Empirical Method Based on Neural Networks for Analog Power Modeling

Suissa, Abraham; Romain, Olivier; Denoulet, Julien; Hachicha, Khalil; Garda, Patrick

doi:10.1109/tcad.2010.2043759

Cited by 22 publications

(11 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another approach is the use of Neural Network to model a system. In [4] the authors use neural networks to model consumption, not of a complete node but only for analog components. For each component, the authors measure the instantaneous energy consumption according to inputs and operating parameters.…”

Section: Related Workmentioning

confidence: 99%

Reinforcement Learning Strategies for Energy Management in Low Power IoT

Rioual

Laurent

Senn

et al. 2017

2017 International Conference on Computational Science and Computational Intelligence (CSCI)

View full text Add to dashboard Cite

Energy management in low power IoT is a difficult problem. Modeling the consumption of a sensor node is complicated, they operate in a stochastic environment. They harvest energy in their environment but energy sources present timevarying behavior. It becomes hazardous to predict in advance the energy behavior of our system. In this paper we propose a new approach using both neural networks to estimate the harvesting energy and reinforcement learning algorithms to find the operating parameters to maximize the node's performance while preserving its energy resources.

show abstract

Section: Related Workmentioning

confidence: 99%

Reinforcement Learning Strategies for Energy Management in Low Power IoT

Rioual

Laurent

Senn

et al. 2017

2017 International Conference on Computational Science and Computational Intelligence (CSCI)

View full text Add to dashboard Cite

show abstract

“…The neural networks have been utilized for modeling the RF subsystem power consumption at the system level of abstraction [13]. In this section, we examine the performance of the neural networks at prototype level.…”

Section: ) Multivariate Polynomial Fittingmentioning

confidence: 99%

LTE RF subsystem power consumption modeling

Musiige¹,

Vincent²,

Anton

et al. 2012

The 1st IEEE Global Conference on Consumer Electronics 2012

View full text Add to dashboard Cite

This paper presents a new power consumption emulation model, for all possible scenarios of the RF subsystem, when transmitting a LTE signal. The model takes the logical interface parameters, Tx power, carrier frequency and bandwidth between the baseband and RF subsystem as inputs to compute the power consumption. An analysis of modeling approaches was conducted and the modeling approach with the least sum of squared errors is used to compute the emulation model. The neural networks applying the Pseudo-Gauss Newton algorithm for optimization proved to have the least sum of squared errors. This approach was validated against a real life scenario with a relative error of 5.77%.

show abstract

“…Therefore, only features of the impact of the RF subsystem on the modem power consumption are modeled. Neural networks [27] have been utilized for modeling the power consumption of the RF subsystem at system level [28]. Hence, a two layer 4 × 5 × 1 neural network is applied to compute the emulation model of the modem in free space.…”

Section: A the Emulation Of Sar Mpe And Rf Signal Strengthmentioning

confidence: 99%

LTE modem power consumption, SAR and RF signal strength emulation

Musiige¹,

Vincent²,

Anton

2012

2012 IV International Congress on Ultra Modern Telecommunications and Control Systems

View full text Add to dashboard Cite

This paper presents a new methodology for emulating the LTE modem power consumption, emitted SAR and RF signal strength when transmitting an LTE signal. The inputs of the methodology are: modem logical/protocol commands, time advance, near-field specifier, and antenna characteristics. The power emulation model(s) are computed by a two layer 4 × 5 × 1 neural network based on physical power measurements. SAR is emulated by polynomial interpolation models based on FDTD simulations. The accuracies of the mathematical function approximations for the emulation models of power and SAR are 5.19% and 3.6% respectively. The RF signal strength is emulated by an analytical model.

show abstract

Empirical Method Based on Neural Networks for Analog Power Modeling

Cited by 22 publications

References 20 publications

Reinforcement Learning Strategies for Energy Management in Low Power IoT

Reinforcement Learning Strategies for Energy Management in Low Power IoT

LTE RF subsystem power consumption modeling

LTE modem power consumption, SAR and RF signal strength emulation

Contact Info

Product

Resources

About