« Назад

Outcomes of training at City University London

 

Name:  Liubov Reva, Master student.

Home University:  National Aerospace University named after N.E.Zhukovsky "KhAI".

Place of traning:  City University London, Center of Software Reliability, London, Great Britain

Date of traning: February 13 – March 16, 2012

Supervisors:  Dr. Peter Popov.

The task: 

The research problem was defined at the beginning of the internship. It consisted in search, analysis and application of ways to increase the performance of existing Bayesian reliability assessment system. The decision was made to change the implementation of method from Matlab program to digital design for FPGA.  
Taking into account the variety of possible ways to implement the method following tasks were set:

1. To carry out the comparative analysis of FPGA circuits characteristics, micro-controllers and ASIC.
2. To justify FPGA implementation selection based on results of analysis.
3. To make a survey on ways of FPGA-based system development.
4. To make a survey on ways of performance enhancement for FPGA-based systems.
5. To select and apply one or more methods of performance enhancement.
6. To implement the Bayesian reliability  assessment system using selected design flow.
7. To carry out the simulation and analyze time characteristics of obtained FPGA-based implementation of Bayesian assessment.
8. To analyze the results and make conclusions.

The results of traning: the research work was made according to the steps presented above. The following decisions were made:
To implement and analyze the system into two steps:

1. VHDL-model development and verification.
2. SoC (System on Chip) development and validation.

To use the pipe-lining as the effective way of increasing the system performance.
The proposed flow of development and analysis created the basis for  the second part of the work. At the development stage the part of HDL-based model was designed. The unfinished  work became the part of MS graduation thesis and will be completed during its execution. The results will be showed at the graduation thesis presentation

 

Среднее (0 Голоса)
Средний рейтинг 0.0 звезд из 5.


Пока нет комментариев. Будь первым.