Design and Implementation of Digital Signal Processing Algorithms for Communication and Biomedical Applications

在过去的几十年里，数字信号处理（DSP）领域经历了爆炸式的增长。DSP技术已经成为我们日常生活中所需要或遇到的产品和服务中不可或缺的一部分。申请人在过去五年中在这一领域的研究工作已经取得了一些非常具体的成果，这些成果已与国际科学界分享，包括学术界和工业界，并产生了需要进一步研究的新思想和新方向。提出的研究计划的总体目标是开发高效的算法和架构，并为可靠处理图像，视频和生物医学信号奠定良好的数学基础，并为通信和生物医学应用提供经济有效的实施。五年来，物联网（IoT）的发展和对机器对机器连接的需求一直是惊人的。到2020年，全球IP网络将必须支持超过500亿台设备。为了在连接物联网的如此众多的设备之间建立可靠、流畅的通信，同时保持高效和可持续的功耗，开发通用、智能、快速、实时的DOA估计算法至关重要。深度学习为理解和识别图像开辟了无限的可能性。图像理解和分析本质上是一个大数据问题。因此，使用先进的深度学习概念可以在图像理解和视觉跟踪方面产生有意义的结果，这在智能城市的检测和识别、人类行为分析、视频索引和检索、医学成像和交通管理等广泛的现实应用中发挥重要作用。阿尔茨海默病是影响全世界老年人的最常见的痴呆症。这种疾病无法治愈，而且随着病情的发展，病情会恶化。早期发现是预防、减缓和阻止阿尔茨海默病的关键。对病人进行全天候观察既昂贵又不方便，在某些情况下甚至根本不可能。身体区域网络技术提出了一种方便实用的解决方案，有望为医生提供随时随地访问患者生理数据的途径。生物医学信号的压缩感知和统计学习的应用可以使这项技术负担得起。在不久的将来，本研究旨在通过开发高效的DSP算法和架构来调查这些问题并寻求解决方案。这项研究的直接受益者将是加拿大的电信和生物医药行业。这项建议还将有助于培训一些技术人员，从而对加拿大工业界和学术界作出贡献。

The field of digital signal processing (DSP) has experienced explosive growth during the past couple of decades. The DSP techniques have become integral parts of the products and services that we need or encounter in our daily lives. The research efforts of the applicant in this area during the past five years have led to some very concrete results that have been shared with the international scientific community, both from academia and industry, and have given rise to new ideas and directions that need to be further investigated. The overall objective of the proposed research program is to develop efficient algorithms and architectures and to lay sound mathematical foundations for reliable processing of image, video and biomedical signals, and cost-effective implementation for communication and biomedical applications. For half a decade now, the growth in the internet of things (IoT) and the demand for machine-to-machine connections have been staggering. By 2020, the global IP networks will have to support more than 50 billion devices. In order to establish a reliable and smooth communication among such a large number of devices connected to IoT, while still maintaining efficient and sustainable power consumption, the development of versatile, smart, fast, and real-time direction of arrival (DOA) estimation algorithms is of paramount importance. Deep learning has opened up the limitless possibilities in understanding and recognition of images. Image understanding and analysis is inherently a big data problem. Hence, the use of advanced deep learning concepts can produce meaningful results in image understanding and visual tracking, which play an important role in a wide range of real-life applications such as detection and recognition, human behavior analysis, video indexing and retrieval, medical imaging, and traffic management of smart cities. Alzheimer’s disease is the most common type of dementia affecting elderly people worldwide. There is no cure for this disease, and the disease worsens as it progresses. Early detection is a key to preventing, slowing and stopping Alzheimer’s disease. Around-the-clock observation of patients is expensive, inconvenient, and in some cases, simply impossible. The technology of body area networks proposes a convenient and useful solution that promises to provide physicians access to the patient’s physiological data anytime and anywhere. Application of compressed sensing and statistical learning of biomedical signals can make this technology affordable. This research in the immediate future is aimed at investigating these problems and seeking their solutions by developing efficient DSP algorithms and architectures. The direct beneficiary of this research will be Canadian telecommunication and biomedical industry. This proposal will also contribute to Canadian industry and academia by enabling the training of a number of skilled personnel.