ITR/Collaborative Research: FPGA Wavelet Image Compression Systems for Mobile Wireless Networks

在许多成像应用中，移动的代理--例如无绳侦察机器人和微型飞行器--通过无线网络传输和接收带宽密集型图像数据。开发小型、移动的代理的一个关键问题是最大限度地降低其功率需求。无线通信系统中最重要的功耗来源是通过无线链路传输数据。在传输前对数据进行压缩是显著降低功耗的直接方法;因此，有效的图像压缩算法对于实现移动、无线、实时、高质量的成像系统至关重要。移动计算可以考虑各种硬件平台。在这些平台中，现场可编程门阵列（FPGA）对于无线、移动的、快速、低功耗、手持/可穿戴/嵌入式图像压缩设备是最引人注目的。FPGA可以作为实现的基础，是低成本，易于更新，并显着快于基于微处理器的implementation.The项目的目标是双重的：设计新的，高效的图像压缩算法和设计，构建，演示和评估FPGA实现的国家的最先进的图像压缩系统。我们紧密团结的跨学科团队的图像压缩和硬件专家将允许端到端的系统性能分析。这种关键的评估几乎从来没有执行，但它是重要的，以了解算法和实现设计选择的综合影响。此外，两个或三个现实世界的跨学科项目，可用于高级数字信号处理和数字设计类将设计和出版。

In many imaging applications, mobile agents--such as untethered reconnaissance robots and micro air vehicles--transmit and receive bandwidth-intensive image data over wireless networks. A key issue in developing small, mobile agents is minimizing their power requirements. The most significant source of power dissipation in a wireless communications system is the transmission of data over the wireless link. Compressing the data before transmission is a direct way to significantly decrease power requirements; thus, effective image compression algorithms are critical to realizing a mobile, wireless, real-time, high quality imaging system.Various hardware platforms can be considered for mobile computing. Of these platforms, field programmable gate arrays (FPGAs) are the most compelling for wireless, mobile, fast, low-power, handheld/wearable/embedded, image compression devices. FPGAs can be used as the basis for implementations that are low cost, easily updated, and significantly faster than microprocessor-based implementations.The objective of this project is twofold: to design new, efficient image compression algorithms and to design, build, demonstrate and evaluate FPGA implementations of state-of-the-art image compression systems. Our close-knit interdisciplinary team of image compression and hardware specialists will allow for end-to-end system performance analysis. This critical assessment is almost never performed and yet it is important to understanding the combined effects of algorithmic and implementation design choices.In addition, two or three real-world interdisciplinary projects that can be utilized in senior-level digital signal processing and digital design classes will be designed and published.