Edge Computing (ARTIFICIAL INTELLIGENCE)
Edge Computing as “pushing the frontier of computing applications, data and services away from centralized nodes to the logical extremes of a network.”
Today, most of AI technology still rely on the data center to execute the inference which will increase the risk of real-time application for applications such as traffic monitoring, security CCTV, etc. Therefore, it’s crucial to implement a low latency, real-time edge computing platform.
The advantage of edge computing:
1. Reduce data center loading, transmit less data, reduce network traffic bottlenecks.
2. Real-time applications, the data is analyzed locally, no need long distant data center.
3. Lower costs, no need to implement a server-grade machine to achieve non-complex applications.
– Sorting by human bare-eyes, the standard cannot fix due to a different person. Low accuracy, low productivity.
– Training and sorting by conventional, machine vision method, need a well-trained engineer to teach and modify the inspection criteria. Low latency, but low accuracy if the product has too many variations.
Deep Learning Server
– Implement deep learning methods for training and sorting, reduce, human bias and machine vision training process. High CTO, high latency, power-hungry, high power consumption.
Deep learning Edge
– Implement deep learning methods for training and sorting, reduce human bias and machine vision training process. Low latency, low CTO, power efficiency, low power consumption.
Applications / Surveillance
The Mustang-F100-A10 edge computing device can be utilized to capture data and send traffic to a control center to optimize a traffic light system. It can also perform license plate recognition (LPR) to help law enforcement if vehicles break traffic laws or help parking services. Identify available parking spaces to assist drivers in congested urban areas.
With the algorithms developed using the Mustang- F100-A10 edge device, trained deep neural networks, now have inference capabilities to identify suspicious persons to alert law enforcement or for security departments to early warning scenarios.
Mustang-F100-A10 solutions help enable intelligent factories to be more efficient on work order schedule arrangements. In today’s production line, sticking to manufacturing schedules are becoming more and more important for business efficiency. From raw material storage to fabrication and complete products, all information from the factory such as manufacturing equipment process time and warehouse storage status is essential to achieve production goals. Solution-based AI technology can produce more detailed, accurate, and meaningful digital models of equipment and processes for product management.
Implementing AI into machine vision makes smart-automation applications easier. Previously, factory AOI needed sophisticated engineers to fine-tune inspection parameters such as length, width, diameters and many other specifications that required many adjustments. The Mustang-F100-A10 powered using AI technologies supports workloads such as defect detection and quality control to improve production yield.
Using the Mustang-F100-A10 for computer vision solutions at the edge of retail sites can quickly recognize the gender and age of the customers and provide relevant product information through digital signage. Display to improve product sales and inventory control. Self-checkout can reduce human resource cost so that retail owners can spend more resources on promoting products and understanding business patterns. Also, it can help to analyze customer’s in-store behavior and provide customer information based on gender and age to facilitate product positioning. Quickly converting the business intelligence gained and help build better business practices and increase profitability.
With AI-based technology, healthcare and medical centers can diagnose, locate and identify suspicious areas such as tumors and other abnormalities more quickly and accurately. Using segmentation technology and accurately. Using segmentation technology and trained models on the Mustang-F100-A10 can be used to locate and identify abnormalities with a high degree of accuracy helping doctors and researchers quickly serve the patient.
1. Half-Height, Half-length, Double-slot.
2. Power-efficiency, low-latency.
3. Supported Open VINO™ toolkit, AI edge computing ready device.
4. FPGAs can be optimized for different deep learning tasks.
5. Intel FPGAs supports multiple float-point and inference workloads.
Model Name : Mustang-F100-A10
Main FPGA : Main FPGA
Operating Systems : Ubuntu 16.04.3 LTS 64-bit, CentOS 7.4 64-bit (Support Windows® 10 at the end of 2018 & more OS are coming soon)
Voltage Regulator and Power Supply : Intel® Esperion® Power Solutions
Memory : 8GB onboard DDR4
Data plane Interface : PCI Express x8
Voltage Regulator and Power Supply
Data plane Interface
Ubuntu 16.04.3 LTS 64-bit, CentOS 7.4 64-bit (Support Windows® 10 at the end of 2018 & more OS are coming soon)
Intel® Esperion® Power Solutions
8GB onboard DDR4
PCI Express x8
Power Consumption : < 60W
Operating Temperature : 5°C ~ 60°C (ambient temperature) Cooling Active fan
Dimensions : Standard Half-Height, Half-Length, Double-Slot
Operating Humidity : 5% ~ 90%
Power Connector : *Preserved PCIe 6-pin 12V external power
Dip Switch/LED indicator : Identify the card number
Dip Switch/LED indicator
5°C ~ 60°C (ambient temperature) Cooling Active fan
Standard Half-Height, Half-Length, Double-Slot
5% ~ 90%
*Preserved PCIe 6-pin 12V external power
Identify the card number