UP Xtreme i11 Edge Kits out Intel® DevCup candidates
Intel’s OpenVINO™ Toolkit x UP Xtreme i11 Edge Compute Enabling Kit
Introduction to Intel® DevCup
In recent years, the integration of Artificial Intelligence (AI) into new and exciting applications has been a topic of intrigue. Therefore, to foster an innovative and collaborative AI community, Intel®, along with MakerPRO, invited developers to submit projects aiming to showcase the wide range of industries that can benefit from AI integration.
This was the concept behind the 2021 Intel® DevCup, wherein candidates used Intel’s OpenVINO™ Toolkit to implement their ideas in one of five categories; smart manufacturing, transportation, healthcare, retail, or security. Two of the candidates were Zhi He Technology and Omnix Reality Interactive Studio, both of whom made it to the finals despite there being over 200 other teams competing for the accolade.
Both candidates implemented their concepts using the learning models provided in Intel’s OpenVINO Toolkit, with additional features being developed from a number of different sources. Both candidates also took advantage of the computing power and inference capabilities of the UP Xtreme i11 Edge Compute Enabling Kit. Powered by 11th Gen Intel® Core™ processors, and with Iris® Xe graphics, the UP Xtreme i11 Edge Compute Enabling Kit is the perfect hardware to compliment Intel’s OpenVINO™ Toolkit.
【Application Story I】 AI Intelligent Driving Behavior Analysis
Based in Kaohsiung City, Zhi He Technology submitted their project to illustrate the positive impact that AI could have on the logistics and transport industries. They proposed the "AI Intelligent Driving Behavior Analysis system", a multifaceted application aiming to expand on the growing use of AI in the logistics sphere. Powered by AAEON’s UP Xtreme i11 Edge Compute Enabling Kit, and developed using Intel’s OpenVINO Toolkit, the idea behind the project is to proactively identify and prevent poor driving behavior.
The team’s motivation was to provide a comprehensive system to monitor driver behavior and mitigate risks from both inside and outside the vehicle. To do this, they utilized two cameras. The first camera system, known as the Driver Monitoring System (DMS), was designed to monitor and identify driving behaviors that could produce sub-optimal outcomes, such as smoking, or texting while operating the vehicle.
The second camera system, their Advanced Driver-Assistance System (ADAS), corresponds with the driver’s view of the road, and is programed to monitor the spatial relationship between the driver and other road users. The purpose of the ADAS-powered camera is to monitor risk factors such as driving too closely to the vehicle in front, or not allowing for sufficient stopping time.
By utilizing the UP Xtreme i11 Edge Compute Enabling Kit in their project, Zhi He Technology were able to achieve excellent collaboration between the edge and the cloud, meaning the webcams could capture the images and send them to the Car Networking AI Platform, which were then analyzed by OpenVINO using AI inference. This was possible because of the UP Xtreme i11 Edge Compute Enabling Kit’s Iris® Xe Graphics and its AI module expandability, which provided the device with powerful AI inferencing capability. The implication of this was that using this inferencing capability, driving violations could be monitored automatically via the driving monitoring platform.
In addition to monitoring the actions of drivers, the DMS system has the potential to provide emergency-prevention contingency features. Because the software was programed to note abnormal driving behavior, it would be able to recognize more subtle precursors to adverse outcomes, such as falling asleep while driving. In the event of a driver’s facial expression changing and showing signs of drowsiness, the AI can be trained to alert the driver or fleet management in real time due to the UP Xtreme i11 Edge Compute Enabling Kit’s superior speed of transferring both data and images. This speed is possible because of its powerful CPU and expandability, with 5G and Wi-Fi 6 compatibility giving users faster connectivity when transferring data.
With regard to driver grading and monitoring, the ADAS is equipped to identify abnormal driving behavior, such as following vehicles too closely, sudden lane changes, and not abiding by the traffic light system. This feature could allow fleet managers to address any issues and implement corrective action plans for employees who exhibit these poor habits while driving before an incident occurs.
【Application Story II】 Object Missing Detection
Omnix Reality Interactive Studio also entered the Intel® DevCup intending to impress with their concept, but aimed to address a very different problem. They wanted to illustrate that using a combination of the UP Xtreme i11 Edge Compute Enabling Kit’s sophisticated AI Edge capability with Intel’s OpenVINO Toolkit, they could greatly reduce manufacturing errors on the assembly line.
To address the issue of missing or ill-placed components when manufacturing box-type hand tool sets, Omnix Reality Interactive Studio proposed to develop a system that was able to spot missing or incorrect items, while also displaying a numerical value to represent the result’s accuracy based on the interpretation of high-resolution images of the product.
Given the incredible rate of output needed in DIY and home improvement manufacturing to keep up with demand in an industry worth €701 billion by the end of 2020, even the most incremental improvement to a company’s assembly error rate would enable them to save a substantial amount of money.
Through an open source object detection model using the OpenVINO Toolkit, Ubuntu, and ScaledYOLOv4-csp, powered by the UP Xtreme i11 Edge, the program used a high-resolution (1280x720) network camera to obtain and transfer images to the ScaledYOLOv4-csp open source object detection model. The AI inference, which analyzed the resulting images and determined whether or not an object was missing or incorrectly placed, was done using the AAEON Xtreme i11 Edge Compute Enabling Kit.
The program was able to not only identify which product was missing from the tool set, but also to discern whether the tool in each compartment accurately reflected the correct tool for that compartment. The implications of this are impressive, and could have the added benefit of reducing manufacturing losses that occur as a result of human error, assembly line malfunctions, or theft.
By choosing the AAEON UP Xtreme i11 Edge Compute Enabling Kit for their projects, both teams were able to take advantage of the speed and consistency of the machine, which is built with Intel® Time Coordinated Computing (TCC) and time-sensitive networking (TSN) technology to reduce latency and jitter.
With Intel® Iris® Xe Graphics, the UP Xtreme i11 Edge Compute Enabling Kit provides the most powerful Intel integrated graphics to ensure detailed yet fast imaging, allowing for accurate inferencing. To compound this, the kit also has up to 64GB of memory with in-band ECC support to maximize processing speeds. This is particularly important when considering the time-sensitive nature of industrial manufacturing, in the case of Omnix Reality Interactive Studio.
The UP Xtreme i11 Edge Compute Enabling Kit’s suitability for rugged applications further illustrates why the teams chose AAEON. For example, the UP Xtreme i11 supports both 5G and Wi-Fi 6, allowing users to take advantage of high-speed connectivity and achieve fast data transference with ultra-low latency. Additionally, the fact that Intel® OpenVino AI software is built-in provides users with vision inferencing solutions, meaning they do not need to develop their own AI software.
The success of these projects illustrates that the power-efficiency of the UP Xtreme i11 Edge Compute Enabling Kit allows the user a more reliable and less expensive option in applying AI Edge capabilities to dynamic settings such as logistics fleets. Not only does the UP Xtreme i11 Edge Compute Enabling Kit Support TPM to enable the reliable transmission of large amounts of data, but also has multiple I/O ports; including PCIE, USB, and eDP ports for extension, supporting Wi-Fi and 4G modules, along with LAN and COM functions. It is also suited to this type of visual application as it achieves 4K resolution at 60 Hz, delivering smooth 4K content streaming on 4 displays simultaneously.
The UP Xtreme i11 Edge Compute Enabling Kit provides the perfect base to allow innovative AI concepts to become a reality. The power provided allows for efficient inference and is paired with the Iris® Xe graphics, the most powerful integrated graphics available.
To get started in making your concept a reality, visit UP Shop for more information.
UP Xtreme i11 Edge Compute Enabling Kit
◆ 11th Gen Intel® Core™ U Processors (formerly Tiger Lake)
◆ Intel® Iris® Xe onboard graphics, twice as powerful as previous generation
◆ Clock speeds up to 4.4 GHz with only 28 Watt TDP
◆ 5G Communication for low latency communication
◆ Intel® i225 Ethernet Controller for up to 2.5 Gbps LAN
◆ NVMe storage, USB4.0 Type C, and USB3.2 Gen 2
◆ 40-pin I/O expansion powered by Intel® Altera MAX 5
◆ Multiple M.2 slots and PCIe [x4]
◆ Read to Connect Support for 5G, WiFi 5/6, Bluetooth
Industrial and Secure
◆ Industry standard 12V lockable power input
◆ Features Intel® TCC, TSN, and supports TPM 2.0 and In-Band ECC
◆ COM Ports (RS232/422/485), Phoenix Connector
◆ Compatible with Intel® Distribution of OpenVINO™ Toolkit
◆ Scalable AI with UP AI Core XM2280
◆ Optimize inferences across CPU, GPU and VPU
Production Ready Solution
◆ Production Ready Solution for Robotics, Automation, and other vertical markets
◆ Intel® Distribution of OpenVINO™ Toolkit pre-loaded
◆ Compatible with Intel® OneAPI
◆ Worldwide Shipping in Seven Days