According to Appendix A in the paper, we introduce the key 00 11 22 33 44 55 66 77 88 99 AA BB CC DD EE FF for every case study, as the AES test vectors are specified. However, I recall that the Verilog implementation of "instructionreplacementtrojan.sv" we used a different key for test purposes, namely F0 E0 D0 C0 B0 A0 90 80 70 60 50 40 30 20 10 00 that is more different to the original key and thus well-suited as a proof of concept. It should though not be a big deal to replace the trojan-key to the other one and test that the implementation still works as expected, it should result in the test vectors given in A.3.
According to Appendix A in the paper, we introduce the key
00 11 22 33 44 55 66 77 88 99 AA BB CC DD EE FFfor every case study, as the AES test vectors are specified. However, I recall that the Verilog implementation of "instructionreplacementtrojan.sv" we used a different key for test purposes, namelyF0 E0 D0 C0 B0 A0 90 80 70 60 50 40 30 20 10 00that is more different to the original key and thus well-suited as a proof of concept. It should though not be a big deal to replace the trojan-key to the other one and test that the implementation still works as expected, it should result in the test vectors given in A.3.