Technical Articles
2025-01-18

Thoughts on the Development of Embedded Technology

Introduction

In the 1970s, the emergence of microprocessors allowed microcomputers to be embedded into an object system, achieving intelligent control of the object system. The computer systems that achieve intelligent control of the object system are called embedded computer systems. Initially, embedded systems were defined as: “application-centered, based on computer technology, software and hardware can be tailored, suitable for dedicated computer systems with strict requirements for functionality, reliability, cost, size, and power consumption.” With the rapid development of microelectronics technology, computer technology is quickly integrated with various industries, making embedded applications increasingly widespread, with various forms of embedded products ubiquitous, and greatly changing the connotation and extension of the embedded system concept. Embedded technology involves microelectronics technology, application technology, computer technology, and software technology, where technologies interact, influence each other, and complement each other. The concept of “ubiquitous computing” continues to make embedded technology a hot topic. This article will elaborate on embedded technology from these four technological perspectives and point out the issues of embedded talent in the country, along with relevant suggestions.

1. Microelectronics Technology is the Foundation of Embedded Systems

The microelectronics industry is a foundational, leading, and strategic industry related to the overall national economy and social development. It is an important indicator of a country’s development level and comprehensive national strength, and it is the core and key to the development of a new generation of information technology. Ultra-high capacity, ultra-small size, ultra-high speed, ultra-high frequency, and ultra-low power consumption are the foundations for solving “deep embedding,” the endless pursuit of information technology, and the eternal driving force behind the rapid development of microelectronics technology and industry.
Looking back at the origin and development of embedded computers, it is clear that microelectronics technology is the foundation for the development of embedded technology. The earliest embedded systems originated after the birth of the Intel 4004 microprocessor, with various manufacturers subsequently launching microprocessors. Embedded systems based on these microprocessors have been widely applied in instruments, medical devices, home appliances, and other industries, forming a vast embedded application market and giving rise to a series of modular and standardized single-board computers centered on embedded processors that are user-friendly.
While single-board computers were being used for embedded applications, microcontrollers were born that integrated basic components such as microprocessors, IO (input/output) interfaces, A/D (analog-to-digital) converters, D/A (digital-to-analog) converters, serial interfaces, RAM (random access memory), and ROM (read-only memory) into a single VLSI (very-large-scale integration) chip. This early microcontroller achieved a degree of miniaturization, low power consumption, and high reliability for embedded applications, marking the initial stage of SoC (system on chip) technology.
Continuous advancements in process technology have significantly increased chip integration levels, enabling the integration of more functions into integrated circuits. As a result, integrated circuits have rapidly advanced to the SoC stage, which brings low power consumption, low cost, miniaturization, intelligence, and high reliability to embedded systems at the chip level, making many previously size-, power-, and weight-restricted embedded applications possible. Thus, the development of SoC technology has further accelerated the upgrade and miniaturization of embedded systems and has determined the universality, application depth, and intelligence levels of embedded systems. With the increase in chip integration levels, SoC chips have become the core of embedded systems.
System-in-package (SiP) technology integrates various IC chips made from different processes, passive components (or passive integrated components), antennas, optical devices, biological devices, and micro-electromechanical systems (MEMS) into a single package to form a micro-system device. SiP technology is an effective way to achieve miniaturization and high reliability for embedded systems based on packaging methods.
In recent years, with the rapid development of AI technology, the complex algorithms driven by application scenarios and the increasingly stringent requirements for “small, low, and light” have posed significant challenges to the integration of Monolithic SoCs (single-chip systems). Scale, integration level, and complexity have increased exponentially. For example, Nvidia’s flagship GPU Volta GV100 released in 2019 has an area of up to 800mm², and the upcoming Ampere series GPU is expected to reach 826mm² using 7nm technology. The AI chip from Silicon Valley startup Cerebras Systems has an area of 46,225mm², with on-chip SRAM reaching 18GB. Such large chips make yield and cost control difficult, and Monolithic SoC chips have reached a dead end in recent years, ushering in an era where SoC may be centered on Chiplet technology.
Chiplet technology breaks down the original large Monolithic SoC single-chip solution into multiple smaller chip combinations, which are then reorganized through advanced packaging. Essentially, it is a 2.5D/3D packaging of SiP; early SiP only satisfied the connection between chips made with different processes, such as heterogeneous integration of CPU/GPU and DRAM (dynamic random access memory), while Chiplet allows different bare chips to be manufactured using different process nodes and even from different suppliers. Third-party Chiplets can significantly reduce design time and manufacturing costs. Although SoCs will remain mainstream for a long time, Chiplet technology enables different components to be designed and implemented on independent bare chips, providing a new idea for solving manufacturability and cost issues in high-complexity, ultra-large-scale heterogeneous systems. Statistics show that the yield rate and chip area size are correlated; for chips smaller than 10mm², the yield difference between monolithic and chiplet solutions is not significant. However, once the chip area exceeds 200mm², the yield of the monolithic solution is more than 20% lower than that of the chiplet solution. It is expected that for areas of 700-800mm², the yield of the monolithic solution may not exceed 10%, while the cost of Chiplet solutions based on mature chips will be much lower than that of monolithic SoC solutions. Chiplet-integrated chips will form a “super” heterogeneous system, providing greater flexibility and new development opportunities for AI computing.
More than Moore’s (超越摩尔) explanation of the two technological paths of SoC/SiP, fully combined with Chiplet and micro-system technology, has provided foundational support for the low power consumption, miniaturization, high reliability, and intelligent development of embedded systems. This has allowed embedded systems to have higher added value and further promoted their leapfrog and widespread development.

2. Applications Drive the Development Direction of Embedded Technology

The endless pursuit of information acquisition, representation, transmission, processing, and usage drives the continuous emergence of hotspots in embedded technology. Each era and different stages within it have distinct characteristics of embedded systems. In the industrial era, instrument control, industrial equipment, and automatic control were the earliest applications of embedded systems; in the information era, home appliances, computers, communication, and networks developed rapidly, and every era is inseparable from embedded technology.
Hot topics of the era, such as virtual reality, big data, cloud computing, the Internet of Things, 5G, blockchain, and artificial intelligence, have led to the emergence of massive applications like live streaming, facial recognition, smart furniture, autonomous driving, and smart cities. A variety of products, including smart phones, multi-purpose drones, smart-assisted cars, and robots, have emerged, and the demand for embedded applications is increasingly rich and diverse. With the rapid implementation of IoT, big data, and AI technologies in the future, embedded systems will penetrate human life more deeply and broadly than ever before.
The continuous expansion and innovation of application scenarios pose more demands on the software and hardware ecology of embedded systems. Early microcontrollers aimed at industrial control quickly gave rise to DSPs (digital signal processors) and GPUs for signal and graphics processing, and in recent years, artificial intelligence has also been making strides. After 2010, with the continuous enrichment of application scenarios and service content, the types of embedded system chips have rapidly increased, and their complexity has risen exponentially. Customized heterogeneous, multi-core embedded SoC system chips have appeared in various application fields such as aircraft, automobiles, mobile phones, and watches. The rapid development and constant segmentation of application scenarios require embedded systems to be more specialized and customized. The gradual implementation of artificial intelligence will further exacerbate the demand for segmentation in application scenarios. Custom processors aimed at specific application scenarios are the future development trend of embedded systems. The increasing complexity of processor functions and the diversification of application scenarios also raise higher demands on the software ecology.
As embedded systems are increasingly applied in high-safety fields such as finance, aviation, automobiles, and nuclear power, higher requirements are placed on the safety, reliability, and trustworthiness of embedded systems. Various industries have produced various software and hardware development specifications, standards, and process control systems, leading to the development of corresponding processors and operating systems. As application complexity continues to rise and the scale of embedded systems expands, design methods that meet safety, reliability, and trustworthiness characteristics still need further exploration. Applications will continue to drive the collaborative and sustainable development of various embedded technologies.

3. Computer Technology is the Core of Embedded Systems

Applications drive the collaborative development of embedded technologies, while different computing architectures and corresponding hardware and software technologies support every stage of embedded computing development. The birth of processors in the 1970s solved control issues, forming microcontrollers centered on CPUs and integrating various IO interfaces, rapidly achieving applications in industrial control and home appliances. The birth of DSPs in the 1980s solved signal processing issues, forming mobile communication control and processing systems centered on CPUs and DSPs, promoting the development of mobile communication devices. Entering the 21st century, the birth of GPUs solved graphics display issues, forming graphics display systems centered on CPUs and GPUs, promoting the widespread application of visual industrial control and electronic instruments. Since 2010, embedded systems centered on CPUs, DSPs, and GPUs have sparked a wave of intelligent mobile communication systems. The birth of GPGPU in 2006 exponentially increased parallel computing capabilities, with Nvidia leading the way in launching embedded systems for autonomous driving and big data processing centered on CPUs and GPGPUs. With the rise of deep learning neural networks, NPU (neural processing unit) was born in 2017, and Huawei was the first to integrate NPU into smartphone SoCs, adding AI elements to embedded systems and greatly enhancing applications such as facial recognition and intelligent photography.
Every generation of computing technology innovation adds new vitality to embedded technology, endowing embedded systems with rich functionalities, powerful performance, and better implementation efficiency, promoting the rapid realization of various applications.

4. Software Technology is the Soul of Embedded Systems

Hardware-software collaboration is a significant feature of embedded systems. With the rapid development of embedded systems, embedded software has also seen tremendous growth. Development languages have evolved from early assembly and C language to now include C++, Python, and JAVA, with a multitude of programming languages expanding the application space of embedded systems while further segmenting expertise, allowing for better exploitation of hardware potential. At the same time, embedded operating systems represented by VxWorks, Android, and embedded Linux have provided strong momentum for the development of embedded systems.
In today’s intelligent era, the development of embedded software has created software ecosystems aimed at various application fields. Ecosystems such as consumer mobile terminals represented by Android, robotics represented by ROS, and autonomous driving represented by Apollo are all centered on embedded software technology, unifying software architecture and user APIs (application programming interfaces), and using hardware abstraction layer technology to build an open hardware support architecture. The emergence of embedded ecosystems not only promotes the orderly development of embedded systems aimed at application fields but also further facilitates the rapid integration of application demands and hardware in industrial development. Throughout the development history of embedded systems’ software and hardware, microelectronics technology provides a robust body for embedded systems, while software technology endows embedded systems with a flexible brain, vitality, and soul.
As embedded systems become increasingly complex, the complexity and scale of software ecosystems have grown exponentially, and the implementation of artificial intelligence has accelerated the increase in software ecosystem complexity. Software engineering, open-source software, and software quality will become focal points for embedded software. Especially in high-safety fields such as aerospace, automobiles, and finance, the design and certification of software systems for safety, reliability, and trustworthiness are particularly important. The country has invested heavily in trusted software, achieving remarkable results in application software development in finance, the Internet, and aerospace. However, in the embedded field, the deep integration of software and hardware makes it difficult to achieve trustworthiness solely through either software or hardware. It requires a combination of software and hardware processing characteristics, mutual cooperation, and collaborative design to jointly build secure, reliable, and trustworthy systems. The collaborative design of trustworthy software and hardware ecosystems will become a research hotspot.

5. Cultivating Embedded Talent is a Domestic Shortcoming

Embedded development is a technology based on knowledge from multiple disciplines, aimed at specific needs, characterized by applications, and is a comprehensive application of various knowledge. Currently, the division of specialties and knowledge transfer in China is overly fragmented, with teaching often emphasizing application software and APP development based on specific software and hardware platforms, remaining at the application level. How to build the system and hardware platforms, as well as the collaborative development of software and hardware, is key and core. The cultivation of these comprehensive capabilities is critical for nurturing innovative talent, but is noticeably insufficient. Systematic thinking, interdisciplinary integration, collaborative software and hardware development, and innovative capabilities are also difficult to cultivate within a single discipline or professional direction. The emphasis on building an embedded discipline system focused on innovative technology and comprehensive application capabilities is in significant discord with the discipline construction and evaluation system guided by SCI and other papers, resulting in a disconnect between university talent cultivation and the urgent need for embedded talent in enterprise innovation and development. It has become common for enterprises to “run universities” to cultivate the required talent, which is also a helpless move.
The embedded industry needs a large number of talents, but more importantly, it needs high-level embedded talents with leading capabilities. Currently, the cultivation of embedded talent in China has serious shortcomings and far from meets the urgent demand for embedded talent at various levels in the industry. How to build a talent cultivation system with embedded characteristics requires the industry and universities to take a moment to seriously think and come up with practical solutions. Schools and research institutions should focus on development needs and “write papers on the land of the motherland.” Enterprises should provide more practical training positions and a broad research platform so that teachers and students can focus on the scientific research practices of enterprises.

6. Conclusion

This article reviews the origin and development of embedded technology, analyzing and summarizing the relationship between technology and embedded systems from four aspects: microelectronics technology, application technology, computer technology, and software technology. Through the evolution of technology, it points out the future development trends of embedded technology and future research focuses. Finally, it analyzes the shortcomings in the cultivation of embedded system talents in the country and suggests that universities and enterprises should base their embedded talent development on their respective needs for collaborative and innovative development.

Author Profile

Thoughts on the Development of Embedded Technology
Tian Ze, PhD, researcher at Xi’an Aviation Computing Technology Research Institute, has long been engaged in SoC design methodology and research and management of dedicated integrated circuit design for the aviation field.
References:
[1]NVIDIA. Nvidia tesla V100 GPU architecture[EB/OL]. [2020-01-30]. https://www.nvidia.com/en-us/data-center/v100/.
[2]Xu Juyan. Returning to the Way – Packaging Transformation in the Chip Industry[J]. Electronics & Packaging, 2019, 10(1): 1-3. XU J Y. Electronics & Packaging, 2019, 10(1): 1-3.
[3]You Zheng. Intelligent Manufacturing and Intelligent Microsystems[J]. China Industry & Information Technology, 2019, 12(8): 50-52. YOU Z. Intelligent manufacturing and intelligent Microsystems[J]. China Industry & Information Technology, 2019, 12(8): 50-52.
[4]NVIDIA W, GENERATION N, COMPUTE C. Whitepaper NVIDIA’s Next Generation CUDA Compute Architecture[J]. ReVision, 2009, 23(6): 1-22. https://doi.org/10.1016/j.immuni.2005.11.006.
[5]He Jifeng. Summary of the major research plan of “basic research of trusted software”[J]. Bulletin of National Natural Science Foundation of China, 2018, 3(9): 291-296.
Citation:
Tian Ze. Some Thoughts on the Development of Embedded Technology[J]. Micro-Nano Electronics and Intelligent Manufacturing, 2020, 2(1): 1-4.
Journal Number: CN10-1594/TN
Supervisory Unit: Beijing Electronics Holdings Co., Ltd.
Organizing Unit: Beijing Institute of Electronic Technology Information Research.
Beijing Fanglue Information Technology Co., Ltd.
Submission Email: [email protected] (Website: www.mneim.org.cn)

*Disclaimer: This article is original by the author. The content of the article reflects the author’s personal views. Lu Ke Verification reprints it only to convey a different viewpoint and does not represent Lu Ke Verification’s endorsement or support of this viewpoint. If there are any objections, please contact Lu Ke Verification.

Related posts

Leave a Comment

Your email address will not be published. Required fields are marked *