deep learning" as an artificial intelligence product, such as integrated speech recognition. In a quiet environment, in the standard Mandarin case, our human listening ability is 98% accurate, and Baidu can achieve 97% accuracy. In the vehicle environment, the general recognition rate reaches 92%;

The accuracy of image recognition capability reached 89.32%, ranking first in the world. This is an important technical basis and accumulation of unmanned vehicles on the road;

Baidu's high-precision map and status ability, GPS positioning error is about 5 meters, Baidu's map positioning accuracy and accuracy last year is about 10 cm.

Three days ago, we signed a demonstration zone with the Wuhu City of Anhui Province, a fully unmanned demonstration zone. We hope to transform Wuhu City into the largest all-wheelless zone in China through three stages.

In addition, the success of China's unmanned vehicles requires the government's support in laws and regulations, and we need to work together in the entire industry chain. Baidu is very willing to take our leading technology out to create a powerful unmanned vehicle ecosystem.

li ion battery cell output voltage is 4.2v highest and 3v cut off, normally called 3.6v/3.7v batttery. 
3.6V-3.7V Li-Ion/Polymer Battery Pack 
3.6V-3.7V 2000~5000mAh Battery Pack
3.6V-3.7V 5000-10Ah Battery Pack
3.6V-3.7V  10Ah-50Ah Battery Pack
3.6V-3.7V 50-100 Ah Battery Pack 
li ion vs li po, 3v Battery, aa lithium batteries, 3.7 volt battery, lithium ion vs lithium polymer, 3.6 volt battery
3.6 v aa lithium battery,li ion cell, high voltage battery, volt battery, lithium ion aa battery, 14250 battery,
er26500, 2aa battery, 3.6v battery, ls14250 battery, cell voltage, lithium ion vs polymer, 3.6 volt aa lithium battery,
3.6 v Lithium Ion Battery, 3.7 volt lithium ion battery, 3.7 battery, 3.6 volt lithium battery, 3 volt rechargeable battery,
3.6 volt rechargeable battery, lithium d batteries, Lipo Battery voltage, 3.6 volt aa battery, 3.6 volt pil,
3.2 volt battery, 3.2 volt rechargeable battery, 3.7 volt lithium battery, 3.6 lithium battery,3.2 volt solar batteries.

3v Battery

3v Battery,3v Lithium Battery,3V Rechargeable Battery,3V Polymer Battery

Asarke Industry Co., Limited , https://www.asarke-industry.com