Md Saikat Ahmed

Mr. Md Saikat Ahmed

Technology, Illinois State University, Normal, Illinois, USA
orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

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

Representative Publications

View All Publications
13 October 2026 in Computers
Verified
Journal Article
Open access
PDF file
Md Tahmid RashidMd Tahmid Rashid
Md Tahmid Rashid
Md Saikat AhmedMd Saikat Ahmed
Md Saikat Ahmed
Nafiz ImtiazNafiz Imtiaz
Nafiz Imtiaz
Federated Artificial Intelligence of Things (AIoT) activity-recognition systems learn from decentralized wearable and smart-home records with potentially incorrect labels, non-IID client distributions, and unequal edge-device resources. Our objective is to improve downstream learning from these labels while limiting harmful corrections and respecting auditing, review, and communication budgets. FedCuraLM addresses this objective by converting sensor records into deterministic Semantic Evidence Cards, obtaining candidate-constrained small language model (SLM) assessments, and calibrating and combining them with task-model, prototype, and temporal evidence. Conflict-aware decisions retain observed labels, apply provisional soft corrections, request trusted review, or quarantine uncertain samples without overwriting annotations. Review-grounded reliability and communication-budgeted multi-prototype exchange support collaboration, while resource-aware scheduling assigns feasible SLM tiers across heterogeneous devices. Extensive evaluation on WISDM-W, WISDM-P, and CASAS using controlled symmetric, semantic-asymmetric, instance-dependent, and temporal-boundary corruption under varying client heterogeneity reveals that FedCuraLM significantly improves downstream performance over recent baselines and improves lower-tail client outcomes, audit precision, selective-correction safety, calibration, and communication efficiency. Edge-device measurements characterize latency, memory, and energy trade-offs. The findings position federated label curation as a semantic, statistical, communication, and systems problem.

13 October 2026 in Computers
Verified
Journal Article
Open access
PDF file
Md Tahmid RashidMd Tahmid Rashid
Md Tahmid Rashid
Md Saikat AhmedMd Saikat Ahmed
Md Saikat Ahmed
MSMd Jawad Siddique
Md Jawad Siddique
Multi-UAV tracking can extend coverage by combining aerial observations with reports from fixed monitors and mobile sources. Intermittent links and heterogeneous sources nevertheless create delayed, duplicated, and unreliable evidence, while broad sharing consumes limited bandwidth and energy. This article presents RAI-CoS, a reliability-calibrated and Age-of-Information-aware collaborative sensing framework for this coupled estimation and communication problem. The framework combines context-conditioned source reliability and innovation consistency with generation-time out-of-sequence processing, provenance-aware fusion, receiver-specific information valuation, virtual-queue scheduling, and recurrent decentralized control. We evaluate RAI-CoS in road-constrained multi-target simulations with replayed network traces and supplement the evaluation with Jetson Orin NX decision-stack trace replay. Under the evaluated conditions, RAI-CoS preserves useful tracks through source and network disturbances while reducing unnecessary communication and energy demand. Calibration, fusion, and component analyses connect these outcomes to conservative source assessment, physical-time reconstruction of delayed evidence, correlation-aware fusion, and destination-specific packet selection. Transfer, scaling, and failure analyses further identify where these mechanisms remain effective and where adaptation is required. The findings show that resilient collaborative tracking benefits from treating reliability, time, provenance, and communication cost as one information-control problem.