Title

Distributed fault-tolerant classification using coding theory

Date of Award

2003

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Electrical Engineering and Computer Science

Advisor(s)

Pramod K. Varshney

Keywords

Fault-tolerant, Coding, Fading channels, Wireless sensors

Subject Categories

Electrical and Computer Engineering | Engineering

Abstract

In this dissertation, we develop two new approaches for distributed multiclass classification using coding theory. In the first approach, the distributed classification fusion using error correcting codes (DCFECC), binary decisions from local sensors, possibly in the presence of faults, are forwarded to the fusion center that determines the final classification result. Classification fusion in our approach is implemented via error correcting codes to incorporate fault-tolerance capability. This new approach not only provides an improved fault-tolerance capability but also reduces bandwidth requirements as well as computation time and memory requirements at the fusion center. Code matrix design is essential for the design of such systems. Two efficient code matrix design algorithms are proposed in this dissertation. The relative merits of both algorithms are also studied. Performance evaluation of the proposed approach in the presence of faults is provided. These results show significant improvement in fault-tolerance capability as compared with conventional parallel fusion networks. A hierarchical classification system is also designed to achieve additional fault-tolerance capability and a lower probability of decision error.

In the second case, data transmission through fading channels is considered during the design of distributed multiclass classification systems. We obtain a new fusion rule, the distributed classification fusion using error correcting codes--soft-decision decoding (DCFECC-SD) fusion rule, by integrating soft decoding into the distributed classification problem. The optimal fusion rule that considers the effect of fading channels and the multiclass equal gain combining fusion rule are also developed. Performance evaluation of these three fusion rules shows that the DCFECC-SD has the best fault-tolerance capability. The DCFECC-SD approach also allows the number of information bits per local decision sent to the fusion center to be adjusted according to the status of the communication channel and observation signals, while the code matrix in the fusion center is a fixed binary matrix. This feature enables more efficient bandwidth management.

Access

Surface provides description only. Full text is available to ProQuest subscribers. Ask your Librarian for assistance.

http://libezproxy.syr.edu/login?url=http://proquest.umi.com/pqdweb?did=765157921&sid=1&Fmt=2&clientId=3739&RQT=309&VName=PQD