Frame mechanics: Dynamical principles for optimal redundant expansions

框架力学：最佳冗余扩展的动力学原理

• 批准号: 1109545
• 负责人: Bernhard Bodmann
• 金额: $ 21.49万
• 依托单位: University of Houston
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2011
• 资助国家: 美国
• 起止时间: 2011-09-15 至 2015-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1109545&HistoricalAwards=false
• 关键词: Frame; mechanics; Dynamical; principles; optimal

BodmannDMS-1109545 The concept of frame mechanics addresses the need for constructing an abundance of optimal redundant, stable expansions with frames, which have become central to applications of mathematics in remote sensing or wireless transmissions, in analog-digital conversion such as audio and video encoding, in packet-based network communications, noise-insensitive quantum computing and recently also in compressive sensing. Despite its popularity, the search for near-optimal frames has been successful mostly in small dimensions, or it had to rely on specific group-representation properties, or the use of randomization principles. In frame mechanics, the investigator is studying an alternative to the conventional, structured or random design methods by letting frames evolve under flows which drive them towards optimality, instead of constructing them directly. The general objectives are to find (1) appropriate frame dynamics, (2) suitable initializations, and to obtain (3) deterministic control of the approximation error. The envisioned outcome of the project includes leveraging recently established numerical results on the construction of equiangular tight frames for the verification of Zauner's conjecture (the existence of maximal Gabor frames in all finite-dimensional Hilbert spaces), constructing controlled approximations of Grassmannian frames and fusion frames for loss-insensitive transmissions in wireless or packet-based network communications, and the design of matrices for compressive sensing based on quantum chaotic dynamics which improve the restricted isometry properties of sensing matrices. The mathematics of redundant signal representations is called frame theory. For practical purposes, a frame is a tool which incorporates or removes repetitive information when data is stored, transmitted or received. Frames have become essential in many data-intensive areas of modern technology, because the repetitive information helps compensate errors of transmission devices and sensors. However, over the last decades, progress in the optimal design of frames has been outpaced by the rapid growth of data generated by our hardware. In frame mechanics, the investigator and his students explore a fundamentally new strategy to overcome this problem: The burden of constructing such optimal frames is put on the computer, which lets frames evolve in a way that drives them towards optimality. The goal of this project is to demonstrate that this dynamic design strategy is mathematically guaranteed to find many optimal frames where previous attempts failed. Frame mechanics allows us to maximize performance in remote sensing, seismic and medical imaging, wireless and fiber-optic communications, and to make internet transmissions robust to network outages.

BodmannDMS-1109545 帧力学的概念解决了用帧构建大量最佳冗余、稳定扩展的需求，这已成为遥感或无线传输、音频和视频编码等模数转换中数学应用的核心、基于数据包的网络通信、噪声不敏感的量子计算以及最近的压缩传感领域。 尽管它很受欢迎，但对近乎最优框架的搜索主要在小维度上取得了成功，或者它必须依赖于特定的组表示属性或随机化原理的使用。 在框架力学中，研究人员正在研究传统的结构化或随机设计方法的替代方法，让框架在驱动它们走向最优的流动下演化，而不是直接构建它们。 一般目标是找到（1）适当的框架动力学，（2）适当的初始化，并获得（3）近似误差的确定性控制。 该项目的预期成果包括利用最近建立的等角紧框架构造的数值结果来验证 Zauner 猜想（所有有限维希尔伯特空间中存在最大 Gabor 框架），构造格拉斯曼框架和融合框架的受控近似用于无线或基于分组的网络通信中的丢失不敏感传输，以及基于量子混沌动力学的压缩感知矩阵的设计，改善了感知矩阵的受限等距特性。 冗余信号表示的数学称为框架理论。 出于实用目的，框架是一种在存储、传输或接收数据时合并或删除重复信息的工具。 帧在现代技术的许多数据密集型领域已变得至关重要，因为重复信息有助于补偿传输设备和传感器的错误。 然而，在过去的几十年里，框架优化设计的进展已经被硬件生成的数据的快速增长所超越。 在框架力学中，研究人员和他的学生探索了一种全新的策略来克服这个问题：构建这种最佳框架的负担被放在计算机上，这让框架以一种驱使它们走向最优的方式进化。 该项目的目标是证明这种动态设计策略在数学上可以保证找到许多先前尝试失败的最佳框架。 框架力学使我们能够最大限度地提高遥感、地震和医学成像、无线和光纤通信的性能，并使互联网传输能够应对网络中断。