TI's open source intelligent hardware platform helps China's graduate electronic design competition start smoothly

In the recently concluded 15th China Graduate Electronic Design Competition National Finals (abbreviated as "Yandian Competition"), a total of 66 participating teams from colleges and universities and research institutes across the country, with the help of the Texas Instruments Sitara series AM5708 Industrial Pie (IndustriPi) open source intelligent hardware development platform and the TI-RSLK Expert Edition with ARM+DSP+GPU heterogeneous multi-core and more powerful "IndustriPi" as the core controller as the basic development tools for the competition, carried out a full-scale competition and created many excellent works. These two hard-core intelligent platforms and tools of Texas Instruments - Industrial Pie and TI-RSLK Expert Edition, help engineering students participating in the Yandian Competition use innovative artificial intelligence technology to achieve advanced system design, which is also the best practice of combining industry, academia and research.

The 15th China Graduate Electronic Design Competition 2020

In this year's research and electronics competition, Texas Instruments ( TI) set up proposition challenges closely related to future industrial development in four major areas, namely, intelligent gateway, target tracking, sound source localization and image recognition, based on "industrial intelligence". Participating students choose to use the Industrial Pie board or TI-RSLK Expert Edition to complete system design based on the interests and advantages of the team, such as a binocular ranging system based on a camera, a sound source localization and target tracking system based on a microphone array, a mobile target detection and target following system based on image processing, or a 4G router gateway system.

TI-RSLK Expert Edition is equipped with ARM+DSP+GPU heterogeneous multi-core and more powerful "industrial school" as the core controller, which can realize high-level practical learning of artificial intelligence majors such as image recognition and processing, sound recognition and processing, edge computing, robot learning, autonomous navigation and path planning based on SLAM, etc. It is an excellent tool to help students have a deeper understanding of the working principle of electronic system design and learn the composition and working mode of robot systems. TI-RSLK Expert Edition has a microphone array, which can realize the sound source positioning and target tracking system; a USB camera module, which can realize the target tracking system based on image analysis; a laser radar, which can realize autonomous navigation and path planning based on SLAM; a WIFI module, which can realize the wireless communication function between the host computer and TI-RSLK Expert Edition. Based on the TIDL deep learning framework + USB camera module + robotic arm, contestants can also realize object recognition and intelligent handling systems based on deep learning. With the help of TI-RSLK Expert Edition, the "Young Pioneers" of Beijing Institute of Technology and the "LW" of Civil Aviation University of China stood out in this research and electrical competition and won the first prize of TI proposition. “Young Pioneers” uses tools such as MATLAB calibration toolbox and visual functions in OpenCV to implement a mobile target detection and tracking system. “LW” uses a redundancy simplification algorithm between multiple MIC receiving data to design a positioning vehicle for sound source positioning and real-time tracking of moving targets, and removes noise reverberation caused by multi-user interference.

TI-RSLK Expert Edition

IndustriPi is a minimal system and open source intelligent hardware development platform based on the AM5708 heterogeneous multi-core processor of TI Sitara series. It is mainly aimed at industrial Internet, intelligent manufacturing, robots, artificial intelligence, edge computing, intelligent human-computer interaction and other application fields. IndustriPi is a basic platform with completely open source software and hardware. It supports 1 Gigabit Ethernet interface, 1 100M industrial Ethernet interface (PRU), CSI high-definition camera interface and other functions. Developers can use it for functional testing, algorithm verification and application development. Due to its rich industrial attributes, it is especially suitable for industrial application fields such as industrial control, industrial communication, industrial human-computer interaction, industrial data acquisition and processing, and real-time control. In addition, IndustriPi supports a rich software development ecosystem, providing support for Processor Software Development Kit (SDK), which can support two versions including Linux and RTOS; support Ubuntu16.04 operating system; support ROS robot operating system; support deep learning architecture-TIDL, implemented through highly optimized CNN/DNN; support Caffe or TensorFlow-slim framework trained models can be imported and converted. The second prize-winning team "Visual02∞" from Shanghai Second Polytechnic University used industrial pie as the core of the entire control and designed a fully functional camera-based binocular ranging system, which received a lot of praise in the competition.

TI Industrial Pi (IndustriPi)

Dr. Wang Chengning, director of TI's global university program, said: "I am very pleased to see that college students have achieved excellent results by using TI's platforms and tools. In the future, we hope to leverage TI's advantages in advanced technology and industrial applications to better promote the integration of industry, academia and research in higher engineering education, help more students understand real industry needs, and help cultivate more professionals who master the world's advanced technologies."

As a world-leading analog and embedded processing semiconductor manufacturer, TI has been working tirelessly to promote education in the field of electronic information technology in Chinese universities. 2020 marks the 24th year that TI's university program has been rooted in China's education. With the support of its rich experience in industry-university cooperation, TI hopes to support the electronic information teaching reform and innovative talent training in Chinese universities through competitions. Currently, TI has established more than 3,000 digital signal processing, analog and microcontroller laboratories in more than 700 universities in China, and more than 300,000 students practice and learn through TI's laboratories and various activities each year.