Edge AI computers (edge AI computers)

AI computers are devices designed for processing data at the edge of the network (Edge Computing), which is where the data is actually generated, without the need to send it to central specialized servers or clouds. AI computers of this type can be used extensively for intelligent CCTV video analysis, autonomous robots, traffic management systems and many more specialized applications. We offer the widest variety of AI PC units adapted from simple analysis for educational purposes to the most advanced edge processing capabilities.

The Edge AI computer acts as a neural network gas pedal, a specialized unit designed to accelerate the calculations involved in training and inference (inference) of neural networks. The neural network gas pedal is optimized for mathematical operations typical of machine learning algorithms, especially deep learning (deep learning), such as matrix operations and convolutions. It is adapted to perform multiple parallel calculations. The use of such computers is to maximize energy efficiency and be confident in the unit's performance.

Nvidia Jetson - AI computer for edge applications

Nvidia Jetson is a series of computer modules designed by Nvidia that are the perfect platform for Edge AI. The key feature that these computers have is their energy efficiency and ease of use, combined with excellent performance achieved through advanced graphics processing units (GPUs) based on the CUDA architecture. Most AI computers dedicated to digital imaging, are based on the Nvidia Jeston processor which forms the world's leading platform for autonomous machines and other embedded system applications.

Supporting data analytics is ten times better performance than other solutions currently in use. Basing your infrastructure on Nvidia Jetson solutions will eliminate latency and accelerate response to emergencies, reduce bandwidth consumption and thus provide operational savings, increase data security and finally scalability and flexibility. In today's world, the increasing miniaturization and high cost of cloud solutions are making edge solutions the industry standard.

The Nvidia Jetson platform is constantly being developed, and solutions based on Nvidia Jetson Nano base boards with a processor from the SoC (System on Chip) segment can now be used. The Edge AI computer, built on an Nvidia Jetson Nano chip, is the base model of the Jetson series, and there are many more devices with even greater capabilities and much higher performance available from Nvidia itself and Stovaris.

Edge AI computer built using the Nvidia Jetson platform:

• Jetson AGX Xavier
  A family of solutions designed for use in autonomous vehicles. Built on Nvidia Carmel processor with cores based on Nvidia Volta architecture.

• Jetson NX/AGX Orin
  Series adapted for use in mainly automotive industry. . Built on ARM processor and Nvidia Ampera graphics architecture, with CUDA and Tensor cores.

Support for multiple AI frameworks

AI frameworks (artificial intelligence frameworks) are sets of libraries and development tools that facilitate the creation, training and implementation of artificial intelligence (AI) models. They are crucial for Machine Learning (ML) and Deep Learning (DL) engineers, due to the fact that they simplify the process of implementing complex algorithms.

With support for a wide range of AI frameworks, Edge AI computers such as Nvidia Jetson NX Orin, Nvidia Jetson AGX Xavier, Nvidia Jetson AGX Orin, etc. are perfectly suited to work with TensorFlow, PyTorch, Caffe and many others. This makes it easy for developers to port their models and applications to the Jetson platform.

AI PC systems in IoT applications

Edge AI-capable industrial computers are perfect for any IoT application. With their help, we will be able to process a huge amount of data on the environment and obtain the functionality required by our project while fully automating the process.

For example, Nvidia computers from Seeed Studio are designed to work in harsh conditions, so we will be able to use for our projects units standing on, for example, production halls where there is a lot of dust. They will be possible to store data the working temperature will not be a problem thanks to the airflow provided by the design of the case. The units usually have a cooling design without fans.

Seeedis an Elite Partner of the NVIDIA Jetson ecosystem. Seeed's Edge AI platforms provide devices, carrier boards, peripherals, software tools and ML solutions. Seeed is an official reseller and elite hardware partner of NVIDIA. By consolidating best-in-class hardware and cutting-edge technology from software and community partners, our goal is to help both developers and enterprises deploy Machine Learning in the real world across industries.

 AI computer applications at the network edge

• Security systems - Through the use of ip cameras connected to a computer/server equipped with the ability to analyze the image in terms of specific behaviors and situations and facial recognition mechanisms, we will be able to relieve the human eye from the work of analyzing in real time the situation in the space monitored by the cameras. We can, with the help of video streams, check whether the people being watched are smoking a cigarette, running, moving the object being watched, and an infinite number of other applications. All this on the client's fully secure infrastructure, without going to the cloud.

• Machine monitoring - With the help of devices equipped with algorithms that allow us to learn the analyzed parameters, we will be able to build advanced systems that allow continuous monitoring of the operation of machines on the production site. Such devices will collect data from sensors in the machines detecting anomalies and responding in advance to any signs of problems, will allow such application to minimize downtime in production and avoid costly repairs.

• Analyzing customer behavior - Shops, museums, amusement parks, etc. can use devices equipped with AI algorithms to check the behavior of their customers. They can analyze the time they spend at specific displays, what catches their attention and where they go.

• Smart agriculture - Drones or other devices with an AI edge can analyze the condition of crops, detecting plant diseases, irrigation status, etc. This allows for early response and no downtime in crop production.

• Autonomous vehicles - autonomous drones and vehicles can be used for a variety of vertical applications where any autonomous operation of the equipment is required. This is an excellent option wherever you want to relieve the burden of controlling equipment on people. Successfully, these types of machines can work well for sorting packages, storing goods, etc.

• Checking patients' vital signs - devices equipped with artificial intelligence algorithms can be successfully used in medicine to monitor patients' health. Such a device taught the parameters to pay attention to, such as saturation, heart rate, etc., will monitor the patient's health and alert medical personnel in real time to emergency situations.

How to get started with Nvidia Jetson?

For AI beginners who would like to gain critical AI skills and delve into Deep Learning. NVIDIA Deep Learning provides hands-on training and certification. We also encourage you to check out the Seeed wiki guide, covering getting started with Jetson Nano hardware, which is ideal for getting started, as well as creating various projects. Stovaris and developer Seeed Studio provide full support throughout the process of creating custom applications for Nvidia Jetson hardware

This group of products is based on the companies' solutions:

Nvidia | Seeed