| Weibu AIBOX-7702, based on Intel artificial Intelligence edge computing technolo |
| Release time:2024-01-05 10:52:22 | Source: | Views： |
The following article is from Haitian Intelligent Internet of Things Laboratory, the author CZH Zhejiang Ocean University Haitian Intelligent Internet of Things Laboratory cooperated with Intel Corporation, the world's largest semiconductor chip manufacturer, to fully contribute to the intelligence of the Marine field, respond to the national "14th Five-Year Plan", actively promote Marine ecological protection, Marine economic development and Marine rights and interests, and accelerate the construction of a Marine power. Laboratory members based on Intel edge computing server to achieve ocean water quality prediction, edge computing, as a rapidly developing new computing model, can provide services near the object or data source, so as to effectively solve the problem of high delay and insufficient bandwidth, help to better improve China's ocean observation monitoring capabilities. Intel Edge AI box-AIBOX7702 With the explosive growth of the current large model, artificial intelligence technology is rapidly migrating to the cloud, and the edge will play a more important role in promoting the deployment of artificial intelligence in the process, so the status of the edge side of the network will continue to improve, but also bring greater market benefits. The growing demand for computing power brings new opportunities and challenges to computing infrastructure.Intel, as one of the industry leaders, adhering to the new infrastructure concept, the use of computing power to serve the traditional energy industry, with a full-stack XPU(CPU,GPU,VPU,FPGA), to provide a variety of platform architectures, with higher performance to adapt to these diverse workloads.
Intel Edge Ai box - Weibu AIBOX 7702 appearance Weibu This AI BOX 7702, is based on Intel artificial intelligence edge computing technology independently developed edge computing box, not only comes with 3-7TOPS computing power, but also supports Intel's official artificial intelligence OpenVINO™ toolkit, and provides I3, I5, I7 three different specifications of CPU configuration. To meet the different price and functional needs of customers, it can quickly deploy applications and solutions that simulate human vision, and bring unlimited possibilities to the design of software service providers! In terms of performance, it is particularly worth mentioning that the Weibu AIBOX 7702 motherboard uses an 8-layer industry difficult pcb stack design, with high-quality passive components, coupled with greatly improved electrical stability, comprehensive safety and service life are guaranteed. In addition, the device has high static immunity, high vibration resistance, high impact resistance and high dust resistance, which can provide stable protection in a variety of harsh environments. In terms of heat dissipation, the whole machine adopts closed fanless design, the shell adopts aluminum alloy extrusion forming, compact structure, the shell doubles as independent heat dissipation; Similarly, the device also completely reconstructs the entire heat dissipation module, extending the length of the copper tube by nearly 300%, increasing the thickness of the aluminum sheet by nearly 200%, and the weight of a single heat dissipation module is up to 1KG, providing higher heat dissipation efficiency and more stable operation support for the stable operation of the software platform.  Internal configuration drawing In Marine monitoring applications, real-time monitoring and early warning of water quality, water quantity and other data can be achieved through the deployment of edge computing equipment. At the same time, edge computing can also analyze and forecast monitoring data to provide more accurate data support for water management. The laboratory uses the following configuration of the edge computing server, which provides real-time monitoring and prediction for Marine water quality management. Marine water quality prediction algorithm Accurate prediction of water quality parameters is very important for effective control of water pollution. However, water quality data are often highly nonlinear and non-stationary, and existing models face great challenges in accurately predicting water quality sequences. The laboratory has developed a novel hybrid prediction method called CEEMDAN-QPSO-BILSTM, which combines three techniques: fully integrated Empirical Pattern decomposition with adaptive noise (CEEMDAN), quantum behavioral particle swarm optimization (QPSO), and bidirectional long and short memory neural network (BiLSTM). The training and testing of the mixed model are all from the real water quality data of the East China Sea of Zhejiang Province, including dissolved oxygen (DO), pH value and chemical oxygen demand (COD) parameters. The test results show that the model can effectively deal with the complex characteristics of water quality data and has high prediction accuracy. At the same time, the experimental results of the HDMI display on the Intel edge computing box -Weibu AIBOX 7702 show that CEEMDAN-QPSO-BiLSTM is superior to the baseline model, which once again verified that the model has excellent error performance and prediction accuracy.
Sum up Haitian Intelligent Internet of Things Laboratory based on Intel edge computing box AIBOX-7702 introduced the Internet of Things technology into the Marine environment, breaking through the limitations of the traditional model, and contributing to the development of Marine Internet of Things technology. The results show that the Marine information processing performance and data computing efficiency based on edge computing have outstanding advantages over the traditional technology. In the future, the Marine Internet of Things system of the laboratory will have more applications to serve smart Marine applications, and the Marine Internet of Things system will also further promote the development of edge computing. |
