With the development of network technology and grid computing, “ubiquitous computing” centered on embedded mobile devices will become a reality. From traditional industrial robots for production and processing to modern entertainment robots that enrich people’s life, they are inseparable from embedded. But the embedded system we often talk about is basically the software and hardware system with microcontroller as the core.
Advantages of embedded
Hardware: low power consumption, good tailoring, high real-time (rich interrupts), small size, rich interface and peripheral resources, some can also carry out parallel computing and so on.
Software: fast operation speed, basically FPU for floating-point operation, real-time operating system (VxWorks μ COS、freeRTOS、 μ Clinux.. at present, there are also some open-source real-time operating systems in China, such as RT thread), GUI, etc.
Embedded system block diagram
Embedded artificial intelligence
Almost all robots and intelligent devices will use embedded devices, such as MCU, arm, FPGA, DSP, ASIC and so on. At present, the fire of embedded seems to be spreading in the Internet and artificial intelligence industries. This trend and current situation are related to the good adaptability of the two fields, and of course, the promotion of arm architecture and various EDA tools. Even at present, embedded artificial intelligence has developed into a new concept in industry and opened up an important branch in the field of AI.
AI technology and algorithms are finally implemented on local embedded devices to realize local real-time environment recognition and perception, human-computer interaction, decision control, etc. Embedded artificial intelligence is a breakthrough away from the cloud and the marginalization of computing. With the increasing AI requirements of mobile devices, many computing will migrate from data center to mobile devices to realize embedded edge computing.
AI plus embedded devices are used in various industries and fields. The most popular image recognition and voice interaction technologies are used in various mobile terminals. The popular concepts and attempts include automatic driving (after all, there is no mature market), virtual reality, etc. the common ones are UAV, multi axis manipulator The application of deep vision recognition equipment and AGV in logistics storage and automatic production industry.
The intelligent warehousing composed of warehousing and logistics robots makes the heavy work in warehousing simple and fast, saves human resources and improves work efficiency.
This paper briefly introduces several common logistics robots
1. Fixed multi axis industrial manipulator robot, combined with depth vision camera to detect materials, carries out mixed destacking operation of roller assembly line.
2. The intelligent handling robot carries out automatic material handling, mainly in the form of backpack, jacking and roller.
3. Forklift AGV is different from traditional manual forklift. Forklift AGV can carry out tasks such as automatic driving and automatic stacking. It can accurately locate and fork goods, and realize machine to machine, machine to ground, ground to ground, stacking pallets and other modes.
4. Intelligent bin picking robot, a highly flexible goods to people robot, is suitable for cargo positions with different heights to automatically complete lifting, grabbing bins and handling, so as to realize warehousing automation.
The hardware of logistics robot is generally composed of several modules: power supply module, motor drive module, sensor module (infrared sensor, ultrasonic sensor, vibration sensor, camera, depth camera, lidar and other sensors), processor, display, speaker module, etc. These modules together form an embedded hardware system, which provides the basis for the realization of various functions for the robot.
Generally, the embedded processor of logistics robot mainly works: one is used to run the robot system, the other is used to collect sensor data, or the two are combined into one, and an embedded chip is used to process the robot system and sensor data.
At present, the performance of embedded chip running robot system may be slightly worse than that of x86 architecture chip, but the general tasks are enough, and even the special embedded AI chip can accelerate the data processing of the system. The more popular robot operating system ROS (a very good robot learning platform) is responsible for the operation of the whole robot task, sensor data reading, slam, navigation, object recognition, speech recognition, robot arm motion planning and Internet connection, etc. (if you have experience or interest in this field, you are welcome to submit your resume to bluecore hr.)
In recent years, the AI industry has developed explosively, and various Internet industries have entered into artificial intelligence physical devices. Ali and Tencent’s industrial Internet applications, Baidu’s Apollo and the driverless industrialization process of car enterprises all point to artificial intelligence.
In 2019, more enterprises will turn from the Internet or other fields to artificial intelligence real economy, promote platform construction, create industrial ecology and accelerate the development of AI industry.
In the future, embedded will get more development in the direction of single core performance and multi-core and multi-threaded, the development of special AI processors will become more and more mature, the processing and computing capacity will be greatly improved, occupy a more important position in the field of human-computer interaction and AI, accelerate the development of AI technology and promote the historical process of AI technology and application.
Source: Intelligent Manufacturing Network