Vehicle-level testing isn’t enough to guarantee completesafety.Ensuring ultra dependable system isn’t possible be-cause of testing systems thoroughly is infeasible.The V”software development model has been used in automotiveindustry for a long time. It was one of the development ref-erence models incorporated with MISRA Guidelines morethan 20 years ago.
Though V framework reects acceptedpractices for automotive safety, autonomous vehicles level5 can trigger challenges to map technical aspects with Vapproach. The bigger challenge with autonomous vehicle isabsence of driver which means absence of external inputsand where driver can’t be countable for exceptional situ-ations.the computer must be ready to handle every faultwhich can occur anytime. Some of the technologies used inautonomous vehicles are purely statistical in nature.Fail operational system has been designed successfully inaerospace but is still dicult to test them for several rea-sons. The obvious one is that redundancy has to be pro-vided in case of failure of one component where another onecan take over.It is dicult to test combinations of exceptions occurringduring unusual operational conditions.
While testers cande ne some test cases, scalability of that testing is still ques-tionable due to the combinatorial explosion of exceptions,scenarios, and other factors involved. Fault injection androbustness testing had been mature technologies to assessthe performance of a system under exceptional conditions, also they can help avoid designer and tester blind spots.After testing is done, security software is implemented incars, software teams regularly update the software to avoidemerging potential new threat.
they need to coordinate theupdating process with third-party suppliers, along with thelong product life cycles of cars. And this process requiresconsistent procedures and strict enforcement of the distri-bution process and all of these are under responsibilities ofthe OEM.
