Kayotee: A Fault Injection-based System to Assess the Safety and Reliability of Autonomous Vehicles to Faults and Errors. Jha, S., Tsai, T., Hari, S., Sullivan, M., Kalbarczyk, Z., Keckler, S. W, & Iyer, R. K 2018.
bibtex   
@Article{jha18kayotee,
  author    = {Jha, Saurabh and Tsai, Timothy and Hari, Siva and Sullivan, Michael and Kalbarczyk, Zbigniew and Keckler, Stephen W and Iyer, Ravishankar K},
  title     = {Kayotee: A Fault Injection-based System to Assess the Safety and Reliability of Autonomous Vehicles to Faults and Errors},
  year      = {2018},
  comment      = {* systematic FI into HW and SW components of autonomous driving systems (e.g., cars) * technology stack by nvidia * in closed-loop environment * inject transient faults (i.e., software state is being modified) * to assess error masking * fault injection points on different levels * CPU/GPU: transient faults in functional units * software: modify program output * workload-generation (i.e., traffic scenarios) * compare with fault-free execution * \# there aren't too much details in the paper (~4 pages)},
  file      = {:jha18kayotee - Kayotee_ A Fault Injection-based System to Assess the Safety and Reliability of Autonomous Vehicles to Faults and Errors.pdf:PDF},
  groups    = {fault injection},
  timestamp = {2019-02-18},
}

Downloads: 0