Inbar Fijalkow

Inbar Fijalkow

Auto-generated profile, based on publicly available information

Inbar Fijalkow

ETIS CY Cergy Paris Univ ENSEA CNRS F-95000 Cergy-Pontoise France

Research Focus & Metrics

Academic Career Length

22Years (2005-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 August 2021 in Energies
Verified
Journal Article
Open access
PDF file
LDLouis Desportes
Louis Desportes
IFInbar Fijalkow
Inbar Fijalkow
PAPierre Andry
Pierre Andry
We address the control of a hybrid energy storage system composed of a lead battery and hydrogen storage. Powered by photovoltaic panels, it feeds a partially islanded building. We aim to minimize building carbon emissions over a long-term period while ensuring that 35% of the building consumption is powered using energy produced on site. To achieve this long-term goal, we propose to learn a control policy as a function of the building and of the storage state using a Deep Reinforcement Learning approach. We reformulate the problem to reduce the action space dimension to one. This highly improves the proposed approach performance. Given the reformulation, we propose a new algorithm, DDPGαrep, using a Deep Deterministic Policy Gradient (DDPG) to learn the policy. Once learned, the storage control is performed using this policy. Simulations show that the higher the hydrogen storage efficiency, the more effective the learning.

8 November 2022 in IEEE Transactions on Communications
Verified
Journal Article
Closed access
JRJoachim Rosseel
Joachim Rosseel
VMValerian Mannoni
Valerian Mannoni
IFInbar Fijalkow
Inbar Fijalkow
VSValentin Savin
Valentin Savin
This paper investigates decoder diversity architectures for short low-density parity-check (LDPC) codes, based on recurrent neural network (RNN) models of the belief-propagation (BP) algorithm. We propose a new approach to achieve decoder diversity in the waterfall region, by specializing BP-RNN decoders to specific classes of errors, with absorbing set support. We further combine our approach with an ordered statistics decoding (OSD) post-processing step, which effectively leverages the bit-error rate optimization deriving from the use of the binary cross-entropy loss function. We show that a single specialized BP-RNN decoder combines better than BP with the OSD post-processing step. Moreover, combining OSD post-processing with the diversity brought by the use of multiple BP-RNN decoders, provides an efficient way to bridge the gap to maximum likelihood decoding.

11 April 2019 in IEEE Transactions on Signal Processing
Verified
Journal Article
Closed access
AMAlexandre Marcastel
Alexandre Marcastel
EBElena Veronica Belmega
Elena Veronica Belmega
PMPanayotis Mertikopoulos
Panayotis Mertikopoulos
IFInbar Fijalkow
Inbar Fijalkow
One of the key challenges in Internet of Things (IoT) networks is to connect many different types of autonomous devices while reducing their individual power consumption. This problem is exacerbated by two main factors: first, the fact that these devices operate in and give rise to a highly dynamic and unpredictable environment where existing solutions (e.g., water-filling algorithms) are no longer relevant; and second, the lack of sufficient information at the device end. To address these issues, we propose a regret-based formulation that accounts for arbitrary network dynamics : this allows us to derive an online power control scheme that is provably capable of adapting to such changes, while relying solely on strictly causal feedback . In so doing, we identify an important tradeoff between the amount of feedback available at the transmitter side and the resulting system performance: if the device has access to unbiased gradient observations, the algorithm's regret after $T$ stages is $\mathcal O(T^{-1/2})$ (up to logarithmic factors); on the other hand, if the device only has access to scalar, utility-based information, this decay rate drops to $\mathcal O(T^{-1/4})$ . The above is validated by an extensive suite of numerical simulations in realistic channel conditions, which clearly exhibit the gains of the proposed online approach over traditional water-filling methods.

16 October 2020 in IEEE Open Journal of the Communications Society
Verified
Journal Article
Open access
MBMouktar Bello
Mouktar Bello
ACArsenia Chorti
Arsenia Chorti
IFInbar Fijalkow
Inbar Fijalkow
WYWenjuan Yu
Wenjuan Yu
LMLeila Musavian
Leila Musavian
In this paper, we study the performance of an uplink non-orthogonal multiple access (NOMA) network under statistical quality of service (QoS) delay constraints, captured through each user’s effective capacity (EC). We first propose novel closed-form expressions for the EC in a two-user NOMA network and show that in the high signal-to-noise ratio (SNR) region, the “strong” NOMA user, referred to as U2, has a limited EC, assuming the same delay constraint as the “weak” user, referred to as U1. We demonstrate that for the weak user U1, OMA and NOMA have comparable performance at low transmit SNRs, while NOMA outperforms OMA in terms of EC at high SNRs. On the other hand, for the strong user U2, NOMA achieves higher EC than OMA at small SNRs, while OMA becomes more beneficial at high SNRs. Furthermore, we show that at high transmit SNRs, irrespective of whether the application is delay tolerant, or not, the performance gains of NOMA over OMA for U1, and OMA over NOMA for U2 remain unchanged. When the delay QoS of one user is fixed, the performance gap between NOMA and OMA in terms of total EC increases with decreasing statistical delay QoS constraints for the other user. Next, by introducing pairing, we show that NOMA with user-pairing outperforms OMA, in terms of total uplink EC. The best pairing strategies are given in the cases of four and six users NOMA, raising once again the importance of power allocation in the optimization of NOMA’s performance.

27 April 2020 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
MPMylene Pischella
Mylene Pischella
ACArsenia Chorti
Arsenia Chorti
IFInbar Fijalkow
Inbar Fijalkow
A new multiple access (MA) strategy, referred to as non orthogonal multiple access -Relevant (NOMA-R), allows selecting NOMA when this increases all individual rates, i.e., it is beneficial for both strong(er) and weak(er) individual users. This letter provides a performance analysis of the NOMA-R strategy in uplink networks with statistical delay constraints. Closed-form expressions of the effective capacity (EC) are provided in two-users networks, showing that the strong user always achieves a higher EC with NOMA-R. Regarding the networks sum EC, there are distinctive gains with NOMA-R, particularly under stringent delay constraints.

9 September 2019
Verified
Book Chapter
Closed access
PDF file
LDLouis Desportes
Louis Desportes
PAPierre Andry
Pierre Andry
IFInbar Fijalkow
Inbar Fijalkow
JDJérôme David
Jérôme David