Diffusion models have demonstrated remarkable generation quality but at the cost of numerous function evaluations. Recently, advanced ODE-based solvers have been developed to mitigate the substantial computational demands of reverse-diffusion solving under limited sampling steps. However, these solvers, heavily inspired by Adams-like multistep methods, rely solely on t-related Lagrange interpolation. We show that t-related Lagrange interpolation is suboptimal for diffusion model and reveal a compact search space comprised of time steps and solver coefficients. Building on our analysis, we propose a novel differentiable solver search algorithm to identify more optimal solver. Equipped with the searched solver, rectified-flow models, e.g., SiT-XL/2 and FlowDCN-XL/2, achieve FID scores of 2.40 and 2.35, respectively, on ImageNet-$256\times256$ with only 10 steps. Meanwhile, DDPM model, DiT-XL/2, reaches a FID score of 2.33 with only 10 steps. Notably, our searched solver outperforms traditional solvers by a significant margin. Moreover, our searched solver demonstrates generality across various model architectures, resolutions, and model sizes.

The Sign Language Production (SLP) project aims to automatically translate spoken languages into sign sequences. Our approach focuses on the transformation of sign gloss sequences into their corresponding sign pose sequences (G2P). In this paper, we present a novel solution for this task by converting the continuous pose space generation problem into a discrete sequence generation problem. We introduce the Pose-VQVAE framework, which combines Variational Autoencoders (VAEs) with vector quantization to produce a discrete latent representation for continuous pose sequences. Additionally, we propose the G2P-DDM model, a discrete denoising diffusion architecture for length-varied discrete sequence data, to model the latent prior. To further enhance the quality of pose sequence generation in the discrete space, we present the CodeUnet model to leverage spatial-temporal information. Lastly, we develop a heuristic sequential clustering method to predict variable lengths of pose sequences for corresponding gloss sequences. Our results show that our model outperforms state-of-the-art G2P models on the public SLP evaluation benchmark. For more generated results, please visit our project page: https://slpdiffusier.github.io/g2p-ddm.

Arbitrary-resolution image generation still remains a challenging task in AIGC, as it requires handling varying resolutions and aspect ratios while maintaining high visual quality. Existing transformer-based diffusion methods suffer from quadratic computation cost and limited resolution extrapolation capabilities, making them less effective for this task. In this paper, we propose FlowDCN, a purely convolution-based generative model with linear time and memory complexity, that can efficiently generate high-quality images at arbitrary resolutions. Equipped with a new design of learnable group-wise deformable convolution block, our FlowDCN yields higher flexibility and capability to handle different resolutions with a single model. FlowDCN achieves the state-of-the-art 4.30 sFID on $256\times256$ ImageNet Benchmark and comparable resolution extrapolation results, surpassing transformer-based counterparts in terms of convergence speed (only $\frac{1}{5}$ images), visual quality, parameters ($8\%$ reduction) and FLOPs ($20\%$ reduction). We believe FlowDCN offers a promising solution to scalable and flexible image synthesis.

Long-term evolution advanced (LTE-A) heterogeneous networks have been observed to offer reliable and service-differentiated communication, thereby enabling numerous mobile applications such as smart meters, remote sensors, and vehicular applications. This fact envisions the trend of Internet of Things (IoT) underlaying heterogeneous small cell networks. On this basis, this paper proposes an energy-efficient framework for such a scenario, where multitier heterogeneous small cell networks provide wireless connection and seamless coverage for mobile users and IoT nodes. In our proposed framework, an elastic cell-zooming algorithm based on the quality of service and traffic loads of end-users is performed by adaptively adjusting the transmission power of small cells in order to reduce energy consumption. In addition, aiming at the high energy efficiency of IoT underlaying small cell networks, a clustering-based IoT structure is used, where a SWIPT-CH selection algorithm is proposed to maximize the average residual energy of IoT nodes and to mitigate resource competition between IoT nodes and mobile users. Extensive simulations demonstrate that our proposed framework can significantly enhance the energy efficiency for IoT underlaying small cell networks with guaranteed outage probability.

Image Signal Processor (ISP) is a crucial component in digital cameras that transforms sensor signals into images for us to perceive and understand. Existing ISP designs always adopt a fixed architecture, e.g., several sequential modules connected in a rigid order. Such a fixed ISP architecture may be suboptimal for real-world applications, where camera sensors, scenes and tasks are diverse. In this study, we propose a novel Reconfigurable ISP (ReconfigISP) whose architecture and parameters can be automatically tailored to specific data and tasks. In particular, we implement several ISP modules, and enable back-propagation for each module by training a differentiable proxy, hence allowing us to leverage the popular differentiable neural architecture search and effectively search for the optimal ISP architecture. A proxy tuning mechanism is adopted to maintain the accuracy of proxy networks in all cases. Extensive experiments conducted on image restoration and object detection, with different sensors, light conditions and efficiency constraints, validate the effectiveness of ReconfigISP. Only hundreds of parameters need tuning for every task1.

We consider mutual discovery of asynchronous wireless transceivers with a fixed activity ratio. On-off activity patterns are slotted, and repeat in discovery frames. For discovery it has to be guaranteed that the activity patterns of two transceivers are not overlapping. We design necklace codes determining activity patterns to guarantee discovery within a discovery frame, so that two asynchronous transceivers always have non-overlapping activity patterns. The number of distinct patterns is analyzed as a function of discovery frame length, and on-off activity ratio. As an application example, we consider the discovery problem for vehicle-to-vehicle communication. To guarantee discovery of far-away vehicles, discovery sequences providing processing gain, and necklace coded activity patterns are needed. We find that billions of discovery code identities can be provided with a repetition frequency that is high enough to guarantee a missed discovery probability less than 10-6.

The fifth generation (5G) of cellular networks is expected to provide connectivity for a wide range of services. This requires the network to encounter novel features. Ultra-reliable communications (URC) is one of the considered features, which provides a certain level of communication service almost all the time. This is essential in order to support mission-critical applications, such as industrial automation, public safety, and vehicular communications. This paper studies link adaptation optimization for URC, considering errors in both data and feedback channels. As the implementation of optimal link adaptation is challenging, particularly, for downlink transmissions due to the limited feedback channel, a simple link adaption scheme is also proposed. Results reveal that the performances of the proposed and optimal link adaptation schemes are close. Hence, the proposed scheme can be utilized to efficiently support URC in cellular networks.

Cellular systems are gaining more attention as a means of connectivity solutions for machine-type communications (MTC). The future cellular systems need to encounter new features in order to meet the diverse requirements of MTC applications. Ultra-reliable with low-latency communications (URLLC) is one of the considered features, required for mission-critical applications. This paper attempts to provide a better understanding of URLLC in the cellular systems. It considers different transmission modes and presents the required signaling and data transmissions. The optimal resource allocations are formulated for data transmissions. In addition, simplified solutions are given in order to reduce the computation complexity of resource allocations.

Fifth generation (5G) wireless systems will provide connectivity for a wide range of new applications with diverse requirements. In part, the network needs to support ultra-reliable communications with low-latency (URLLC) for mission-critical applications. For these applications, the generated data should be delivered with a limited number of transmission attempts with high success probability. This paper considers the optimal transmission and resource allocations for URLLC in cellular systems. The resource allocations are derived for the fixed and adaptive transmission attempt assignments. The analysis results reveal that both fixed and adaptive transmission assignments, applicable to automatic repeat request (ARQ) and hybrid ARQ (HARQ) schemes, can reduce the required resources compared to the equal transmission assignment.

The focus of this paper on mission-critical communications in a 5G cellular communication system. Technologies to provide ultra-reliable communication, with 99.999 % availability in a factory environment are studied. We have analysed the feasibility requirements for ultra-reliable communication and obtained the loss margins against path loss, shadow and fast fading. We also study the effect of increased interference due to higher deployment density on offered reliable rates and packet delays. Resource allocation schemes based on full and orthogonal resource sharing, as well as power control are compared. Last, the importance of multi-hop communication and multi-point coordination schemes are highlighted to improve the reliable communication in presence of interference and clutter.

Evolving 5G cellular communication is envisioned to enable connectivity for a wide range of new use cases. The focus on mission-critical communications, such as factory automation, public safety and vehicular safety applications presses the demand for ultra-reliable communication. In this paper, we investigate by various means of multi-hop relaying schemes to improve the reliable communication in a factory environment. Impacts due to half- and full-duplex operation at relay nodes are studied together with resource allocation. No retransmissions are considered and the relaying schemes are limited to two hops to meet the low latency constraint. Performance evaluations in a factory scenario are shown. We illustrate that in a network of uncoordinated cells one can exploit the varying fading profiles and can improve the offered reliable rate using multi-hop communication.

In order to harvest the business potential of deviceto-device (D2D) communication, direct communication between devices subscribed to different mobile operators should be supported. This would also support meeting requirements resulting from D2D relevant scenarios, like vehicle-to-vehicle communication. In this paper, we propose to allocate the multi-operator D2D communication over dedicated cellular spectral resources contributed from both operators. Ideally, the operators should negotiate about the amount of spectrum to contribute, without revealing proprietary information to each other and/or to other parties. One possible way to do that is to use the sequence of operators' best responses, i.e., the operators make offers about the amount of spectrum to contribute using a sequential updating procedure until reaching consensus. Besides spectrum allocation, we need a mode selection scheme for the multi-operator D2D users. We use a stochastic geometry framework to capture the impact of mode selection on the distribution of D2D users and assess the performance of the best response iteration algorithm. With the performance metrics considered in the paper, we show that the best response iteration has a unique Nash equilibrium that can be reached from any initial strategy. In general, asymmetric operators would contribute unequal amounts of spectrum for multi-operator D2D communication. Provided that the multi-operator D2D density is not negligible, we show that both operators may experience significant performance gains as compared to the scheme without spectrum sharing.

The explosive growth of the demand for higher data rates in mobile networks have been mainly driven by the increasing use of cloud based applications by smartphones. This has led the industry to investigate new radio access technologies to be deployed as part of 5G networks, while providing mechanisms to manage user mobility and traffic in a more efficient manner. In this paper, we consider a mobility and traffic management mechanism that proposes a close interaction between the cloud data servers and the radio access network to enable efficient network operation. Such a management mechanism is enabled by utilizing the application-dependent delay tolerance properties of the cloud data, with the delay values conveyed to the radio access network and UE to manage the service requests for the cloud data. The mechanism was evaluated using LTE-Advanced heterogeneous network scenario and 5G dense-urban information society scenario from EU FP7 METIS project, and relative gains in terms of packet delays and throughput values are presented. The results indicate significant gains using the proposed management mechanism as compared to the reference case where no such enhancements are used.

Device-to-device (D2D) proximity services has been one of the key topics in Release 12 of 3GPP LTE and most of its technical specifications were recently concluded. Among them, synchronization procedure is a crucial step required for devices to proceed with any further operation such as discovery and communication. Once the devices are synchronized, other operations can be initiated. Technical details pertaining to synchronization procedure and its signal structure were finalized in the latest 3GPP meeting. In this paper, we address the implementation constraints and propose a detection method for primary synchronization signal (PSS) in D2D. D2D primary synchronization signals (D2DPSS) are primarily based on the synchronization signals in downlink, but with required modifications in terms of signal position, periodicity, power, etc. D2D signals are single-carrier frequency division multiplexed (SCFDM), transmitted at lower power from user equipment and hence the detection is much more difficult. Therefore, it becomes necessary to take these constraints into consideration and develop a robust detection method for PD2DSS. The proposed detection method referred to as selective summing of correlations provides the target performance while operating within constraints and requirements of D2D. In comparison to existing methods, this method reaches the target hit-rate at lower SNR values, while maintaining low false-alarm rate.

In this paper, a novel simple generalized Chase algorithm (GCA) based multiuser detector (MUD) is proposed for MC-CDMA systems. After getting the initial solution from the front-end frequency domain linear minimum mean-squared error (LMMSE) receiver, different Chase-type algorithms are adopted to extend the potential solution list. Among the list, the candidate which can achieve the minimum distance to the received signal, is selected as the final solution. Both bit- and symbol-level receivers for high level modulation schemes are considered. Compared to the conventional MMSE receiver, the proposed receiver can provide superior performance with little increase in complexity. Moreover, the robustness of the proposed receiver is studied by incorporating a semi-blind subspace based channel estimation.

The spectral resources of high capacity cellular networks are desired to be utilized efficiently using aggressive frequency reuse. The consequence of the frequency reuse in each cell is inter-cell interference. In this paper, we consider downlink multi-carrier code-division multiple-access (MC-CDMA) in a multi-cell environment. In order to achieve good error rate performance and high spectral efficiency, multiple transmit and receive antennas with strong space-frequency channel coding and layered space-time (LST) architecture are applied. We study the performance of the minimum mean squared error (MMSE) based receivers in interference limited environment. The inter-cell interference suppression in MMSE detection and iterative detection method with soft co-antenna interference (CAI) cancellation are considered. The results indicate that the considered MC-CDMA system requiring no channel knowledge at the transmitter is suitable for high date rate downlink packet transmission in a cellular environment with relative high user load. The inter-cell interference suppression in MMSE detection provides remarkable performance improvement, especially in the case of spatially correlated antennas.

The performance of an interleaved multicarrier code-division multiple access system with a maximal-ratio combining receiver is studied with the help of the characteristic function of correlated Nakagami-m random variables. Computer simulations are used to demonstrate the accuracy of the analysis. Based on the analytical results, with an exhaustive search of different subcarrier interleaving schemes, the optimal subcarrier interleaver which provides the minimum bit-error rate (BER) can be achieved. Compared to a random interleaver, the optimal subcarrier interleaver offers significant performance improvement.

In this paper, a novel space–time-frequency minimum mean squared error (STF-MMSE)-based parallel interference cancellation receiver is proposed for space–time block-coded multicarrier code division multiple access systems in time-varying fading channels. The signal processing of this new detector is jointly implemented in space, time, and frequency domains, which leads to a powerful capability of combating interference coming from different sources. An adaptive implementation based on subspace estimation is proposed for slow-varying fading channels. Furthermore, based on the characteristic function of a complex Gaussian random vector, an analytical method to calculate the bit error probability of the proposed STF-MMSE receiver is presented. Representative examples of the detector are provided to demonstrate its superior performance.

For the single-user communications pragmatic yet powerful methods known as layered space-time (LST) architectures provide means to increase the user data rate of a multiple-input multiple-output (MIMO) antenna system dramatically. To achieve better error rate performance, the LST transmission can be further accompanied with forward error correction coding. Future wireless communication systems, however, require high data rates also in multiuser environments. For broadband multiuser wireless communications, multicarrier code-division multiple-access (MC-CDMA) has emerged as an attractive technique due to its low equalization complexity and robust performance in multipath fading channels. Utilizing strong channel coding and LST architectures with MC-CDMA, high spectral efficiency and good error rate performance can be obtained in diverse environments. In this paper, single and multi-antenna turbo coded downlink MC-CDMA is combined with the concept of LST architectures. Due to the inevitable complexity restrictions, a suboptimal receiver interface for the underlying system is designed and its performance is evaluated in fading channels. The results demonstrate that significant improvement in both error rate performance and the system throughput can be achieved also with multiuser communications by using these MIMO techniques.

We present a novel joint multiuser detection method based on sphere packing lattice decoding and semi-blind channel estimation for multicarrier code-division multiple-access (MC-CDMA) systems. After modelling MC-CDMA as a sphere packing lattice, a low-complexity maximum-likelihood (ML) detection, sphere decoding algorithm, is applied to jointly detect all users. The impacts of channel estimation errors are studied by incorporating a semi-blind subspace based channel estimation in the receiver. The selection of search radius and the complexity of the receiver are also investigated. Another promising detection technique, genetic algorithm (GA) based multiuser detector (MLJD) is also studied. Simulation results demonstrate the superior performance of the semi-blind receiver compared to the conventional receivers for MC-CDMA and its robustness to channel estimation errors.

We evaluate the performance of single- and multi-antenna multi-carrier code division multiple access (MC-CDMA) downlink (base station to mobile terminal) systems in multi-cell environments. We first propose a minimum mean square error (MMSE) filter for a single input single output (SISO) MC-CDMA downlink system. Then, we apply it to a SIMO (single input multiple output) system with a conventional turbo coding. Furthermore, we compare the performance of SISO (1 /spl times/ 1) and SIMO (1 /spl times/ 2) MC-CDMA systems with that of a multiple input multiple output (MIMO) (2 /spl times/ 2) system employing space-time turbo coded modulation (STTuCM) in a multi-cell environment with 7 cells by computer simulation. Based on the computer simulation results, it is found that the considered MIMO system can achieve twofold capacity with the same transmission power in the multi-cell environment.

Multicarrier code-division multiple-access (MC-CDMA) is a potentially attractive multiple access technique for future wireless communication systems. In this paper space-frequency minimum mean squared error based parallel interference cancellation receiver (SF-MMSE/PIC) is developed for MC-CDMA systems which produces the SF-MMSE combining as a byproduct. The signal processing of this new detector is jointly carried out in space and frequency domain which leads to the advantage of combating the interference from different sources simultaneously. Several representative simulation examples of the proposed SF-MMSE/PIC receiver are provided to demonstrate its powerful capability to suppress interference. In the fully loaded system, the proposed receiver can offer single user performance.

In this paper, a new practical method for determining the bit-error rate (BER) of equal gain combining (EGC) receivers for multicarrier code-division multiple-access (MC-CDMA) systems is presented. This new method is based on an alternative Gaussian approximation (AGA) and it applies to the system with multiple active users in frequency selective Nakagami-m fading channels. This new procedure takes advantage of an alternate integral representation of the conditional BER along with characteristic functions. The AGA method has the advantage of simplicity in expression and computational efficiency. The accuracy of the proposed approach is demonstrated with computer simulations in a fading channel.