Home
Industry News
Solutions
About Us
| EN

Zynq UltraScale+ Edge Computing Board Industry Solution

1. 5G/ORAN Distributed Unit (DU) Solution

Customer Pain Points

In 5G network deployment, distributed units (DUs) need to process massive amounts of data, requiring extremely high computing performance, data transmission real-time capability, and stability. Traditional solutions often face problems such as insufficient processing power, high transmission latency, and difficulty meeting the needs of complex network environments.

Solution Design

Using the Zynq UltraScale+ Edge Computing Board as the core processing unit, equipped with the Xilinx Zynq UltraScale+ MPSoC ZU9EG chip, combined with 600K LUT and 2562 DSP Slices, and providing about 5TOPS INT8 computing power after loading Vitis AI DPU, it can efficiently process 5G network data. 4×SFP+ 10GbE and 1×PCIe Gen2 x4 Root Port and other high-speed I/O interfaces achieve 100MHz bandwidth, 4T4R configuration, 25% margin in a 100µs cycle, and front-haul link latency less than 5µs, ensuring fast and stable data transmission.

Application Effect

Greatly improves 5G network data processing efficiency and transmission speed, meets the needs of low latency and high bandwidth in 5G networks, ensures network service quality, and helps operators deploy 5G networks quickly and stably.

2. 8K/4K Real-time Video Processing and Intelligent Transcoding Solution

Customer Pain Points

With the continuous improvement of video resolution, 8K/4K video processing faces problems such as low codec efficiency, high transmission latency, and difficulty in image quality optimization, making it difficult to achieve real-time, high-quality video playback and distribution.

Solution Design

Utilizing the multimedia processing capability of the Zynq UltraScale+ Edge Computing Board, video sources are accessed through 2×HDMI and 1×DP interfaces, combined with Xilinx VCU + DPU Pipeline, to achieve the complete process of encoding/decoding → AI super-resolution/denoising → packaging. It can support one channel of 4K60 or two channels of 4K30 video processing. With the OBS/FFmpeg plugin sample project, it achieves efficient processing with end-to-end latency of 80ms. The onboard 4GB DDR4-2400 (PS side) and 512MB DDR4-2133 (PL side) ensure data storage and fast reading.

Application Effect

Achieves real-time intelligent processing and transcoding of 8K/4K video, reduces video transmission latency, improves image quality, suitable for online video platforms, live broadcast industry, etc., bringing users a smooth and high-definition video experience.

3. Robot and Autonomous Driving Sensor Fusion Solution

Customer Pain Points

The fields of robotics and autonomous driving require real-time fusion of multiple sensor data, with strict requirements for data processing speed, synchronization, and accuracy. Traditional solutions are difficult to meet the efficient operation needs in complex scenarios.

Solution Design

The Zynq UltraScale+ Edge Computing Board is applied to the sensor fusion stage. 2×MIPI cameras and 10GbE LiDAR data can directly enter the PL, and full hardware timestamp and synchronization technology are used to ensure precise data alignment. The powerful computing power supports DPU running the YOLOv8-tiny model, achieving 120fps at 960×544 resolution, with remaining computing power used for VIO/SLAM DSP cores, enabling fast processing and analysis of multi-sensor data.

Application Effect

Improves the perception and decision-making speed of robots and autonomous driving systems for the environment, ensures the safety and stability of the system in complex scenarios, and promotes the development and application of robotics and autonomous driving technology.

4. Industrial Vision and High-speed Data Acquisition Solution

Customer Pain Points

In industrial production, industrial vision inspection and high-speed data acquisition require high-resolution image acquisition, fast data processing, and stable storage. Traditional equipment has problems such as slow acquisition speed, data loss, and low processing efficiency, affecting production quality and efficiency.

Solution Design

Through the 2×FMC HPC interface expansion of the Zynq UltraScale+ Edge Computing Board, Camera Link, CoaXPress, or 16-lane ADC can be connected to achieve high-speed image and data acquisition. PL DDR4 ring buffer combined with PCIe DMA technology continuously writes to NVMe at a speed of 2GB/s, ensuring no frame loss of data, and at the same time, its powerful computing power is used for real-time analysis and processing of acquired data.

Application Effect

Improves the accuracy and efficiency of industrial vision inspection, achieves stable acquisition and processing of high-speed data, helps the automation and intelligent upgrade of industrial production, reduces production costs, and improves product quality.

5. Solution Support and Services

We provide resources such as XSA/DSA, Petalinux images, and reference Bitstreams to facilitate rapid customer development. At the same time, we provide commercial services such as hardware customization, algorithm adaptation, mass production testing, CE/FCC/CCC certification assistance, etc. We can also combine our self-developed RISC-V+DPU development board to form a heterogeneous cluster, and use unified gRPC/REST scheduling to meet customers' diversified and complex business needs, providing customers with comprehensive one-stop solutions.

WeChat
- Scan WeChat for Consultation -