Date of Award

1998

Publication Type

Master Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

Keywords

Engineering, Electronics and Electrical.

Supervisor

Miller, W. C.,

Rights

info:eu-repo/semantics/openAccess

Abstract

The purpose of this thesis is to introduce and demonstrate an implementation of the intelligent sensor in a 0.8mu BiCMOS process. The intelligent sensor is a multilayer programmable optically coupled neural network which can be used for pattern recognition for many applications. The intelligent sensor contains an 8x8 array of photosensitive elements as the inputs to this multilayer neural network. The design presented here marks an improvement over an existing design done in a 1.2mu CAMS process Performance and area efficiency are both unproved as a result of the design practices used and the benefits of the BiCMOS process used for this design. The photosensitive element offers a great performance increase over the previous design by using a true device as opposed to a parasitic device. An optical cell test chip containing several optical cell designs was fabricated and tested. The results of these tests were used to choose the optical cell used in the intelligent sensor. As well, a new training method which was tailored for this intelligent sensor design was written. This new training algorithm reduces the error of normal training methods by trying to more closely match the actual network during training. As well, a complete graphical interface was designed for this training tool. Both a 5.0V design and a low power 3.3V design of the intelligent sensor were completed and fabricated. The low power design offers a great power savings over the 5.0V design. The building blocks for the designs were designed and characterized using Cadence 4.3 design tools and were simulated using Hspice. The two chips which were fabricated, were tested using several VLSI test tools and verified the simulation results achieved during the design process.Dept. of Electrical and Computer Engineering. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1998 .M33. Source: Masters Abstracts International, Volume: 39-02, page: 0565. Adviser: W. C. Miller. Thesis (M.A.Sc.)--University of Windsor (Canada), 1998.

Share

COinS