ML Estimation and MAP Estimation for Device Activities in Grant-Free Random Access with Interference

Jiang, Dongdong; Cui, Ying

doi:10.1109/wcnc45663.2020.9120542

Cited by 18 publications

(28 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This paper aims to detect the activity status {a n } N n=1 based on the received signal Y at the BS. Since the IoT devices are stationary in many practical deployment scenarios, the largescale fading components can be obtained in advance and hence assumed to be known [21]- [23]. In order to reduce the channel gain variations among different devices, the transmit power of each device can be controlled based on the large-scale channel gain [8].…”

Section: A System Modelmentioning

confidence: 99%

“…CN (0, Σ) with Σ = E y m y H m = SGAS H + σ 2 I Lp . Consequently, the activity status {a n } N n=1 can be detected by maximizing the likelihood function [21]- [23] p(Y;…”

Section: B Existing Approachesmentioning

confidence: 99%

“…This allows the neural network to mimic the optimal mapping function f (•, •) directly from the ground-truth labels [35], [36]. Therefore, there is an opportunity to achieve better detection performance than state-of-the-art covariance based methods that only obtain a stationary point of the relaxed problems [21]- [23]. Moreover, after the training procedure, the neural networks can infer the solution of any new problem in a real-time manner due to the computationally cheap inference.…”

Section: B Parametrization By Neural Networkmentioning

confidence: 99%

See 2 more Smart Citations

Heterogeneous Transformer: A Scale Adaptable Neural Network Architecture for Device Activity Detection

Li¹,

Chen²,

Wang³

et al. 2021

Preprint

View full text Add to dashboard Cite

To support the modern machine-type communications, a crucial task during the random access phase is device activity detection, which is to detect the active devices from a large number of potential devices based on the received signal at the access point. By utilizing the statistical properties of the channel, state-of-the-art covariance based methods have been demonstrated to achieve better activity detection performance than compressed sensing based methods. However, covariance based methods require to solve a high dimensional nonconvex optimization problem by updating the estimate of the activity status of each device sequentially. Since the number of updates is proportional to the device number, the computational complexity and delay make the iterative updates difficult for real-time implementation especially when the device number scales up. Inspired by the success of deep learning for realtime inference, this paper proposes a learning based method with a customized heterogeneous transformer architecture for device activity detection. By adopting an attention mechanism in the architecture design, the proposed method is able to extract the relevance between device pilots and received signal, is permutation equivariant with respect to devices, and is scale adaptable to different numbers of devices. Simulation results demonstrate that the proposed method achieves better activity detection performance with much shorter computation time than state-of-the-art covariance approach, and generalizes well to different numbers of devices, BS-antennas, and different signalto-noise ratios.

show abstract

Section: A System Modelmentioning

confidence: 99%

“…CN (0, Σ) with Σ = E y m y H m = SGAS H + σ 2 I Lp . Consequently, the activity status {a n } N n=1 can be detected by maximizing the likelihood function [21]- [23] p(Y;…”

Section: B Existing Approachesmentioning

confidence: 99%

Section: B Parametrization By Neural Networkmentioning

confidence: 99%

See 1 more Smart Citation

Heterogeneous Transformer: A Scale Adaptable Neural Network Architecture for Device Activity Detection

Li¹,

Chen²,

Wang³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…It is necessary to pre-process the collected data to protect it from falsified data that leads to inaccurate results. Maximum likelihood (ML) or maximum a posterior (MAP) are used iteratively under Expectation-Maximization (EM) algorithm in the initial step of data pre-processing [29] [30]. The log-likelihood corrects the falsified data.…”

Section: Data Classification Layer (Dcl)mentioning

confidence: 99%

IoT Inspired Intelligent Monitoring and Reporting Framework for Education 4.0

et al. 2021

View full text Add to dashboard Cite

Education 4.0 contemplates the involvement of emerging technologies like IoT (Internet of Things), Fog and cloud computing. Ubiquitous real-time monitoring under the challenging environment of an educational institution is the key requirement for proper implementation of Education 4.0 around the world. Conspicuously, this research presents a novel monitoring and irregularity detection framework for educational institutions. In this research, acquisition of data while deploying IoT is being suggested in the comprehensive environs of the educational institutions. Pre-processing of acquired data for feature extraction is implemented using the Fog-Cloud nodes. Subsequently, the Temporal Chunk (TC)which is articulated with Temporal Data mining, is further employed to detect irregularities based on SoTL (Set of threshold limits)and HAA (Historical Adversity Approximation). Successively, Multi-Layered Bi-Directional Long Short Term Memory (M-Bi-LSTM) oriented irregularity prediction model is deployed. Furthermore, a fog-inspired alert generation and reporting module are employed for real-time reporting to notify the concerned stakeholder for in-time preventive action corresponding to predicted irregularity. To authenticate the proposed framework and to avoid the experimental implementation cost, the experimental simulations are performed. The experimental results verify that the proposed framework is capable to perform better in comparison to other contemporary decision-making methods for Delay Efficiency, Data Classification, Irregularity Prediction and System Stability. Moreover, the proposed framework can successfully estimate most of the irregularities in an educational environment and provide support for the proper implementation of Education 4.0 in a proficient and realistic manner.

show abstract

“…In the covariance approach, the coordinate descent (CD) algorithm that iteratively updates the activity indicator of each device is commonly used since it can achieve excellent detection performance; see [8], [12], [13] for more details. In the single-cell scenario, the CD algorithm is also computationally efficient because it admits a closed-form solution for each update.…”

Section: Introductionmentioning

confidence: 99%

Accelerating Coordinate Descent via Active Set Selection for Device Activity Detection for Multi-Cell Massive Random Access

Wang

Liu

Chen

et al. 2021

Preprint

View full text Add to dashboard Cite

We propose a computationally efficient algorithm for the device activity detection problem in the multi-cell massive multi-input multi-output (MIMO) system, where the active devices transmit their signature sequences to multiple BSs in multiple cells and all the BSs cooperate to detect the active devices. The device activity detection problem has been formulated as a maximum likelihood maximization (MLE) problem in the literature. The state-of-the-art algorithm for solving the problem is the (random) coordinate descent (CD) algorithm. However, the CD algorithm fails to exploit the special sparsity structure of the solution of the device activity detection problem, i.e., most of devices are not active in each time slot. In this paper, we propose a novel active set selection strategy to accelerate the CD algorithm and propose an efficient active set CD algorithm for solving the considered problem. Specifically, at each iteration, the proposed active set CD algorithm first selects a small subset of all devices, namely the active set, which contains a few devices that contribute the most to the deviation from the first-order optimality condition of the MLE problem thus potentially can provide the most improvement to the objective function, then applies the CD algorithm to perform the detection for the devices in the active set. Simulation results show that the proposed active set CD algorithm significantly outperforms the state-of-the-art CD algorithm in terms of the computational efficiency.

show abstract

ML Estimation and MAP Estimation for Device Activities in Grant-Free Random Access with Interference

Cited by 18 publications

References 9 publications

Heterogeneous Transformer: A Scale Adaptable Neural Network Architecture for Device Activity Detection

Heterogeneous Transformer: A Scale Adaptable Neural Network Architecture for Device Activity Detection

IoT Inspired Intelligent Monitoring and Reporting Framework for Education 4.0

Accelerating Coordinate Descent via Active Set Selection for Device Activity Detection for Multi-Cell Massive Random Access

Contact Info

Product

Resources

About