It is as warm and cute as a donkey, reliable and humane, and has a "brute strength" to accomplish its mission. It is not afraid of high temperatures, rain, snow, thunderstorms and lightning, extreme cold and heat.
Article丨AutoR smart driving? Nuoyi
Yesterday we built a rhinoceros, today we build a little donkey. You will never expect what is in Ali’s zoo.
At the Yunqi Conference, Alibaba released the first logistics robot "Little Man Donkey" and also released a robot platform, officially entering the robot track.
This little donkey robot integrates the most cutting-edge artificial intelligence and autonomous driving technology of DAMO Academy. It has human-like cognitive intelligence and its brain emergency response speed is 7 times faster than that of humans.
Alibaba explained that donkeys were once the most commonly used tools for Chinese people to work, travel, go to markets, and carry goods. Nowadays, people’s production and life are becoming increasingly cloud-based and digital, but the core needs of social operation have never been Change, the "Little Man Donkey" who specializes in end-of-line logistics services can be said to be a market tool in the digital native era.
At the same time, it is as cute, reliable and humane as a donkey, and has a "brute strength" to accomplish its mission. It is not afraid of high temperatures, rain, snow, thunderstorms and lightning, extreme cold and heat. In Jiangsu, Zhejiang and Shanghai dialects, "man" is often used as an adverb of degree. Our little barbarian donkey born in Zhejiang is quite smart, quite capable, and quite safe.
It is reported that Alibaba has registered and established Xiaomanlu Intelligent Technology Co., Ltd. to promote the research and development and mass production of robots.
The Little Man Donkey adopts a wheeled structure and a laser radar on its head. It has a cool appearance and is flexible. It can easily handle complex road conditions and intelligently choose the optimal path. When encountering an emergency, its "brain" can react The speed reaches 7 times that of the human brain.
Zhang Jianfeng, president of Alibaba Damo Academy, said, “This robot will soon be used on a large scale in communities, schools, and office parks.?”
It can run 102 kilometers on 4 kilowatt-hours of electricity. , the autonomous driving rate reaches 99.9999
Let’s look at the hardware configuration first.
In terms of appearance size, the size of the robot is 2100*900*1200mm (1445mm tall including lidar).
The appearance of the car body adopts a silver-gray tone, with rounded lines and a cute feeling. The compartment grille can be freely customized. Based on the maximum load of 50 regular-sized express delivery items per car and 10 deliveries per day, the robot’s peak transportation capacity is It can handle up to 500 orders a day, and its performance will not be affected by extreme environments such as thunderstorms, lightning, high temperatures, rain and snow, as well as garages and tunnels.
Considering the safety needs of the end scene, the maximum speed of the robot is set to 20km/h, which means that the speed of this vehicle is consistent with the speed of a normal electric vehicle.
In terms of battery life, the "Little Man Donkey" robot uses a pull-out rechargeable battery, which can be charged 4 degrees per time and has a range of 102 kilometers. For example, the power consumed by the robot when traveling 100 kilometers is less than The electricity used to eat hot pot for two hours.
In terms of power consumption, the robot's power is only 615w, less than half of the Dyson hair dryer (1600w), and less than one-third of the conventional household induction cooker (about 2000w).
In terms of sensors, it is equipped with RoboSense customized lidar, surround-view cameras, millimeter-wave radar, inertial navigation and other sensors, which are similar to the current most advanced driverless car sensor solutions.
With powerful hardware configuration, the next step is DAMO Academy’s cutting-edge artificial intelligence and autonomous driving technology.
The first is intelligence, with human-like cognitive decision-making capabilities.
The "Little Man Donkey" robot can drive freely in complex terminal scenes, avoid obstacles reliably, and smoothly handle turns, emergency stops, meeting vehicles, reversing, etc., with an autonomous driving rate of 99.9999.
It only takes 0.01 seconds for the robot to identify the intentions of hundreds of pedestrians and vehicles. When it encounters danger and needs to make an emergency stop, it only takes 0.1s for the brain to complete decision-making, planning and issue control instructions.
The second is safety, with a five-layer safety design.
In order to ensure the safety and stability of the vehicle, the robot's system architecture introduces a five-fold redundant design, including brain decision-making, redundant cerebellum, abnormality detection brake, contact protection brake, and remote protection.
Once again, it can be mass-produced and has large-scale mass production capabilities.
Through self-research and in-depth customization of algorithms, the manufacturing cost of robots has been reduced to a level that can be mass-produced. This includes independent design of appearance, possession of design patents, and self-developed embedded heterogeneous computing platforms with high performance. , low cost, low power consumption and other characteristics.
The "Little Man Donkey" is just the beginning. In the future, he will go to high buildings and down to the basement.
The "Little Man Donkey" is just the beginning.
Alibaba Damo Academy also launched an intelligent robot platform simultaneously, opening general robot technology and capability modules to the outside world. Users can quickly develop various types of robots in different scenarios on this platform.
The Autonomous Driving Laboratory of DAMO Academy has accumulated the general capabilities of mobile robots (hardware solutions, software algorithms, terminal-cloud integrated service capabilities) into the intelligent robot platform.
The underlying technologies that support these capabilities mainly include multi-modal artificial intelligence technology and autonomous driving technology above L4.
Therefore, based on this platform, mobile robots for multiple scenarios can be developed as quickly as assembling Lego, such as security patrol robots, airport service robots, epidemic prevention and disinfection robots, and scenic area guide robots.
In addition to wheeled robots, Alibaba revealed that in the future, it will also develop different types of robot products such as bipedal and crawler robots, allowing robots to climb mountains and wade in rivers, chat and accompany them, and go up high buildings and down to basements.
Why is Damo Academy’s logistics robot so smart?
First of all, at the algorithm level, DAMO Academy proposed a "small front desk, large middle desk" algorithm architecture, which greatly improves the effect of the algorithm model and the efficiency of algorithm iteration.
DAMO Academy has developed the industry's unique autonomous driving machine learning platform AutoDrive, which uses machines to replace manual labor for algorithm parameter adjustment and model optimization, improving algorithm development efficiency.
Based on the support of the AutoDrive platform, "small front desks" such as perception, positioning, and decision-making planning continue to propose lighter and faster algorithm models.
In terms of positioning, the robot can achieve centimeter-level high-precision positioning in environments without or with weak GPS.
In terms of perception, the robot can identify "centimeter-level" obstacles with the help of the perception algorithm self-developed by DAMO Academy.
Secondly, in terms of computing power, DAMO Academy is backed by Alibaba Cloud and has large-scale cloud computing resources, large-scale computing cluster engineering operation and maintenance, large-scale cloud data storage and other resource support.
Based on the AutoDrive platform, DAMO Academy has built its own autonomous driving cloud platform to move massive data (scenario database, autonomous vehicle data, data collection vehicle data) to Alibaba Cloud, opening up data collection , data annotation, simulation, model training, evaluation and other complete systems.
Finally, at the level of software and hardware collaborative design, the logistics robot’s embedded computing platform, sensors, positioning unit and other hardware all adopt an integrated design of software and hardware.
DAMO Academy has independently developed a high-performance, low-power, low-cost embedded heterogeneous computing unit, which can achieve the same level of intelligence with 1/3 the computing power.
Through software and hardware collaborative optimization, the power consumption of the computing unit is reduced by 72%, the cost is reduced by 50%, and the volume is compressed by 62%. In terms of sensors, DAMO Academy has proposed a multi-level and multi-sensor fusion technical solution, including lidar, cameras, millimeter wave radar, inertial navigation and other sensors.
In the era of human-machine emergence, there is a competition for a trillion-dollar market
Why did you first choose to develop robots for logistics scenarios?
Alibaba gave three explanations:
First, society has needs. In the digital native era, the demand for logistics and distribution is exploding rapidly. It is expected that in the near future, China will generate 1 billion delivery orders.
Second, the scenario requires it. In the entire logistics system, terminal logistics has always been the most costly and least efficient link. We hope to use technology to provide terminal logistics solutions.
Third, technology has been accumulated. DAMO Academy and the entire Alibaba economy have explored multi-modal artificial intelligence technology and autonomous driving technology for many years, providing a profound technical foundation for the research and development of robots.
It can be said that with the development of new generation information technologies such as big data, cloud computing, and 5G, robotics technology will evolve rapidly, software and hardware technologies will become increasingly mature, costs will continue to decrease, and performance will continue to improve, eventually triggering The advent of the robot age.
Of course, in terms of terminal logistics and distribution, Alibaba is not the only one. Prior to this, companies such as Zhixingzhe, Neolithic, JD.com, and Suning have all launched low-speed unmanned logistics and distribution vehicles.
According to the prediction of the International Federation of Robotics, the robot revolution will create a market worth trillions of dollars, and China will be the largest robot market in the world. An era of human-machine union is coming, where humans and robots The collaborative interaction has just begun.
This article comes from the author of Autohome Chejiahao and does not represent the views and positions of Autohome.