The simplified joint channel estimation and symbol detection based on the EM (expectation-maximization) algorithm for space-time block code (STBC) are proposed. By assuming channel to be invariant within only one STBC...The simplified joint channel estimation and symbol detection based on the EM (expectation-maximization) algorithm for space-time block code (STBC) are proposed. By assuming channel to be invariant within only one STBC word and utilizing the orthogonal structure of STBC, the computational complexity and cost of this algorithm are both very low, so it is very suitable to implementation in real systems.展开更多
无线图像传输面临着带宽和计算资源的双重挑战,在节点计算能力有限的物联网等应用场景中尤为突出.联合信源信道编码(Joint Source-Channel Coding,JSCC)能够同时优化信源和信道编码,逐渐成为无线图像传输中一个重要研究方向.近年来,基...无线图像传输面临着带宽和计算资源的双重挑战,在节点计算能力有限的物联网等应用场景中尤为突出.联合信源信道编码(Joint Source-Channel Coding,JSCC)能够同时优化信源和信道编码,逐渐成为无线图像传输中一个重要研究方向.近年来,基于深度学习的JSCC方法受到广泛关注,其通过端到端训练实现编码器与解码器的联合优化.然而,大多数基于深度学习的JSCC方法的编码器涉及大量的线性与非线性运算,导致计算复杂度较高,难以应用于物联网边缘计算节点等计算资源受限的设备.为实现轻量化的编码过程,本文提出了一种基于深度压缩感知的联合信源信道编码方法BCS-JSCC(Block Compressive Sensing-Joint Source Channel Coding),实现对于编解码器的端到端优化.该方法在编码端设计可学习尺度二值化测量的压缩感知采样,实现噪声环境下匹配解码器的轻量化编码方法;在解码端,基于MMSE(Minimum Mean Squared Error)准则求解测量值传输的线性逆问题,获得信道噪声敏感的初始重建,抑制噪声对参数复用重建网络的影响.与现有的基于深度学习的JSCC方法相比,在保持编码端每像素浮点计算次数(FLOating Point operations per pixel,FLOPs per pixel)相同的条件下,本文所提出的BCS-JSCC方法在高信噪比条件下可以取得更好的传输性能.在低算力(0.10 K FLOPs/pixel)情况下,优势更为明显.本文提出的BCS-JSCC方法编码器构造简单、计算量低,适用于物联网边缘计算节点等低算力设备部署.展开更多
基金This project was supported by the National Natural Science Foundation of China (60272079).
文摘The simplified joint channel estimation and symbol detection based on the EM (expectation-maximization) algorithm for space-time block code (STBC) are proposed. By assuming channel to be invariant within only one STBC word and utilizing the orthogonal structure of STBC, the computational complexity and cost of this algorithm are both very low, so it is very suitable to implementation in real systems.
文摘无线图像传输面临着带宽和计算资源的双重挑战,在节点计算能力有限的物联网等应用场景中尤为突出.联合信源信道编码(Joint Source-Channel Coding,JSCC)能够同时优化信源和信道编码,逐渐成为无线图像传输中一个重要研究方向.近年来,基于深度学习的JSCC方法受到广泛关注,其通过端到端训练实现编码器与解码器的联合优化.然而,大多数基于深度学习的JSCC方法的编码器涉及大量的线性与非线性运算,导致计算复杂度较高,难以应用于物联网边缘计算节点等计算资源受限的设备.为实现轻量化的编码过程,本文提出了一种基于深度压缩感知的联合信源信道编码方法BCS-JSCC(Block Compressive Sensing-Joint Source Channel Coding),实现对于编解码器的端到端优化.该方法在编码端设计可学习尺度二值化测量的压缩感知采样,实现噪声环境下匹配解码器的轻量化编码方法;在解码端,基于MMSE(Minimum Mean Squared Error)准则求解测量值传输的线性逆问题,获得信道噪声敏感的初始重建,抑制噪声对参数复用重建网络的影响.与现有的基于深度学习的JSCC方法相比,在保持编码端每像素浮点计算次数(FLOating Point operations per pixel,FLOPs per pixel)相同的条件下,本文所提出的BCS-JSCC方法在高信噪比条件下可以取得更好的传输性能.在低算力(0.10 K FLOPs/pixel)情况下,优势更为明显.本文提出的BCS-JSCC方法编码器构造简单、计算量低,适用于物联网边缘计算节点等低算力设备部署.
