While simulation must be used there are several issues that need to be addressed.

It is 500 billion to one trillion mils of shadow driving needed not 11 billion. I believe you got 11 billion from RAND? That is the study where they said it is impossible to shadow drive the miles. (At an expense I conservatively calculate at $300B. And you also have the issue with causing many needless casualties when thousands of complex and dangerous scenarios are run thousands of times each.) That 11B is for the system to be only 20% better than a human. If that is extrapolated out to the 10X folks think is needed that is 500B miles. Toyota said 1 trillion miles. Nonetheless you have to use simulation.

However the simulation that is needed is the aerospace/DoD type. Not what is being used by AV makers or OEMs. Those systems have significant gaps. Many that will lead to crippling false confidence. See my article here for more. (Cloud based AV sensors systems are especially concerning). Do yourselves a BIG favor. Stop thinking this industry is the end-all-be-all of tech, that no one else has done a thing and you innovate everything. the FACT is DoD had the sim tech needed to do this right 20 years ago. I know because I witnessed it. (And don’t think commercial air. DoD war-games involved many entities donated from simulations and simulators in a real real time network with FAA level D simulators. Many running models more complicated than vehicles run.)

The Dangers of Inferior Simulation for Autonomous Vehicles


Written by

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store