Autonomous Vehicles Scenario Testing Framework and Model of Computation
Abstract:
Autonomous Vehicle (AV) technology has the potential to fundamentally transform
the automotive industry, reorient transportation infrastructure, and
significantly impact the energy sector. Rapid progress is being made in the core
artificial intelligence engines that form the basis of AV technology. However,
without a quantum leap in testing and verification, the full capabilities of AV
technology will not be realized. Critical issues include finding and testing
complex functional scenarios, verifying that sensor and object recognition
systems accurately detect the external environment independent of weather
conditions, and building a regulatory regime that enables accumulative learning.
The significant contribution of this article is to outline a novel methodology
for solving these issues by using the Florida Poly AV Verification Framework
(FLPolyVF).