Back to search

National and Kapodistrian University of Athens

In-network processing based hardware acceleration for situational awareness

Abstract

dc:description

WSN and IoT are the two key information technologies in modern situational based awareness system. Although advances in information technology, current situational awareness operation is shifting from traditional large command and control centers to small groups working together in distributed manner through the use of information technology. Data collection to central databases, as used in typical WSNs for monitoring, creates overhead, potential bottlenecks and offers limited resilience. In real-time communication, situational information needs to deliver in a timely manner and would prefer to receive information tailored to their current information needs directly from the network to minimize delays and dependence on central infrastructure.In this research, an In-network processing (INP) based computational framework has been proposed that gradually refine the captured information while moving it upstream to the application. The refinement can go up to the point of knowledge extraction from the primitive or even fused data. It shows that the demanding tasks that previously were simply undertaken on the fixed infrastructure are now possible on the mobile end. It relies on a sophisticated in-network processing scheme that gradually refines the information captured by sensing elements to the level of application-exploitable knowledge. With this process, the scaling problem is tackled much more efficiently through a problem segmentation and exploitation of physics-based and human-based sources.Energy is the dominant constraint because the quantities of information to process are much larger and the processing algorithms are much heavier. In-network processing (INP) is a useful technique to reduce the energy consumption significantly. The components executing on INP nodes are structured appropriately so as to delegate the demanding subtasks to some onboard accelerator. The core program executes on the central processing unit and exploits the particular characteristics of the side processor. Special system-on-chip (SoC) hardware for computational acceleration like central processor unit (CPU), digital signal processor (DSP) or field-programmable gate arrays (FPGAs) are desirable. These accelerated hardware supports software defined process on-chip memory that expedites all the advanced processing with a minimum energy overhead. In-network software defined processing (ISDP) capabilities has been considered that allows dynamic reconfiguration of the network topology. It is advantageous for the network to possess reliable and complete end-to-end network connectivity; however, even when the network is not fully connected, the system may act as conduits of information — either by connectivity gaps, or by distributing information from the network space.To validate the argument, a real maintenance work at ATLAS detector in Large Hadron Collider at European Organization for Nuclear Research (CERN) has been considered. The challenges lie in the development of real-time data-transmission, instantaneous analysis of data coming from different inputs, local intelligences in low power embedded system, interaction with multiple on-site users, complex interfaces, modularity and portability. An integrated prototype system with augmented and virtual reality based information management system has been developed. The collaborative execution of INP nodes also provides system modularity and improves the network security and privacy.The results of this research are interesting for the scientific and industrial communities and also very appealing for the general public. The knowledge of this research will be helpful to the students, learners, researchers. It will support the domains of on-board decision support systems, context- and situation-awareness. Achieved results influences further optimization of the INP architecture for implicit parallelism and are all fairly similar in computational models, syntax, and ease of use. The thesis discusses different low power accelerated hardware-software solutions for INP process however; there are several issues that need to be addressed before it can offer a safe and secure communication over wireless sensor and actuator networks. Other important features, for example, interoperability between different wireless systems, availability and meeting real-time guarantees where latency is an important parameter to pay attention to. Further investigations will be carried out in these domains.

Degree

thesis:*
Grantor dc:publisher
National and Kapodistrian University of Athens
Year dc:date
2018

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Alam, Md Fasiul
  • Αλάμ, Φασιούλ

Subjects

dc:subject × 11

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/44547

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Alam, Md Fasiul; Αλάμ, Φασιούλ. In-network processing based hardware acceleration for situational awareness. National and Kapodistrian University of Athens, 2018. http://hdl.handle.net/10442/hedi/44547