@article{pervej_jin_dai_2023, title={Resource Constrained Vehicular Edge Federated Learning With Highly Mobile Connected Vehicles}, volume={41}, ISSN={["1558-0008"]}, DOI={10.1109/jsac.2023.3273700}, abstractNote={This paper proposes a vehicular edge federated learning (VEFL) solution, where an edge server leverages highly mobile connected vehicles’ (CVs’) onboard central processing units (CPUs) and local datasets to train a global model. Convergence analysis reveals that the VEFL training loss depends on the successful receptions of the CVs’ trained models over the intermittent vehicle-to-infrastructure (V2I) wireless links. Owing to high mobility, in the full device participation case (FDPC), the edge server aggregates client model parameters based on a weighted combination according to the CVs’ dataset sizes and sojourn periods, while it selects a subset of CVs in the partial device participation case (PDPC). We then devise joint VEFL and radio access technology (RAT) parameters optimization problems under delay, energy and cost constraints to maximize the probability of successful reception of the locally trained models. Considering that the optimization problem is NP-hard, we decompose it into a VEFL parameter optimization sub-problem, given the estimated worst-case sojourn period, delay and energy expense, and an online RAT parameter optimization sub-problem. Finally, extensive simulations are conducted to validate the effectiveness of the proposed solutions with a practical 5G new radio (5G-NR) RAT under a realistic microscopic mobility model.}, number={6}, journal={IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Pervej, Md Ferdous and Jin, Richeng and Dai, Huaiyu}, year={2023}, month={Jun}, pages={1825–1844} }
 @article{pervej_guo_kim_parsons_orlik_di cairano_menner_berntorp_nagai_dai_2022, title={Mobility, Communication and Computation Aware Federated Learning for Internet of Vehicles}, ISSN={["1931-0587"]}, DOI={10.1109/IV51971.2022.9827190}, abstractNote={While privacy concerns entice connected and automated vehicles to incorporate on-board federated learning (FL) solutions, an integrated vehicle-to-everything communication with heterogeneous computation power aware learning platform is urgently necessary to make it a reality. Motivated by this, we propose a novel mobility, communication and computation aware online FL platform that uses on-road vehicles as learning agents. Thanks to the advanced features of modern vehicles, the on-board sensors can collect data as vehicles travel along their trajectories, while the on-board processors can train machine learning models using the collected data. To take the high mobility of vehicles into account, we consider the delay as a learning parameter and restrict it to be less than a tolerable threshold. To satisfy this threshold, the central server accepts partially trained models, the distributed roadside units (a) perform downlink multicast beamforming to minimize global model distribution delay and (b) allocate optimal uplink radio resources to minimize local model offloading delay, and the vehicle agents conduct heterogeneous local model training. Using real-world vehicle trace datasets, we validate our FL solutions. Simulation shows that the proposed integrated FL platform is robust and outperforms baseline models. With reasonable local training episodes, it can effectively satisfy all constraints and deliver near ground truth multi-horizon velocity and vehicle-specific power predictions.}, journal={2022 IEEE INTELLIGENT VEHICLES SYMPOSIUM (IV)}, author={Pervej, Md Ferdous and Guo, Jianlin and Kim, Kyeong Jin and Parsons, Kieran and Orlik, Philip and Di Cairano, Stefano and Menner, Marcel and Berntorp, Karl and Nagai, Yukimasa and Dai, Huaiyu}, year={2022}, pages={750–757} }
 @article{pervej_tan_hu_2020, title={Artificial Intelligence Assisted Collaborative Edge Caching in Small Cell Networks}, ISSN={["2576-6813"]}, DOI={10.1109/GLOBECOM42002.2020.9322101}, abstractNote={Edge caching is a new paradigm that has been exploited over the past several years to reduce the load for the core network and to enhance the content delivery performance. Many existing caching solutions only consider homogeneous caching placement due to the immense complexity associated with the heterogeneous caching models. Unlike these legacy modeling paradigms, this paper considers heterogeneous content preference of the users with heterogeneous caching models at the edge nodes. Besides, aiming to maximize the cache hit ratio (CHR) in a two-tier heterogeneous network, we let the edge nodes collaborate. However, due to complex combinatorial decision variables, the formulated problem is hard to solve in the polynomial time. Moreover, there does not even exist a ready-touse tool or software to solve the problem. We propose a modified particle swarm optimization (M-PSO) algorithm that efficiently solves the complex constraint problem in a reasonable time. Using numerical analysis and simulation, we validate that the proposed algorithm significantly enhances the CHR performance when comparing to that of the existing baseline caching schemes.}, journal={2020 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM)}, publisher={IEEE}, author={Pervej, Md Ferdous and Tan, Le Thanh and Hu, Rose Qingyang}, year={2020} }
 @inproceedings{pervej_lin_2020, title={Dynamic power allocation and virtual cell formation for throughput-optimal vehicular edge networks in highway transportation}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85090279811&partnerID=MN8TOARS}, DOI={10.1109/ICCWorkshops49005.2020.9145348}, abstractNote={This paper investigates highly mobile vehicular networks from users' perspectives in highway transportation. Particularly, a centralized software-defined architecture is introduced in which centralized resources can be assigned, programmed, and controlled using the anchor nodes (ANs) of the edge servers. Unlike the legacy networks, where a typical user is served from only one access point (AP), in the proposed system model, a vehicle user is served from multiple APs simultaneously. While this increases the reliability and the spectral efficiency of the assisted users, it also necessitates an accurate power allocation in all transmission time slots. As such, a joint user association and power allocation problem is formulated to achieve enhanced reliability and weighted user sum rate. However, the formulated problem is a complex combinatorial problem, remarkably hard to solve. Therefore, fine-grained machine learning algorithms are used to efficiently optimize joint user associations and power allocations of the APs in a highly mobile vehicular network. Furthermore, a distributed single-agent reinforcement learning algorithm, namely SARL-MARL, is proposed which obtains nearly identical genie-aided optimal solutions within a nominal number of training episodes than the baseline solution. Simulation results validate that our solution outperforms existing schemes and can attain genie-aided optimal performances.}, booktitle={2020 IEEE International Conference on Communications Workshops, ICC Workshops 2020 - Proceedings}, publisher={IEEE}, author={Pervej, M.F. and Lin, S.-C.}, year={2020} }
 @article{pervej_lin_2020, title={Eco-Vehicular Edge Networks for Connected Transportation: A Distributed Multi-Agent Reinforcement Learning Approach}, DOI={10.1109/VTC2020-Fall49728.2020.9348507}, abstractNote={This paper introduces an energy-efficient, software-defined vehicular edge network for the growing intelligent connected transportation system. A joint user-centric virtual cell formation and resource allocation problem is investigated to bring eco-solutions at the edge. This joint problem aims to combat against the power-hungry edge nodes while maintaining assured reliability and data rate. More specifically, by prioritizing the downlink communication of dynamic eco-routing, highly mobile autonomous vehicles are served with multiple low-powered access points (APs) simultaneously for ubiquitous connectivity and guaranteed reliability of the network. The formulated optimization is exceptionally troublesome to solve within a polynomial time, due to its complicated combinatorial structure. Hence, a distributed multi-agent reinforcement learning (D-MARL) algorithm is proposed for eco-vehicular edges, where multiple agents cooperatively learn to receive the best reward. First, the algorithm segments the centralized action space into multiple smaller groups. Based on the model-free distributed Q learner, each edge agent takes its actions from the respective group. Also, in each learning state, a software-defined controller chooses the global best action from individual bests of the distributed agents. Numerical results validate that our learning solution achieves near-optimal performances within a small number of training episodes as compared with existing baselines.}, journal={2020 IEEE 92ND VEHICULAR TECHNOLOGY CONFERENCE (VTC2020-FALL)}, publisher={IEEE}, author={Pervej, Md Ferdous and Lin, Shih-Chun}, year={2020} }
 @article{pervej_tan_hu_2020, title={User Preference Learning-Aided Collaborative Edge Caching for Small Cell Networks}, ISSN={["2576-6813"]}, DOI={10.1109/GLOBECOM42002.2020.9322208}, abstractNote={While next-generation wireless networks intend leveraging edge caching for enhanced spectral efficiency, quality of service, end-to-end latency, content sharing cost, etc., several aspects of it are yet to be addressed to make it a reality. One of the fundamental mysteries in a cache-enabled network is predicting what content to cache and where to cache so that high caching content availability is accomplished. For simplicity, most of the legacy systems utilize a static estimation - based on Zipf distribution, which, in reality, may not be adequate to capture the dynamic behaviors of the contents popularities. Forecasting user’s preferences can proactively allocate caching resources and cache the needed contents, which is especially important in a dynamic environment with real-time service needs. Motivated by this, we propose a long short-term memory (LSTM) based sequential model that is capable of capturing the temporal dynamics of the users’ preferences for the available contents in the content library. Besides, for a more efficient edge caching solution, different nodes in proximity can collaborate to help each other. Based on the forecast, a non-convex optimization problem is formulated to minimize content sharing costs among these nodes. Moreover, a greedy algorithm is used to achieve a sub-optimal solution. Using extensive simulation and analysis, we validate that the proposed algorithm performs better than other existing schemes.}, journal={2020 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM)}, publisher={IEEE}, author={Pervej, Md Ferdous and Tan, Le Thanh and Hu, Rose Qingyang}, year={2020} }
 @inproceedings{roy_pervej_morshed_2016, title={Performance comparison of three optimized alternative pulse shaping filters with the raised cosine filter for wireless applications}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84966455395&partnerID=MN8TOARS}, DOI={10.1109/CCIE.2015.7399305}, abstractNote={This paper is concerned with the performance and cost analysis of different pulse shaping filters such as raised cosine, finite impulse response (FIR) Nyquist, FIR half-band and infinite impulse response (IIR) linear phase half-band for wireless applications. Note that the pulse shaping is the process of changing the waveform of transmitting pulses to reduce the interferences keeping a signal in an allotted bandwidth. This paper proposes an optimal design for the pulse shaping filters to reduce their effective costs in the area of wireless communication applications. Here, we analyze the performance of pulse shaping filters based on the minimum stop band attenuation required to suppress the inter-symbol interference (ISI) and inter-channel interference (ICI). Finally, we compare the performance of raised cosine, FIR Nyquist, FIR half-band and IIR linear phase half-band filters in terms of bit-error rate (BER) and hardware requirements. Our results demonstrate that the hardware implementation costs for the FIR Nyquist, FIR half-band and IIR half-band filters are about 60% to 90% less than the raised cosine filter.}, booktitle={1st International Conference on Computer and Information Engineering, ICCIE 2015}, author={Roy, T.K. and Pervej, M.F. and Morshed, M.}, year={2016}, pages={9–12} }
 @inproceedings{hossen_chowdhury_anower_hossen_pervej_hasan_2015, title={Effect of signal length in cross-correlation based underwater network size estimation}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84962182191&partnerID=MN8TOARS}, DOI={10.1109/ICEEICT.2015.7307476}, abstractNote={Signal length possesses a very important role in size estimation of underwater wireless sensor network (UWSN). As size estimation is very tough in UWSNs using conventional protocol techniques, a cross-correlation based technique is introduced to estimate the number of nodes. In UWSN, The greater the signal length, the more energy is required to perform the estimation. In this paper, we observe the effect of signal length (Ns) over the number of estimated nodes (N) in a spherical region of an underwater acoustic sensor network (UASN) using three sensors and investigate the error in node estimation for different Ns. As the theoretical required Ns is infinity (Ns = 106 samples is used in simulation), using minimum Ns how it is possible to go through a successful estimation process i.e., accurate node estimation is being observed and discussed.}, booktitle={2nd International Conference on Electrical Engineering and Information and Communication Technology, iCEEiCT 2015}, author={Hossen, M.A. and Chowdhury, S.A.H. and Anower, M.S. and Hossen, S. and Pervej, M.F. and Hasan, M.M.}, year={2015} }
 @inproceedings{pervej_sarkar_islam_2015, title={Impact analysis of different inputs and outputs block sizes of DFT-SCFDMA system}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84962159109&partnerID=MN8TOARS}, DOI={10.1109/ICEEICT.2015.7307372}, abstractNote={The Bit Error Rate (BER) and Peak to Average Power Ratio (PAPR) are the major concerns for wireless communication. This paper aims to analyze these two important parameters for different inputs and outputs block sizes of the discrete Fourier transform (DFT) based single carrier frequency division multiple access (SCFDMA) system by using the conventional raised cosine (RC) and square root raised cosine (RRC) pulse shaping filters for different sub-carrier mapping schemes (e.g. interleaved frequency division multiple access (IFDMA), localized frequency division multiple access (LFDMA) and distributed frequency division multiple access (DFDMA)). Note that DFT based SCFDMA system is being adopted for uplink communication in the long term evolution (LTE) for its lower PAPR. The suitable block size of SCFDMA is crucial to improve wireless communication performances in terms of BER and PAPR, because if the block sizes vary then the PAPR also changes. Therefore, this paper analyzes the impact of the input and output block sizes of the system on enhancement of the system performances. Taking help from the numerical analysis and evaluation, this paper shows the way to choose the proper input and output block size to have a significantly lower PAPR and BER.}, booktitle={2nd International Conference on Electrical Engineering and Information and Communication Technology, iCEEiCT 2015}, author={Pervej, M.F. and Sarkar, M.Z.I. and Islam, M.T.}, year={2015} }
 @inproceedings{pervej_sarkar_roy_hasan_rahman_bain_2003, title={Analysis of PAPR reduction of DFT-SCFDMA system using different sub-carrier mapping schemes}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84949924382&partnerID=MN8TOARS}, DOI={10.1109/ICCITechn.2014.7073067}, abstractNote={This paper is concerned with the reduction of peak to average power ratio (PAPR) of 3rd generation partnership project (3GPP) LTE standard using the single carrier-frequency division multiple access (SCFDMA) for mobile uplink transmission. The PAPR reduction is done based on the discrete Fourier transform (DFT) spreading technique which is well known as DFT - SCFDMA and in case of LTE which can reduce the PAPR of OFDM signal to a level of single carrier transmission. Note that the needed power of the signals for uplink transmission of LTE is varied depending on the different sub-carrier mapping and modulation techniques. This paper discusses different sub-carrier mapping techniques such as-localized FDMA (IFDMA), distributed FDMA (DFDMA) and interleaved FDMA (IFDMA). To show the effectiveness of DFT-SCFDMA in the reduction of PAPR with different sub-carrier mapping and modulation techniques the results are compared with the conventional OFDMA system. This paper also shows the comparative results of different sub-carrier mapping techniques in terms of symbol error rate (SER). The comparative results show that the DFT-SCFDMA greatly enhances the PAPR reduction for LTE uplink transmission over the conventional OFDMA.}, booktitle={2014 17th International Conference on Computer and Information Technology, ICCIT 2014}, author={Pervej, M.F. and Sarkar, M.Z.I. and Roy, T.K. and Hasan, M.M. and Rahman, M.M. and Bain, S.K.}, year={2003}, pages={435–439} }
 @inproceedings{hasan_paul_haque_chowdhury_pervej_hossen_2003, title={Effect of node number in range based node localization technique for underwater communications network}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84949924869&partnerID=MN8TOARS}, DOI={10.1109/ICCITechn.2014.7073099}, abstractNote={In underwater network, it is very important to know the location of unmanned underwater vehicle (UUV). The unknown position of UUV considered as an unknown node that can be determined from the known node locations through underwater acoustic network. The known node locations considered here as fixed positioned node. The range data is measured between unknown node and known nodes. By exploiting these measured range data and combination among fixed positioned nodes on this network, it is possible to perform underwater positioning of unknown node (UUV), similar to that of the satellite-based GPS program. The aim of this paper is to determine the position of UUV in an underwater acoustic network using combination method with different number of nodes, whose locations are known, to analyze the effect of node number in position estimation.}, booktitle={2014 17th International Conference on Computer and Information Technology, ICCIT 2014}, author={Hasan, M.M. and Paul, A. and Haque, S.S. and Chowdhury, S.A.H. and Pervej, M.F. and Hossen, M.A.}, year={2003}, pages={413–417} }
 @inproceedings{pervej_sarkar_roy_koli_2003, title={Impact analysis of input and output block size of DCT-SCFDMA system}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84949924614&partnerID=MN8TOARS}, DOI={10.1109/ICCITechn.2014.7073068}, abstractNote={This paper aims to analyze the bit error rate (BER) and peak to average power ratio (PAPR) for discrete cosine transform (DCT) based single carrier frequency division multiple access (SCFDMA) system. Note that BER and PAPR are the major concerns for any wireless communication and for having lower PAPR till now discrete Fourier transform (DFT) based SCFDMA is implemented for the up-link communication for long term evolution (LTE). This paper focuses to analyze the effect of the input and output block size of DCT based SCFDMA system which is a supplementary of the DFT-SCFDMA system. Using the conventional raised cosine (RC) and square root raised cosine (RRC) pulse shaping filters with different sub-carrier mapping schemes (e.g. interleaved frequency division multiple access (IFDMA), localized frequency division multiple access (LFDMA) and distributed frequency division multiple access (DFDMA)) the paper analyze the PAPR for the DCT based SCFDMA system. The suitable block size of DCT-SCFDMA is very important to improve the BER and PAPR performances of the system because if the block sizes vary then the PAPR also changes. Therefore, this paper also analyzes the impact of the input and output block size of the system. Finally, the numerical evaluation is done using the pulse shaping filter for different sub-carrier mapping schemes based on the different block sizes. From the numerical analysis and evaluation this paper shows the way to chose the input and output block size to have a significantly low PAPR and BER.}, booktitle={2014 17th International Conference on Computer and Information Technology, ICCIT 2014}, author={Pervej, M.F. and Sarkar, M.Z.I. and Roy, T.K. and Koli, M.N.Y.}, year={2003}, pages={440–445} }
 @inproceedings{bain_chowdhury_asif_anower_pervej_haque_2003, title={Impact of underwater bandwidth on cross-correlation based node estimation technique}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84949924255&partnerID=MN8TOARS}, DOI={10.1109/ICCITechn.2014.7073137}, abstractNote={Bandwidth is the most monumental topic both for the terrestrial and underwater communications. For proper network operation the estimation of number of signal sources (N) is very important. Most of the protocol based techniques are failed to give the desired results of node estimation due to underwater properties (long propagation delay, high absorptions and dispersion). Node estimation technique using cross-correlation is not based on protocol, where infinite band Gaussian signals are used. But the underwater bandwidth is finite (1-15 kHz). This limited bandwidth can encumber the signal bandwidth which might invade the estimation process. In this paper, the number of estimated nodes is being obtained considering the effect of underwater bandwidth using three sensors.}, booktitle={2014 17th International Conference on Computer and Information Technology, ICCIT 2014}, author={Bain, S.K. and Chowdhury, S.A.H. and Asif, A.H.M. and Anower, M.S. and Pervej, M.F. and Haque, S.S.}, year={2003}, pages={494–497} }
 @inproceedings{koli_sarkar_tazrin_pervej_2003, title={On the optimum secrecy capacity of multiple relay networks}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84949924107&partnerID=MN8TOARS}, DOI={10.1109/ICCITechn.2014.7073091}, abstractNote={We consider a confidential communication system in which a source sends a confidential information to the destination in the presence of an eavesdropper. Multiple relays are used to provide cooperative diversity to the destination. The destination and the eavesdropper are equipped with multiple antennas while each relay and source are equipped with single antenna. We are interested to protect the transmitted information from eavesdropping and to find the effect of receive diversity on the secrecy capacity of the proposed model. We consider the maximal ratio combining (MRC) technique at the destination and derive the expression of combiner output signal-to-noise ratio (SNR) using optimum weighting vector so that the eavesdropper is unable to decode a single bit from the original massage. Finally, we derive the expression of optimum secrecy capacity in terms of the number of relays and the number of antennas at the destination. Our results show that the optimization of weighting vector at the receiver enhances secrecy capacity of a wiretap channel.}, booktitle={2014 17th International Conference on Computer and Information Technology, ICCIT 2014}, author={Koli, M.N.Y. and Sarkar, M.Z.I. and Tazrin, S. and Pervej, M.F.}, year={2003}, pages={514–517} }