HSINCHU SCIENCE PARK, Taiwan, Aug. 22, 2022 /PRNewswire/ -- With the rapid development of semiconductor technology, the Internet of Things (IoT) is increasingly becoming integrated with artificial intelligence (AI). In such context, the IoT- enabled equipment automation, fast data acquisition and remote control is working closely with AI to produce what is the so-called AI+IoT technology. The term of AIoT is literally the best combination of the IoT with artificial intelligence system. At present, such technology is being extensively used in multiple sectors from the industrial applications to people's daily lives, with a view to providing more innovative solutions to various industries. With the AIoT technology, the deep learning is possible by devices collecting and analyzing a large amount of external data. The multi-level neural network training systems and outstanding computing power it boasts is helpful to behavior forecast and decision-making like human. All this, however, requires a large number of internal memories and CPU resources. The traditional clouding service mode takes huge cost as it has to port data to the cloud side for processing. On the other hand, the whole process is not free from data transmission latency and data security risk. The birth and growth of AIoT is aimed at providing better user experience in terms of security and convenience.
As the digital economy grows rapidly, many enterprises are increasingly concerned about delay in data transmission, security and reliability of information. In response, the use of edge computing solution is on the rise. Unlike traditional computing mode, the edge computing is designed to make computing as close to data sources as possible to reduce latency and bandwidth usage, thus offering better transmission speed, lower operation costs and enhanced reliability and security. In the course of digital transformation, the MCU has played a big role in achieving Edge AI or Endpoint AI thanks to its various advantages such as low-power consumption, fast and low-cost development. We can see more and more MCUs engaged in the embedded machine learning (Embedded ML) and tiny machine learning (Tiny ML). The MCUs are equipped with varying performance to cater to different level of AI computation requirements.
The MCU-based AI system focuses on real-time decision-making and response speed because of its strengths in low-power consumption, low latency, low cost development and higher security. With the help of digital signal processor (DSP) and machine learning (ML), the system can be used for the classification, identification, prediction and inference judgment. It is found that many applications like sensor detection, motor vibration analysis and voice recognition, to name a few, are being widely used in industrial control, motor control and consumer electronics, among others. Besides, there are more high-end MCUs offering specific solutions to the complex computer vision and imaging applications, such as fingerprint analysis, facial recognition and collaborative robot. The continuous upgrading in AI algorithms has driven the MCU industry to expand its functions to adapt to the rapid growth of AIoT.
With the growing demand of the device networking, the sector sees fast development, and a large number of human-machine interactions and high-speed communications keeps expanding, causing systems to be much more complex than before. Thus we are gradually turning to the artificial intelligence technology to attain a high level of smart management in this regard.
As an innovative leader in the 32-bit general-purpose MCU sector, ARTERY Tech offers up to 12 product lines covering low power line, value line, mainstream line, wireless BLE and high performance line, a total of nearly 200 parts. In addition to its 55nm advanced process, ARTERY MCU family all operate on ARM®-Cortex®-M4 or M0+ core and are equipped with a complete set of development tool kits, thus speeding up development and shortening production cycles. When it comes to the AIoT applications, AT32 MCUs developed by ARTERY stand out in four respects:
AT32 MCU Family
High performance reduces latency and increases efficiency
The low latency mechanism makes real-time control and faster response possible. The mainstream line AT32F403A/407 and high-performance line AT32F435/437 are capable of offering low-latency solutions as they provide a maximum frequency of 240MHz~288MHz, the first of its kind in the MCU field, and embed floating point unit (FPU), 256~4032KB Flash and 96~512KB SRAM, in addition to rich peripherals, XMC memory extension function and multi-channel 12-bit high-speed ADC and DAC. All these features can help greatly increase development efficiency and lower costs. Among AT32 MCUs, the AT32F435/437 series is the most competitive as its CPU operate up to 360DMIPS, and its CoreMark up to 1002.74 points (3.482 CoreMark/MHz). On the other hand, ensuring stability in communication transfer is regarded as a key feature to the smooth operation of the whole IoT system. The AT32F407 and AT32F437 series are compatible with IEEE-802.3 10/100Mbps Ethernet controller. The AT32F435/437 series, in particular, embeds an independent 5.33Msps ADC engine to guarantee fast and stable transmission, thereby meeting the demand of motion control and intelligent control.
High accuracy and reliability
The accuracy of measurement is often related to the use of sensors or IoT nodes. The sensor can be used for wireless communication and data processing. By using AI algorithms that give reasoning and behavior prediction, the sensor make its own judgment. And this is what the MCUs are working on. For example, the AT32F415 series operates up to 150MHz and supports a wide range of communication interfaces and peripherals including UART, SPI, I2C, SDIO, USB OTG and CAN. At present, such device has been applied in smart home area as a partner of Amazon solutions. Following the Amazon Connect Kit (ACK) protocol, the AT32F415 series provides voice control and smart dimmer light switch control system through Amazon Echo or APP speaker in Wi-Fi environment. Speaking of the AT32F425 series, it features USB OTG, up to 96MHz, 120DMIPS CPU, and CoreMark up to 326.57 (3.402 CoreMark/MHz). The device embeds a maximum of 68KB Flash and 20KB SRAM, OTG controller, CAN, IR timer, four USARTs and 2Msps 12-bit ADC. Therefore the AT32F425 series is capable of achieving high-speed data acquisition, mixed-signal processing at shortened development cycles and lower costs while ensuring accuracy and data integrity.
Low power and fast wake-up capability
For a terminal device, it is required to perform fast data collection and transfer in a short period of time. A low-power MCU is designed to better manage memory resources and communication time while performing edge computing. In most terminal devices, batteries serve as the primary power source. To extend the power supply, ARTERY AT32 MCU series supports three sleep modes: Deepsleep, Standby, and Sleep. Additionally, multiple WKUP pins are available to wake up from Standby mode so as to save power consumption as best as it can.
The IoT leaves its footprint in almost all corners of society from households to urban network. A smooth wireless communication is regarded as the most important link in the entire IoT system. For the signal collection devices scattered in various venues, it is not viable for workers to replace batteries for them frequently. Thus a low-power MCU is needed to ensure these devices to operate for a long period of time. In this case, the AT32WB415 wireless series is highly recommended as it adopts BLE 5.0 technology, a maximum frequency of up to 150MHz, 187.5DMIPS CPU, and 400.84-point CoreMark (2.672 CoreMark/MHz). The device features rich communication interfaces and memory resources, and embeds RF transceiver and baseband. Besides, its Bluetooth RX value can be up to -97 dBm, and transmit power (TX) between -20 dBm and +4 dBm. The antenna embedded in the device can cover as far as 30m, up to 2Mbps, for full connectivity. Moreover, the AT32WB415 enjoys a competitive price which makes it a great choice for networks that are cost-sensitive.
Improve data security and privacy
It is important to ensure user privacy and data security when terminal devices are connected to the Internet because a large amount of sensitive data are generated. Therefore, it is essential for MCUs to provide strong protection for data storage, which is often achieved by hardware access control mechanism in data storage area, strict programming procedures and access protection configuration. The security Library (sLib), developed by Artery, allows users to set any part of the Flash memory as a security area that is code-executable only but non-readable. The "sLib" is a mechanism designed to protect the intelligence of solution vendors and facilitate the second-level development by customers.
In the meantime, Artery has established a complete set of ecosystems, including development tools, real-time operating system (RTOS), and third-party software resources compliant with IEC-60730 by the International Electrotechnical Commission (IEC). These resources are useful for the rapid development and launch of products.
Looking ahead, more networking devices are needed to meet diverse demands from more applications, in particular the IIoT, consumer electronics and wireless products. In order to support complex computing requirements and diversified communications in the future, we should do more to enhance 32-bit MCUs performance in computation, data transfer and data processing to keep abreast of the time.
At present, ARTERY is working on a full-fledged MCU platform ecosystem. Such platform focuses on MCUs based on ARM®-Cortex®-M4/M0+ core, which incorporate digital signal processor (DSP), single-precision floating-point arithmetic (FPU), high-speed CPU performance, larger memories, rich peripherals and self-developed security Library (sLib). And there are various packages available for users to select the desired MCU product. At the same time, ARTERU is speeding up its efforts to move toward 28nm/40nm process, strengthen research and development of high performance, lower consumption and offer diverse packages, aiming to bring about better user experience.
Source: Artery Technology
- Fibocom Launches AI Smart Module SCA825-W, Unleashing the Potential of AIoT with Superior Computing Power
- Axiomtek Unveils AIE900-XNX Fanless AI-Powered Box PC Using NVIDIA Jetson Edge AI Platform for 5G and AIoT Applications
- Milesight Blossoms into Mighty AIoT Solution Provider with Leading-Edge Products and Solutions
- MediaTek Unveils New AIoT Platform Stack and Introduces the Genio 1200 AIoT Chip
- Three Vietnamese papers accepted at the International Conference on Machine Learning
- Three Vietnamese research works accepted at International Conference on Machine Learning
- LITEON Technology Reports the First Half Net profits of NT$6.1 Billion, EPS of NT$2.66
- CES 2022: KaiKuTeK reveals world's first single system-on-chip contactless 3D gesture recognition system
- Innodisk Announces Its New Business Focus on the Edge AI Computing Market