-
Convergence Rate Maximization for Split Learning-based Control of EMG Prosthetic Devices
Authors:
Matea Marinova,
Daniel Denkovski,
Hristijan Gjoreski,
Zoran Hadzi-Velkov,
Valentin Rakovic
Abstract:
Split Learning (SL) is a promising Distributed Learning approach in electromyography (EMG) based prosthetic control, due to its applicability within resource-constrained environments. Other learning approaches, such as Deep Learning and Federated Learning (FL), provide suboptimal solutions, since prosthetic devices are extremely limited in terms of processing power and battery life. The viability…
▽ More
Split Learning (SL) is a promising Distributed Learning approach in electromyography (EMG) based prosthetic control, due to its applicability within resource-constrained environments. Other learning approaches, such as Deep Learning and Federated Learning (FL), provide suboptimal solutions, since prosthetic devices are extremely limited in terms of processing power and battery life. The viability of implementing SL in such scenarios is caused by its inherent model partitioning, with clients executing the smaller model segment. However, selecting an inadequate cut layer hinders the training process in SL systems. This paper presents an algorithm for optimal cut layer selection in terms of maximizing the convergence rate of the model. The performance evaluation demonstrates that the proposed algorithm substantially accelerates the convergence in an EMG pattern recognition task for improving prosthetic device control.
△ Less
Submitted 12 May, 2024; v1 submitted 6 January, 2024;
originally announced January 2024.
-
Optimal time sharing in underlay cognitive radio systems with RF energy harvesting
Authors:
Valentin Rakovic,
Daniel Denkovski,
Zoran Hadzi-Velkov,
Liljana Gavrilovska
Abstract:
Due to the fundamental tradeoffs, achieving spectrum efficiency and energy efficiency are two contending design challenges for the future wireless networks. However, applying radio-frequency (RF) energy harvesting (EH) in a cognitive radio system could potentially circumvent this tradeoff, resulting in a secondary system with limitless power supply and meaningful achievable information rates. This…
▽ More
Due to the fundamental tradeoffs, achieving spectrum efficiency and energy efficiency are two contending design challenges for the future wireless networks. However, applying radio-frequency (RF) energy harvesting (EH) in a cognitive radio system could potentially circumvent this tradeoff, resulting in a secondary system with limitless power supply and meaningful achievable information rates. This paper proposes an online solution for the optimal time allocation (time sharing) between the EH phase and the information transmission (IT) phase in an underlay cognitive radio system, which harvests the RF energy originating from the primary system. The proposed online solution maximizes the average achievable rate of the cognitive radio system, subject to the $\varepsilon$-percentile protection criteria for the primary system. The optimal time sharing achieves significant gains compared to equal time allocation between the EH and IT phases.
△ Less
Submitted 30 June, 2015;
originally announced July 2015.
-
Generic Multiuser Coordinated Beamforming for Underlay Spectrum Sharing
Authors:
Daniel Denkovski,
Valentin Rakovic,
Vladimir Atanasovski,
Liljana Gavrilovska,
Petri Mähönen
Abstract:
The beamforming techniques have been recently studied as possible enablers for underlay spectrum sharing. The existing beamforming techniques have several common limitations: they are usually system model specific, cannot operate with arbitrary number of transmit/receive antennas, and cannot serve arbitrary number of users. Moreover, the beamforming techniques for underlay spectrum sharing do not…
▽ More
The beamforming techniques have been recently studied as possible enablers for underlay spectrum sharing. The existing beamforming techniques have several common limitations: they are usually system model specific, cannot operate with arbitrary number of transmit/receive antennas, and cannot serve arbitrary number of users. Moreover, the beamforming techniques for underlay spectrum sharing do not consider the interference originating from the incumbent primary system. This work extends the common underlay sharing model by incorporating the interference originating from the incumbent system into generic combined beamforming design that can be applied on interference, broadcast or multiple access channels. The paper proposes two novel multiuser beamforming algorithms for user fairness and sum rate maximization, utilizing newly derived convex optimization problems for transmit and receive beamformers calculation in a recursive optimization. Both beamforming algorithms provide efficient operation for the interference, broadcast and multiple access channels, as well as for arbitrary number of antennas and secondary users in the system. Furthermore, the paper proposes a successive transmit/receive optimization approach that reduces the computational complexity of the proposed recursive algorithms. The results show that the proposed complexity reduction significantly improves the convergence rates and can facilitate their operation in scenarios which require agile beamformers computation.
△ Less
Submitted 24 April, 2015;
originally announced April 2015.
