06-02-2021 10:16 AM
I am working in the area of Fault-Tolerant AI processing for safety-critical applications. I would like to get your opinion on the topic of fault injection in FPGA design. Which method should i follow to do fault analysis of the FPGA design (in my case the design is a CNN)
I have CNN (convolutional neural network) implemented on ZCU102 FPGA. I would like to do a fault analysis of this CNN. And after fault analysis, I would like to work on different methods to make the CNN fault-tolerant. So there are these two methods which I am currently looking for fault injection
I would like to get some guidance on which one I should choose.
Please refer me any material/userguides/reference designs/links related to this ?
Are there any better suggestions/methods to do the fault analysis ?
some more info:
Vivado Version: 2019.1
FPGA platform : ZCU102 and VCU118.
Thanks a lot