Collaborative Research: CIF: Small: Beyond Compressed Sensing: Analog Coding for Communications

数字系统是现代通信的基石。然而，它们对于信道条件的变化并不总是鲁棒的，并且需要长数据块的传输来实现它们的全部潜力，这可能转化为相当大的延迟和高的编码/解码复杂度。这些问题可以通过使用端到端模拟编码器来缓解，端到端模拟编码器使用非线性变换直接处理所采集的样本，非线性变换的输出在适当调制后直接通过信道传输。在相同的性能下，这些方案比传统的数字系统具有更高的鲁棒性和更低的编解码复杂度。作为交换，模拟系统比数字系统更不灵活，更难设计。该项目旨在通过系统地开发低延迟、低复杂度的模拟方案来缩小这一差距，同时这些方案具有鲁棒性、灵活性和易于设计性。具体而言，该项目研究了模拟联合源信道编码领域中非常新颖的开发成果在点对点和多终端通信中的应用，考虑在感兴趣的源中存在时间或空间相关性的情况，并研究图像和视频编码中的实际应用。当使用数字方案时，原始源中存在的相关性可能会丢失，因此降低了端到端系统性能，这在所提出的模拟联合信源信道编码框架中不太可能发生。这一点，连同模拟方案对信道变化的鲁棒性，同时要求较低的处理复杂性和短得多的块长度，将使模拟编码成为满足现代通信系统不断增长的要求的一个极好的候选者，如果在该提议中考虑的一些开放的研究问题，例如能够实现任何期望速率的模拟速率兼容码的开发，该奖项反映了NSF的法定使命，并被认为是值得通过使用基金会的知识价值和更广泛的影响审查标准进行评估的支持。

Digital systems have been the cornerstone of modern communications. They are not always robust to changes in channel conditions, however, and require the transmission of long data blocks to achieve their full potential, which can translate to substantial delays and high encoding/decoding complexity. These problems can be alleviated by the use of end-to-end analog encoders, which directly process the acquired samples using a non-linear transformation whose output is transmitted directly through the channel after proper modulation. For the same performance, these schemes claim greater robustness and feature lower encoding/decoding complexity than traditional digital systems. In exchange, analog systems are less flexible and more difficult to design than their digital counterparts. This project aims at closing this gap by systematically developing low-latency, low-complexity analog schemes that at the same time are robust, flexible and easy to design.Specifically, this project investigates the application of very novel developments in the field of analog joint source-channel coding for point-to-point and multiterminal communications, considering the case in which temporal or spatial correlations are present in the sources of interest, and studying practical applications in image and video coding. Correlations existing in the original sources may be lost when digital schemes are utilized, therefore degrading the end-to-end system performance, something which is less likely to occur in the proposed analog joint source-channel coding framework. This, together with the robustness of analog schemes to channel variations, while requiring less processing complexity and much shorter block lengths, would make analog coding an excellent candidate to meet the ever growing requirements of modern communications systems if some of the open research problems considered in this proposal, such as the development of analog rate compatible codes capable of achieving any desired rate, are solved.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.