What kind of edge AI system is successful?" "Precise perception, fast decision making, human-machine collaboration, high efficiency and energy saving, safe and reliable," said Howard Jiang, Director of Embedded product Systems and Applications, Texas Instruments (TI) China. As we all know, perception, decision-making and execution are the three links of edge artificial intelligence, and with the development of edge artificial intelligence, the requirements for embedded perception and decision-making technology are more stringent and differentiated compared with non-artificial intelligence.
Howard Jiang, Director of Embedded Product Systems and Applications, Texas Instruments (TI) China
Perception -- the data source for edge AI
Data is the foundation of edge AI, and perception is the source of data. Just as a person needs not only eyes to feel the world, but also ears, which are important organs for perceiving the natural world, a machine also needs to be able to see and hear, and various sensors emerge with the development of technology. The single-chip millimeter wave radar introduced by TI can avoid the disadvantages of traditional cameras in many applications, and at the same time support multiple data fusion of the system, so that the machine can better obtain data and achieve accurate perception of the target.
Millimeter wave radar can measure the distance and relative velocity of objects and obstacles in its field of vision with high precision by sending and receiving millimeter wave radar. An important advantage of millimeter-wave sensors over sensors based on vision and lidar is that they are not susceptible to environmental conditions such as rain, dust, smoke, fog or frost. In addition, millimeter-wave sensors can work in complete darkness or in direct sunlight. These sensors can be mounted directly behind a plastic housing without external lenses, vents or sensor surfaces, and are extremely durable and meet the protection Class (IP) 69K standard.
TI's single-chip millimeter-wave radar realizes the advantages of high cost performance that traditional radar does not have through CMOS manufacturing technology. Meanwhile, it combines ASIC back-end processing, which can directly reduce BOM cost, reduce product size, and reduce the dependence on processor. The TI millimeter-wave radar design is one-third the size and half the weight of a miniature lidar rangefinder.
More importantly, in addition to autonomous driving, MILLImeter-wave radar can also be applied to a wider range of industrial and intelligent home, intelligent buildings, medical treatment and other fields. For example, through the combination of millimeter wave radar and air conditioning, it can realize a number of intelligent functions such as wind moving with people, posture perception of the target human body, automatic switch. In other applications, such as mechanical arm for operator's safety monitoring, logistics/unmanned aerial vehicle (uav) obstacle avoidance robot detection, the old man fell, such as monitoring, millimeter wave radar image sensor has ever does not have the advantage of accurate and quick perception, and at the same time meet the requirements of many applications of data desensitization (can be installed in the bedroom, bathroom, etc).
In addition to millimeter wave radar, TI also provides a wide range of products such as temperature sensors, DLP technology, ToF and so on, thus further enriching the way of interaction between machine and human.
Decision making - edge ai brain
Edge AI devices need a smart "brain" to process data and make decisions. Integrated soCs are generally a good choice for edge AI, because in addition to housing various processing components capable of performing deep learning reasoning, soCs integrate many of the components necessary for an entire embedded application. Some integrated soCs include display, graphics, video acceleration, and industrial networking capabilities, making single-chip solutions more than just running ML/AI.
TI's Jacinto™ 7 series processors are just such a highly integrated SoC, with high-performance computing, a deep learning engine, a dedicated accelerator for signal and image processing, and functional safety ASIL-D/SIL-3 standards. In addition to advanced Driver Assistance systems (ADAS), the processor can also be used in robotics, machine vision, radar and other fields.
Integrated dedicated accelerators include the "C7x" new generation DSP kernel, which combines TI's industry-leading DSP and EVE kernel, and adds vector floating-point computing capabilities and support for backward compatible code. With the rise of edge artificial intelligence, DSP, based on Harvard architecture, can significantly improve the efficiency of matrix computing and is very suitable for neural network computing acceleration. Meanwhile, the newly added "MMA" deep learning accelerator can achieve 8 TOPS computing performance at low power under typical operating conditions.
General-purpose cores include the multi-core Arm Cortex-A72, Cortex-R5F, and 8XE GE8430 GPU.
Jacinto 7 series multi-core heterogeneous processor architecture is designed to maximize task-specific selection and optimization for improved performance and cost control. In addition, TI also hardware mature algorithms, coupled with the semiconductor process evolution, so as to achieve the best cost performance and power ratio. TI's ISP, for example, can automatically implement wide dynamic adjustment, image pyramid scaling, stereo depth vision and dense optical flow algorithm acceleration based on embedded hardware acceleration units.
The Jacinto 7 series processors provide a comprehensive security solution involving both hardware and software, which is an important focus for the automotive and industrial markets. The Jacinto 7 series processors were designed for ASIL-D functionality using a hardware development process certified by an independent functional safety evaluation body, such as TV-SD. In response to the new challenges of ADAS data convergence, the Jacinto 7 series also integrates multi-channel ports, such as CSI-2, to ensure interconnection with multi-channel sensors and support high-bandwidth data requirements. The Jacinto 7 Series integrates both PCIe hubs and gigabit Ethernet switches, which can be used by domain controllers to achieve a higher level of integration.
To facilitate user development, TI launched the TI-Edge-AI-Cloud, Cloud tools for AI reasoning on Jacinto processors evaluate and support many industry-common and popular deep learning frameworks (including TensorFlow Lite, onNX Runtime, OpenGL ES, etc.) to help easily compile and deploy models and accelerate reasoning.
In addition to CNN, commonly used for visual recognition, and RNN required for edge AI scenarios such as predictive maintenance, the Jacinto 7 processor also provides support. In addition, TI's SitaraTM family of industrial processors, which integrate the Arm CortexA family of cores, can also be used to implement edge ARTIFICIAL intelligence applications with relatively low computational power requirements, such as predictive maintenance in industrial applications.
Manly Battery which located in Shenzhen,China. A leading Lithium battery manufacturer over 12
years ,pls feel free to send email to info@manlybatteries.com
