Geza Koczian

Geza Koczian

Auto-generated profile, based on publicly available information

Geza Koczian

Access Network Group (ANG), School of Computer Science and Electronic Engineering, University of Essex, Colchester, U.K.

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

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

Representative Publications

View All Publications
14 September 2023 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Open access
FNFelix Ngobigha
Felix Ngobigha
GKGeza Koczian
Geza Koczian
SWStuart Walker
Stuart Walker
The next generation of railway customer-oriented services are expected to generate a large volume of data ( $\approx$ 10s of TB). As a result, passengers' applications, safety, security, and Internet-on-Board (IoB) sensors challenge current Train Communication Networks. With the present Ethernet Train Backbone (ETB) specification of just 100 Mbit/s in total, railway passenger services will not support intelligent, seamlessly connected and mobile media on-board trains. In this paper, we propose a novel ETB design with experimental results demonstrating end-to-end 10 Gbit/s and 40 Gbit/s throughput results over existing conducting media on commercial railway carriages. This is equivalent to 100 Mbit/s per user on real-world railway rolling stock and shows that standard RailCat 5e cabling and new rail-approved 10 Gbit/s ETB active nodes (switches) fully support emerging trends and future-proof ETB configurations.

1 January 2026 in IEEE Open Journal of the Communications Society
Verified
Journal Article
Open access
Michael ParkerMichael Parker
Michael Parker
GKGeza Koczian
Geza Koczian
MTManoj Thakur
Manoj Thakur
NBNabeel Nisar Bhat
Nabeel Nisar Bhat
JSJakob Struye
Jakob Struye
Jeroen FamaeyJeroen Famaey
Jeroen Famaey

26 August 2024 in IEEE Access
Verified
Journal Article
Open access
FNFelix Ngobigha
Felix Ngobigha
MEMatthew Evans
Matthew Evans
ICIain Cummines
Iain Cummines
GKGeza Koczian
Geza Koczian
SWStuart Walker
Stuart Walker
The migration of professional audio systems towards Ethernet-based networking allows optimisation software to maximise key performance indicators (KPIs) such as the power required for particular applications. In this study, we focus on minimising peak power requirements by the use of supercapacitors (SCs) for demand smoothing with Class D audio amplification of at least 600 W root mean square (rms) output. Depending on the content, a 600 W maximum r.m.s output music programme is shown to draw just 60 W (10% of peak rms) on average. Our studies used music programme statistical analysis and probability theory, which enabled the minimum value of SC for a predetermined voltage drop to be found exactly and hence minimise costs/size requirements. The use of energy storage in audio (and other contexts) is not new of course, but we believe the probabilistic calculation and use of unprecedented SC reservoir capacitance values is original. This power averaging technique then allows Power-over-Ethernet (PoE) to be used, particularly the IEEE802.3 bt type 4 variant and AES67 Ethernet over powerline. As commercial PoE systems are now available for this standard, we have constructed a complete 600 W r.m.s class D sound system which has type 4 PoE as its sole power input. Again, PoE sound systems are not new but the extension to very high powers has not been reported to the best of our knowledge. In the powerline context, we describe a first-of-its-kind, outdoor venue, 5000 W peak rms, Class D audio system with > 1 km of standard 5 A mains cable, AES67 audio transport over powerline and lithium-ion energy storage at the amplifier location. An added benefit is that solar assistance can be used.