Joint Source-channel Coding Scheme Research Articles

CNN and Attention-Based Joint Source Channel Coding for Semantic Communications in WSNs

Wireless Sensor Networks (WSNs) have emerged as an efficient solution for numerous real-time applications, attributable to their compactness, cost-effectiveness, and ease of deployment. The rapid advancement of 5G technology and mobile edge computing (MEC) in recent years has catalyzed the transition towards large-scale deployment of WSN devices. However, the resulting data proliferation and the dynamics of communication environments introduce new challenges for WSN communication: (1) ensuring robust communication in adverse environments and (2) effectively alleviating bandwidth pressure from massive data transmission. In response to the aforementioned challenges, this paper proposes a semantic communication solution. Specifically, considering the limited computational and storage resources of WSN devices, we propose a flexible Attention-based Adaptive Coding (AAC) module. This module integrates window and channel attention mechanisms, dynamically adjusts semantic information in response to the current channel state, and facilitates adaptation of a single model across various Signal-to-Noise Ratio (SNR) environments. Furthermore, to validate the effectiveness of this approach, the paper introduces an end-to-end Joint Source Channel Coding (JSCC) scheme for image semantic communication, employing the AAC module. Experimental results demonstrate that the proposed scheme surpasses existing deep JSCC schemes across datasets of varying resolutions; furthermore, they validate the efficacy of the proposed AAC module, which is capable of dynamically adjusting critical information according to the current channel state. This enables the model to be trained over a range of SNRs and obtain better results.

An Effecient Joint Source-Channel Coding Scheme for Wireless Hierarchical Federated Learning and Its Information-Theoretic Analysis

An Effecient Joint Source-Channel Coding Scheme for Wireless Hierarchical Federated Learning and Its Information-Theoretic Analysis

Deep Joint Source-Channel Coding for Wireless Image Transmission with Adaptive Models

The implementation of joint source-channel coding (JSCC) schemes using deep learning has accelerated the development of semantic communication research. Existing JSCC schemes based on deep learning (DL) are trained on a fixed signal-to-noise ratio (SNR); however, these trained models are not designed for scenarios in which the SNR is dynamic. Therefore, a novel semantic adaptive model for semantic communication—called joint source-channel coding with adaptive models (AMJSCC)—that has a semantic adaptive model selection (SAMS) module is proposed. The joint source-channel encoding (JSCE) model and the joint source-channel decoding (JSCD) model adapt according to both real-time channel conditions and system available computational power resources. Furthermore, residual networks with different layers are investigated to further improve the accuracy of information recovery. Simulation results demonstrate that our model can achieve higher recovery similarity and is more robust and adaptive to the SNR and communication resources. Meanwhile, compared to the state-of-the-art deep JSCC methods, it reduces storage space and communication resource consumption.

Generative Joint Source-Channel Coding for Semantic Image Transmission

Recent works have shown that joint source-channel coding (JSCC) schemes using deep neural networks (DNNs), called DeepJSCC, provide promising results in wireless image transmission. However, these methods mostly focus on the distortion of the reconstructed signals with respect to the input image, rather than their perception by humans. However, focusing on traditional distortion metrics alone does not necessarily result in high perceptual quality, especially in extreme physical conditions, such as very low bandwidth compression ratio (BCR) and low signal-to-noise ratio (SNR) regimes. In this work, we propose two novel JSCC schemes that leverage the perceptual quality of deep generative models (DGMs) for wireless image transmission, namely InverseJSCC and GenerativeJSCC. While the former is an inverse problem approach to DeepJSCC, the latter is an end-to-end optimized JSCC scheme. In both, we optimize a weighted sum of mean squared error (MSE) and learned perceptual image patch similarity (LPIPS) losses, which capture more semantic similarities than other distortion metrics. InverseJSCC performs denoising on the distorted reconstructions of a DeepJSCC model by solving an inverse optimization problem using the pre-trained style-based generative adversarial network (StyleGAN). Our simulation results show that InverseJSCC significantly improves the state-of-the-art DeepJSCC in terms of perceptual quality in edge cases. In GenerativeJSCC, we carry out end-to-end training of an encoder and a StyleGAN-based decoder, and show that GenerativeJSCC significantly outperforms DeepJSCC both in terms of distortion and perceptual quality.

Analysis and Optimization of a General Linking Matrix for JSCC Scheme Based on Double LDPC Codes

A key component of the joint source-channel coding (JSCC) scheme based on double low-density parity-check (D-LDPC) codes is the introduction of a linking matrix between the source LDPC code and channel LDPC code, by which the decoding information including the source redundancy and channel state information can be transferred iteratively. However, the linking matrix is a fixed one-to-one mapping, i.e., an identity matrix in a conventional D-LDPC code system, which may not take full advantage of the decoding information. Therefore, this paper introduces a general linking matrix, i.e., a non-identity linking matrix, connecting the check nodes (CNs) of the source LDPC code and the variable nodes (VNs) of the channel LDPC code. Further, the encoding and decoding algorithms of the proposed D-LDPC coding system are generalized. A joint extrinsic information transfer (JEXIT) algorithm is derived for calculating the decoding threshold of the proposed system with a general linking matrix. In addition, several general linking matrices are optimized with the aid of the JEXIT algorithm. Finally, the simulation results demonstrate the superiority of the proposed D-LDPC coding system with general linking matrices.

Low-Delay Analog Joint Source-Channel Coding With Deep Learning

We consider the design of low-delay joint source-channel coding (JSCC) schemes for the transmission of discrete-time analog sources over noisy channels based on deep neural networks. The design problem is addressed as optimization of an autoencoder model, and several scenarios are discussed. For point-to-point communication of independent and identically distributed (i.i.d) Gaussian sources and Gauss-Markov sources over additive-white Gaussian noise (AWGN) channels, the encoder and decoder are constructed using recurrent neural networks (RNNs). With minimum prior knowledge used for design, the performance of these RNNs-based models is optimized using fine tuning techniques during training. Sinusoidal representation networks (SIRENs)-based models are proposed and optimized for three JSCC problems namely, transmitting multivariate Gaussian sources over AWGN channels, transmitting i.i.d Gaussian sources with side information at the decoder, and for communicating correlated sources over orthogonal Gaussian channels. We show that these deep learning-based JSCC schemes perform comparably or better than state-of-the-art (SOTA) traditional schemes. The proposed scheme can extend flexibly to different pairs of source and channel dimensions. Moreover, the spontaneously learned encoder mappings exhibit structured patterns that are interpretable.

Joint source-channel coding for 6G communications

In order to provide ultra low-latency and high energy-efficient communication for intelligences, the sixth generation (6G) wireless communication networks need to break out of the dilemma of the depleting gain of the separated optimization paradigm. In this context, this paper provides a comprehensive tutorial that overview how joint source-channel coding (JSCC) can be employed for improving overall system performance. For the purpose, we first introduce the communication requirements and performance metrics for 6G. Then, we provide an overview of the source-channel separation theorem and why it may not hold in practical applications. In addition, we focus on two new JSCC schemes called the double low-density parity-check (LDPC) codes and the double polar codes, respectively, giving their detailed coding and decoding processes and corresponding performance simulations. In a nutshell, this paper constitutes a tutorial on the JSCC scheme tailored to the needs of future 6G communications.

Concatenated Spatially Coupled LDPC Codes With Sliding Window Decoding for Joint Source-Channel Coding

In this paper, a method for joint source-channel coding (JSCC) based on concatenated spatially coupled low-density parity-check (SC-LDPC) codes is investigated. A construction consisting of two SC-LDPC codes is proposed: one for source coding and the other for channel coding, with a joint belief propagation-based decoder. Also, a novel windowed decoding (WD) scheme is presented with significantly reduced latency and complexity requirements. The asymptotic behavior for various graph node degrees is analyzed using a protograph-based Extrinsic Information Transfer (EXIT) chart analysis for both LDPC block codes with block decoding and for SC-LDPC codes with the WD scheme, showing robust performance for concatenated SC-LDPC codes. Simulation results show a notable performance improvement compared to existing state-of-the-art JSCC schemes based on LDPC codes with comparable latency and complexity constraints.

Bandwidth-Agile Image Transmission With Deep Joint Source-Channel Coding

We propose deep learning based communication methods for adaptive-bandwidth transmission of images over wireless channels. We consider the scenario in which images are transmitted progressively in layers over time or frequency, and such layers can be aggregated by receivers in order to increase the quality of their reconstructions. We investigate two scenarios, one in which the layers are sent sequentially, and incrementally contribute to the refinement of a reconstruction, and another in which the layers are independent and can be retrieved in any order. Those scenarios correspond to the well known problems of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">successive refinement</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">multiple descriptions</i> , respectively, in the context of joint source-channel coding (JSCC). We propose DeepJSCC- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$l$ </tex-math></inline-formula> , an innovative solution that uses convolutional autoencoders, and present three architectures with different complexity trade-offs. To the best of our knowledge, this is the first practical multiple-description JSCC scheme developed and tested for practical information sources and channels. Numerical results show that DeepJSCC- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$l$ </tex-math></inline-formula> can learn to transmit the source progressively with negligible losses in the end-to-end performance compared with a single transmission. Moreover, DeepJSCC- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$l$ </tex-math></inline-formula> has comparable performance with state of the art digital progressive transmission schemes in the challenging low signal-to-noise ratio (SNR) and small bandwidth regimes, with the additional advantage of graceful degradation with channel SNR.

A Type-Aware Coding Approach to Joint Source-Channel Coding

In this letter, we propose a type-aware coding approach to joint source-channel coding (JSCC) and present a concrete construction with polarization-adjusted convolutional (PAC) codes. A data sequence is first encoded by enumerative coding and then encoded by a data-dependent PAC code, resulting in a transmitted codeword. At the receiver, a type-search decoding algorithm is employed, and the implementation complexity can be reduced by searching only those typical types. Simulation results show that the presented JSCC scheme can outperform the conventional tandem coding scheme with a fixed-to-fixed-length source code and achieve a performance within the gap between the upper and lower bound on the minimum FER of finite-length JSCCs.

Two-Way Source-Channel Coding

We propose an adaptive lossy joint source-channel coding (JSCC) scheme for sending correlated sources over two-terminal discrete-memoryless two-way channels (DM-TWCs). The main idea is to couple the independent operations of the terminals via an adaptive coding mechanism, which can mitigate cross-interference resulting from simultaneous channel transmissions and concurrently exploit the sources' correlation to reduce the end-to-end reconstruction distortions. Our adaptive JSCC scheme not only subsumes existing lossy coding methods for two-way simultaneous communication but also improves their performance. Furthermore, we derive outer bounds for our two-way lossy transmission problem and establish complete JSCC theorems in some special settings. In these special cases, a non-adaptive separate source-channel coding (SSCC) scheme achieves the optimal performance, thus simplifying the design of the source-channel communication system.

Analysis and Improvement of Error-Floor Performance for JSCC Scheme Based on Double Protograph LDPC Codes

In a joint source-channel coding (JSCC) system, excessive source compression can cause an error floor. In the original JSCC scheme based on double protograph LDPC (DP-LDPC) codes, only connections exist between check nodes (CNs) of the source protograph and variable nodes (VNs) of the channel protograph, and error floors are observed. In this paper, we investigate the joint protograph that forms the basis of a DP-LDPC code. The joint protograph consists of a source protograph, a channel protograph, and connections between the two protographs. Our main focus is on the complete-source-variable-linking (CSVL) protomatrix of the joint protograph, which determines the error floor of the DP-LDPC code. We propose a generalized source protograph extrinsic information transfer (GSP-EXIT) algorithm for evaluating the source decoding thresholds of the CSVL protomatrix. Based on the proposed algorithm, we analyze CSVL protomatrices with regular or irregular source protographs. We present design criteria for connections between VNs of the source protograph and CNs of the channel protograph. Such design rules will result in a higher source decoding threshold, which implies a lower error floor. We also propose a differential evolution algorithm for optimizing the source protograph. We present/compare analytical results and/with simulation results, and conclude that they are consistent.

LQG control and linear policies for noisy communication links with synchronized side information at the decoder

We consider a multidimensional closed-loop control problem where a stochastic linear time-invariant plant is connected via a communication system that includes an encoder, vector additive Gaussian noise (AGN) channel, decoder to a controller. For this setup, we first examine the impact of the communication link when causal side information (CSI) that is correlated to the plant is (possibly) allowed at the decoder and the performance criterion is the classical linear quadratic (LQ) cost. The CSI is modeled as a time-invariant vector-valued Gaussian process and when available at the decoder, it happens concurrently to the signal obtained from the vector AGN channel. Under the assumption that our encoder is linear, we leverage the separation principle of control and estimation, to obtain the following new results. (i) For multivariate Gaussian processes, we completely characterize a lower bound on the communication cost term of the optimal LQ Gaussian (LQG) cost in the form of a convex optimization problem and identify the best linear policies that achieve this bound. (ii) For scalar-valued Gaussian processes, we derive an analytical solution to the lower bound of the optimal communication cost term and show that for linear policies, CSI at the encoder does not improve the specific bound. (iii) For scalar-valued Gaussian processes, we also derive a linear time-invariant low delay joint source-channel coding (JSCC) scheme without bandwidth constraints, when CSI is available at the encoder, which demonstrates the case at which the specific lower bound communication cost expression is achievable. Further, we give a suboptimal linear time-invariant JSCC scheme when CSI is available only at the decoder. We supplement our results with various simulation studies.

Design and Analysis of Joint Source Channel Coding Schemes Over Non-Standard Coding Channels

Although a mountain of work on joint source-channel coding (JSCC) scheme over standard coding channels (a combination of BPSK modulation and AWGN channels) have been studied deeply, there has been no knowledge about JSCC over non-standard coding channels. In this paper, a JSCC scheme based on double protograph low-density parity-check (DP-LDPC) codes is proposed over non-standard coding channels which consists of M -ary differential chaos shift keying (DCSK) and Rayleigh fading channels is proposed, named as JSCC-DCSK system. A joint protograph extrinsic information transform (JPEXIT) algorithm is proposed to analyze the decoding threshold of DP-LDPC codes on the non-standard coding channels. Considering the effects from the source redundancy left by source coding and the non-linear property of non-standard coding channels, the design of DP-LDPC structure is discussed, and new design principles from a global perspective are proposed. Several DP-LDPC codes for different source statistic are designed by the aid of JPEXIT algorithm and verified by Monte Carlo simulation. The appropriate number of the turbo iteration is determined to keep the bit error rate (BER) of system performance and the complexity of the receiver at a low level.

DeepJSCC-f: Deep Joint Source-Channel Coding of Images With Feedback

We consider wireless transmission of images in the presence of channel output feedback. From a Shannon theoretic perspective feedback does not improve the asymptotic end-to-end performance, and separate source coding followed by capacity-achieving channel coding, which ignores the feedback signal, achieves the optimal performance. It is well known that separation is not optimal in the practical finite blocklength regime; however, there are no known practical joint source-channel coding (JSCC) schemes that can exploit the feedback signal and surpass the performance of separation-based schemes. Inspired by the recent success of deep learning methods for JSCC, we investigate how noiseless or noisy channel output feedback can be incorporated into the transmission system to improve the reconstruction quality at the receiver. We introduce an autoencoder-based JSCC scheme, which we call DeepJSCC- $f$ , that exploits the channel output feedback, and provides considerable improvements in terms of the end-to-end reconstruction quality for fixed-length transmission, or in terms of the average delay for variable-length transmission. To the best of our knowledge, this is the first practical JSCC scheme that can fully exploit channel output feedback, demonstrating yet another setting in which modern machine learning techniques can enable the design of new and efficient communication methods that surpass the performance of traditional structured coding-based designs.

Transmission of Spatio-Temporal Correlated Sources Over Fading Multiple Access Channels With DQLC Mappings

The design of zero-delay Joint Source-Channel Coding (JSCC) schemes for the transmission of correlated information over fading Multiple Access Channels (MACs) is an interesting problem for many communication scenarios like Wireless Sensor Networks (WSNs). Among the different JSCC schemes so far proposed for this scenario, Distributed Quantizer Linear Coding (DQLC) represents an appealing solution since it is able to outperform uncoded transmissions for any correlation level at high Signal-to-Noise Ratios (SNRs) with a low computational cost. In this paper, we extend the design of DQLC-based schemes for fading MACs considering sphere decoding to make the optimal Minimum Mean Squared Error (MMSE) estimation computationally affordable for an arbitrary number of transmit users. The use of sphere decoding also allows to formulate a practical algorithm for the optimization of DQLC-based systems. Finally, non-linear Kalman Filtering for the DQLC is considered to jointly exploit the temporal and spatial correlation of the source symbols. The results of computer experiments show that the proposed DQLC scheme with the Kalman Filter decoding approach clearly outperforms uncoded transmissions for medium and high SNRs.

The Dispersion of Mismatched Joint Source-Channel Coding for Arbitrary Sources and Additive Channels

We consider a joint source channel coding (JSCC) problem in which we desire to transmit an arbitrary memoryless source over an arbitrary additive channel. We propose a mismatched coding architecture that consists of Gaussian codebooks for both the source reproduction sequences and channel codewords. The natural nearest neighbor encoder and decoder, however, need to be judiciously modified to obtain the highest communication rates at finite blocklength. In particular, we consider an unequal error protection scheme in which all sources are partitioned into disjoint power-type classes. We also regularize the nearest neighbor decoder so that an appropriate measure of the size of each power type class is taken into account in the decoding strategy. For such an architecture, we derive ensemble-tight second-order and moderate deviations results. Our first-order (optimal bandwidth expansion ratio) result generalizes the seminal results by Lapidoth (1996 and 1997). The dispersion of our JSCC scheme is a linear combination of the mismatched dispersions for the channel coding saddle-point problem by Scarlett, Tan, and Durisi (2017) and the rate-distortion saddle-point problem by the present authors, thus also generalizing these results.

Design of Linking Matrix in JSCC Scheme Based on Double Protograph LDPC Codes

A joint source channel coding (JSCC) scheme based on source and channel low density parity check (LDPC) codes, which is named as the D-LDPC system, has attracted a lot of attention recently. However, it suffers from a high error-floor. By introducing a linking matrix connecting variable nodes of source LDPC and check nodes of channel LDPC, the error-floor can be lowered effectively. However, the water-fall performance was lost with the increasing of the size of linking matrix. In this paper, a detailed analysis about the impact of this linking matrix on the water-fall region is conducted. Some design principles for the linking matrix are proposed, by which several linking matrices for different source statistics will be designed to improve the water-fall performance from the perspective of both the source and channel LDPC codes. The extrinsic information transfer (EXIT) analysis and numerical simulations will be compared to verify the superiority of the proposed linking matrices.

Joint Source-Channel Polarization With Side Information

As an extension of source polarization and channel polarization, this paper considers joint source-channel polarization, which results in a joint source-channel coding (JSCC) scheme using a quasi-uniform systematic polar code (SPC). In this JSCC scheme, the source with side information is encoded as a systematic polar codeword and only parity bits are transmitted through the channel. The indices of systematic bits are quasi-uniform, which enable the source and the channel to be jointly polarized to either a high entropy part or a low entropy part. The analysis reveals that the quasi-uniform SPC cannot be constructed via original polar coding. To solve this problem, additional bit-swap coding is introduced to modify original polar coding and construct this kind of SPCs. The proposed JSCC scheme can asymptotically approach the information-theoretical limit. For the noiseless channel, the proposed scheme is degraded into classic Slepain-Wolf coding or lossless source coding based on parity approach.

Analog Transmission of Correlated Sources Over BC With Distortion Balancing

Analog transmission of correlated information over Gaussian broadcast channels (BCs) using analog joint source–channel coding (JSCC) is addressed. This communication strategy has attractive advantages, such as low complexity, minimal delay, graceful degradation, and adaptation to time-varying environments. In this paper, we focus on the optimization of parametric continuous mappings that satisfy individual quality of service requirements over Gaussian BCs. This optimization is based on the balancing of the user distortions to ensure the feasibility of the resulting optimization problems. An analysis of the overall distortion corresponding to the considered mappings is carried out to design algorithms that determine the optimal values for the mappings parameters. Results show that the proposed analog JSCC scheme provides near optimal performance and an adequate balancing of the individual distortions.