Kaushik Chowdhury

Kaushik Chowdhury

Auto-generated profile, based on publicly available information

Kaushik Chowdhury

The University of Texas at Austin

Research Focus & Metrics

Academic Career Length

21Years (2006-Present)

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

Representative Publications

View All Publications
11 May 2018 in IEEE Transactions on Network Science and Engineering
Verified
Journal Article
Closed access
WLWei Li
Wei Li
FZFan Zhou
Fan Zhou
KCKaushik Roy Chowdhury
Kaushik Roy Chowdhury
WMWaleed M. Meleis
Waleed M. Meleis
Next generation network access technologies and Internet applications have increased the challenge of providing satisfactory quality of experience for users with traditional congestion control protocols. Efforts on optimizing the performance of TCP by modifying the core congestion control method depending on specific network architectures or apps do not generalize well under a wide range of network scenarios. This limitation arises from the rule-based design principle, where the performance is linked to a pre-decided mapping between the observed state of the network to the corresponding actions. Therefore, these protocols are unable to adapt their behavior in new environments or learn from experience for better performance. We address this problem by integrating a reinforcement-based Q-learning framework with TCP design in our approach called QTCP. QTCP enables senders to gradually learn the optimal congestion control policy in an on-line manner. QTCP does not need hard-coded rules, and can therefore generalize to a variety of different networking scenarios. Moreover, we develop a generalized Kanerva coding function approximation algorithm, which reduces the computation complexity of value functions and the searchable size of the state space. We show that QTCP outperforms the traditional rule-based TCP by providing 59.5% higher throughput while maintaining low transmission latency.

12 June 2018 in IEEE Transactions on Robotics
Verified
Journal Article
Closed access
Angelo TrottaAngelo Trotta
Angelo Trotta
MFMarco Di Felice
Marco Di Felice
Federico MontoriFederico Montori
Federico Montori
KCKaushik R. Chowdhury
Kaushik R. Chowdhury
Luciano BononiLuciano Bononi
Luciano Bononi
This paper proposes deployment strategies for consumer unmanned aerial vehicles (UAVs) to maximize the stationary coverage of a target area and to guarantee the continuity of the service through energy replenishment operations at ground charging stations. The three main contributions of our work are as follows. 1) A centralized optimal solution is proposed for the joint problem of UAV positioning for a target coverage ratio and scheduling the charging operations of the UAVs that involves travel to the ground station. 2) A distributed game-theory-based scheduling strategy is proposed using normal-form games with rigorous analysis on performance bounds. Furthermore, a bio-inspired scheme using attractive/repulsive spring actions are used for distributed positioning of the UAVs. 3) The cost-benefit tradeoffs of different levels of cooperation among the UAVs for the distributed charging operations is analyzed. This paper demonstrates that the distributed deployment using only 1-hop messaging achieves approximation of the centrally computed optimum, in terms of coverage and lifetime.

2 December 2020 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
NSNasim Soltani
Nasim Soltani
GRGuillem Reus-Muns
Guillem Reus-Muns
BSBatool Salehihikouei
Batool Salehihikouei
JDJennifer Dy
Jennifer Dy
SIStratis Ioannidis
Stratis Ioannidis
KCKaushik Chowdhury
Kaushik Chowdhury
The universal availability of unmanned aerial vehicles (UAVs) has resulted in many applications where the same make/model can be deployed by multiple parties. Thus, identifying a specific UAV in a given swarm in a manner that cannot be spoofed by software methods becomes important. We propose RF fingerprinting for this purpose, where a neural network learns subtle imperfections present in the transmitted waveform. For UAVs, the constant hovering motion raises a key challenge, which remains a fundamental problem in previous works on RF fingerprinting: Since the wireless channel changes constantly, the network trained with a previously collected dataset performs poorly on the test data. The main contribution of this paper is to address this problem by: (i) proposing a multi-classifier scheme with a two-step score-based aggregation method, (ii) using RF data augmentation to increase neural network robustness to hovering-induced variations, and (iii) extending the multi-classifier scheme for detecting a new UAV, not seen earlier during training. Importantly, our approach permits RF fingerprinting on manufacturer-proprietary waveforms that cannot be decoded or altered by the end-user. Results reveal a near two-fold accuracy in UAV classification through our multi-classifier method over the single-classifier case, with an overall accuracy of 95% when tested with data under unseen channel. Our multi-classifier scheme also improves new UAV detection accuracy to a near perfect 99%, up from 68% for a single neural network approach.

7 March 2018 in IEEE Internet of Things Journal
Verified
Journal Article
Open access
UMUfuk Muncuk
Ufuk Muncuk
KAKubra Alemdar
Kubra Alemdar
JSJayesh D. Sarode
Jayesh D. Sarode
KCKaushik Roy Chowdhury
Kaushik Roy Chowdhury
Ambient RF energy harvesting allows powering low-power electronic devices without wires, batteries, and dedicated energy sources. Current RF energy harvesting circuit designs for ambient RF harvesting are optimized and fabricated for a predetermined frequency band. Thus, a single circuit is tuned for a given band with simple extensions to multiple circuits operating individually in distinct bands. Our approach is different in the sense that it designs and implements a common circuit design that can operate on multiple different RF cellular and ISM bands. This paper makes two contributions. First, it presents a study of ambient RF signal strength distribution conducted in Boston city, indicating locations and associated RF bands that can point towards the practicality of ambient RF energy harvesting. Second, it demonstrates an adjustable circuit for harvesting from LTE 700 MHz, GSM 850 MHz, and ISM 900 MHz bands with one single circuit. Our circuit design is fabricated on printed circuit board with comprehensive evaluations at each associated frequency to test the power conversion efficiency. In addition, we characterize the charging performance, and feasibility of powering sensors outdoors such as TIez430-RF2500. Results reveal more than 45% power conversion efficiency (PCE) for our prototype.

18 June 2018 in IEEE Communications Magazine
Verified
Journal Article
Open access
ACAngela Sara Cacciapuoti
Angela Sara Cacciapuoti
KSKunal Sankhe
Kunal Sankhe
MCMarcello Caleffi
Marcello Caleffi
KCKaushik Roy Chowdhury
Kaushik Roy Chowdhury
This article presents a network architecture for the next generation of MHNs, where mmW, terahertz, and conventional mW bands coexist, with cost-benefit trade-offs of each type of link. We envision a radically different communication paradigm and outline a MAC protocol design that switches among the aforementioned bands for data transmissions, falling back on the slower link each time for the reverse channel ACKs. The use of the higher-capacity link in the forward direction for data communication and the slower reverse channel for the returning ACKs allows for uninterrupted unidirectional communication and efficient use of the channel. The article discusses the challenges in analyzing and parameter setting for the various features of the protocol, and identifies candidate solutions. A performance evaluation of the approach is undertaken using a realistic scenario of vehicle-to-infrastructure communication enabling data center traffic backhauling. This performance shows that by adopting the proposed MAC protocol data transfer of around 100 Terabits, it is possible for typical vehicle-to-infrastructure contact times.

25 January 2022 in IEEE Transactions on Multimedia
Verified
Journal Article
Closed access
DRDebashri Roy
Debashri Roy
YLYuanyuan Li
Yuanyuan Li
TJTong Jian
Tong Jian
PTPeng Tian
Peng Tian
KCKaushik Roy Chowdhury
Kaushik Roy Chowdhury
SIStratis Ioannidis
Stratis Ioannidis
With the recent surge in autonomous driving vehicles, the need for accurate vehicle detection and tracking is critical now more than ever. Detecting vehicles from visual sensors fails in non-line of sight (NLOS) settings. This can be compensated by the inclusion of other modalities in an multi-domain sensing environment. In this paper, we propose different deep learning based frameworks for fusing different modalities (image, radar, acoustic, seismic) through the exploitation of complementary latent embeddings among different modalities, and incorporating different state-of-the-art fusion strategies. Our proposed fusion frameworks perform considerably better than unimodal detection. Moreover, fusion between image and non-image modalities helps to track vehicles, as well as detect them, when direct line of sight (LOS) is not available. We validate our models on the real-world multimodal ESCAPE dataset, showing 33.16% improvement in vehicle detection performance by fusion (over visual inference alone) while 30-42% of the test scenarios have vehicles in NLOS. To showcase the generalization aspect of the proposed frameworks over multiple datasets, we also validate our models on multimodal NuScene dataset, showing ~22% improvement than the competing methods.