FedFog: Network-Aware Optimization of Federated Learning over Wireless Fog-Cloud System

in IEEE Transactions on Wireless Communications (in press)

Federated learning (FL) is capable of performing large distributed machine learning tasks across multiple edge users by periodically aggregating trained local parameters. To address key challenges of ... [more ▼]

Federated learning (FL) is capable of performing large distributed machine learning tasks across multiple edge users by periodically aggregating trained local parameters. To address key challenges of enabling FL over a wireless fogcloud system (e.g., non-i.i.d. data, users’ heterogeneity), we first propose an efficient FL algorithm based on Federated Averaging (called FedFog) to perform the local aggregation of gradient parameters at fog servers and global training update at the cloud. Next, we employ FedFog in wireless fog-cloud systems by investigating a novel network-aware FL optimization problem that strikes the balance between the global loss and completion time. An iterative algorithm is then developed to obtain a precise measurement of the system performance, which helps design an efficient stopping criteria to output an appropriate number of global rounds. To mitigate the straggler effect, we propose a flexible user aggregation strategy that trains fast users first to obtain a certain level of accuracy before allowing slow users to join the global training updates. Extensive numerical results using several real-world FL tasks are provided to verify the theoretical convergence of FedFog. We also show that the proposed co-design of FL and communication is essential to substantially improve resource utilization while achieving comparable accuracy of the learning model. [less ▲]

Efficient Content Caching for Delivery Time Minimization in the LEO Satellite Networks

Scientific Conference (2023, June 01)

Next-generation multi-spot beam satellite systems open a new way to design low earth orbit (LEO) satellite communication systems with full flexibility in managing bandwidth, transmit power, and spot beam ... [more ▼]

Next-generation multi-spot beam satellite systems open a new way to design low earth orbit (LEO) satellite communication systems with full flexibility in managing bandwidth, transmit power, and spot beam coverage, enabling the adoption of spatial multiplexing techniques to meet the unprecedented demand for future mobile traffic. However, conventional spatial multiplexing techniques perform poorly in satellite systems due to high correlation between the satellite channels, resulting in inefficient mitigation of inter-user interference. In this paper, we exploit the flexibility of multi-spot beam LEO satellites and consider the geographic distribution of users to improve the performance of LEO satellite-assisted edge caching systems. Our goal is to jointly optimize the beam coverage, bandwidth and transmit power and minimize the cache delivery time. In particular, the spot beam coverage is optimized by using the K-means algorithm applied to the realistic user demands, followed by a proposed successive convex approximation (SCA)-based iterative algorithm for optimizing the radio resources. Simulations shows that our optimal approach outperforms the conventional precoding-based approach and also shows a significant improvement in the minimization of the maximum content delivery time. [less ▲]

Energy-Efficient RIS-Enabled NOMA Communication for 6G LEO Satellite Networks

E-print/Working paper (2023)

Reconfigurable Intelligent surfaces (RIS) have the potential to significantly improve the performance of future 6G LEO satellite networks. In particular, RIS can improve the signal quality of ground ... [more ▼]

Reconfigurable Intelligent surfaces (RIS) have the potential to significantly improve the performance of future 6G LEO satellite networks. In particular, RIS can improve the signal quality of ground terminal, reduce power consumption of satellite and increase spectral efficiency of overall network. This paper proposes an energy-efficient RIS-enabled NOMA communication for LEO satellite networks. The proposed framework simultaneously optimizes the transmit power of ground terminals at LEO satellite and passive beamforming at RIS while ensuring the quality of services. Due to the nature of the considered system and optimization variables, the problem of energy efficiency maximization is formulated as non-convex. In practice, it is very challenging to obtain the optimal solution for such problems. Therefore, we adopt alternating optimization methods to handle the joint optimization in two steps. In step 1, for any given phase shift vector, we calculate efficient power for ground terminals at satellite using Lagrangian dual method. Then, in step 2, given the transmit power, we design passive beamforming for RIS by solving the semi-definite programming. To validate the proposed solution, numerical results are also provided to demonstrate the benefits of the proposed optimization framework. [less ▲]

Integration of NOMA with Reflecting Intelligent Surfaces: A Multi-cell Optimization with SIC Decoding Errors

E-print/Working paper (2023)

Reflecting intelligent surfaces (RIS) has gained significant attention due to its high energy and spectral efficiency in next-generation wireless networks. By using low-cost passive reflecting elements ... [more ▼]

Reflecting intelligent surfaces (RIS) has gained significant attention due to its high energy and spectral efficiency in next-generation wireless networks. By using low-cost passive reflecting elements, RIS can smartly reconfigure the signal propagation to extend the wireless communication coverage. On the other hand, non-orthogonal multiple access (NOMA) has been proven as a key air interface technique for supporting massive connections over limited resources. Utilizing the superposition coding and successive interference cancellation (SIC) techniques, NOMA can multiplex multiple users over the same spectrum and time resources by allocating different power levels. This paper proposes a new optimization scheme in a multi-cell RIS-NOMA network to enhance the spectral efficiency under SIC decoding errors. In particular, the power budget of the base station and the transmit power of NOMA users while the passive beamforming of RIS is simultaneously optimized in each cell. Due to objective function and quality of service constraints, the joint problem is formulated as non-convex, which is very complex and challenging to obtain the optimal global solution. To reduce the complexity and make the problem tractable, we first decouple the original problem into two sub-problems for power allocation and passive beamforming. Then, the efficient solution of each sub-problem is obtained in two-steps. In the first-step of For power allocation sub-problem, we transform it to a convex problem by inner approximation method and then solve it through a standard convex optimization solver in the second-step. Accordingly, in the first-step of passive beamforming, it is transformed to a standard semidefinite programming problem by successive convex approximation and different of convex programming methods. Then, penalty based method is used to achieve a Rank-1 solution for passive beamforming in second-step. Numerical results demonstrate the benefits of the proposed optimization scheme in the multi-cell RIS-NOMA network. [less ▲]

Beam Splash Mitigation for NGSO Spectrum Coexistence between Feeder and User Downlink

in International ITG 26th Workshop on Smart Antennas (WSA), Braunschweig, Germany, 27 Feb - 03 Mar 2023. (2023)

Optimizing Computational and Communication Resources for MEC Network Empowered UAV-RIS Communication

Scientific Conference (2022, December)

With the technological evolution and new applications, user equipment (UEs) has become a vital part of our lives. However, limited computational capabilities and finite battery life bottleneck the ... [more ▼]

With the technological evolution and new applications, user equipment (UEs) has become a vital part of our lives. However, limited computational capabilities and finite battery life bottleneck the performance of computationally demanding applications. A practical solution to enhance the quality of experience (QoE) is to offload the extensive computation to the mobile edge cloud (MEC). Moreover, the network’s performance can be further improved by deploying an unmanned aerial vehicle (UAV) integrated with intelligent reflective surfaces (IRS): an effective alternative to massive antenna systems to enhance the signal quality and suppress interference. In this work, the MEC network architecture is assisted by UAV-IRS to provide computational services to the UEs. To do so, a cost minimization problem in terms of computing time and hovering energy consumption is formulated. Furthermore, to achieve an efficient solution to a formulated challenging problem, the original optimization problem is decoupled into sub-problems using the block-coordinate decent method. Moreover, numerical results are compared to baseline schemes to determine the effectiveness of the proposed scheme. Simulation results demonstrate that the optimal allocation of local computational resources results in minimizing tasks’ computational time and hovering energy consumption. [less ▲]

Power Allocation for Space-Terrestrial Cooperation Systems with Statistical CSI

in IEEE Global Communications Conference (IEEE Globecom), Rio de Janeiro, Brazil, Dec. 2022 (2022, December)

Multiuser-MISO Precoding under Channel Phase Uncertainty in Satellite Communication Systems

in IEEE Open Journal of Vehicular Technology (2022)

Linear and symbol-level precoding in satellite communications have received increasing research attention thanks to their ability to tackle inter-beam interference, allowing the use of spectral resources ... [more ▼]

Linear and symbol-level precoding in satellite communications have received increasing research attention thanks to their ability to tackle inter-beam interference, allowing the use of spectral resources more efficiently. However, there are still challenges and open questions regarding the implementation of practical precoding systems taking the phase uncertainties in estimating the channel state information into account. This work assesses the impact of phase variations and uncertainties inherent to the satellite communication system operating a precoded forward link. Specifically, we address the inability to measure at the user terminal, the absolute phase rotation introduced by the channel, and the transponder local oscillator phase noise effects on the precoding operations considering the use of frequency division multiplexing in the forward-uplink transmission. We formally demonstrate that the system performance for linear and non-linear precoding operations is not affected by the uncertainty in the phase measurements at the user terminal. Additionally, we show that using a single frequency reference for all the local oscillators at the transponder does not avoid the phase variations related to the frequency division multiplexing in the forward-uplink. This work demonstrates that these phase variations would not affect the system performance for an ideal zero-delay precoding loop. However, this is not feasible in practical scenarios, where the phase noise of the frequency reference at the transponder and the loop delay determine the impact on the system performance. We validate our results by simulations considering three frequency references with different stability levels in a typical GEO satellite system. Our results suggest that practical implementations of multiuser-MISO precoding systems must include a differential phase synchronization loop to compensate for this performance degradation. [less ▲]

Integration of backscatter communication with multi-cell NOMA: a spectral efficiency optimization under imperfect SIC

Poster (2022, November 03)

Future wireless networks are expected to connect large-scale low-powered communication devices using the available spectrum resources. Backscatter communications (BC) is an emerging technology towards ... [more ▼]

Future wireless networks are expected to connect large-scale low-powered communication devices using the available spectrum resources. Backscatter communications (BC) is an emerging technology towards battery-free transmission in future wireless networks by leveraging ambient radio frequency (RF) waves that enable communications among wireless devices. Non-orthogonal multiple access (NOMA) has recently drawn significant attention due to its high spectral efficiency. The combination of these two technologies can play an important role in the development of future networks. This paper proposes a new optimization approach to enhance the spectral efficiency of nonorthogonal multiple access (NOMA)-BC network. Our framework simultaneously optimizes the power allocation of base station and reflection coefficient (RC) of the backscatter device in each cell under the assumption of imperfect signal decoding. The problem of spectral efficiency maximization is coupled on power and RC which is challenging to solve. To make this problem tractable, we first decouple it into two subproblems and then apply the decomposition method and Karush-Kuhn-Tucker conditions to obtain the efficient solution. Numerical results show the performance of the proposed NOMA-BC network over the pure NOMA network without BC. [less ▲]

Task-Oriented Data Compression for Multi-Agent Communications Over Bit-Budgeted Channels

in IEEE Open Journal of the Communications Society (2022)

Various applications for inter-machine communications are on the rise. Whether it is for autonomous driving vehicles or the internet of everything, machines are more connected than ever to improve their ... [more ▼]

Various applications for inter-machine communications are on the rise. Whether it is for autonomous driving vehicles or the internet of everything, machines are more connected than ever to improve their performance in fulfilling a given task. While in traditional communications the goal has often been to reconstruct the underlying message, under the emerging task-oriented paradigm, the goal of communication is to enable the receiving end to make more informed decisions or more precise estimates/computations. Motivated by these recent developments, in this paper, we perform an indirect design of the communications in a multi-agent system (MAS) in which agents cooperate to maximize the averaged sum of discounted one-stage rewards of a collaborative task. Due to the bit-budgeted communications between the agents, each agent should efficiently represent its local observation and communicate an abstracted version of the observations to improve the collaborative task performance. We first show that this problem can be approximated as a form of data-quantization problem which we call task-oriented data compression (TODC). We then introduce the state-aggregation for information compression algorithm (SAIC) to solve the formulated TODC problem. It is shown that SAIC is able to achieve near-optimal performance in terms of the achieved sum of discounted rewards. The proposed algorithm is applied to a geometric consensus problem and its performance is compared with several benchmarks. Numerical experiments confirm the promise of this indirect design approach for task-oriented multi-agent communications. [less ▲]