SPEAKER: Victor Voronin, PhD student.
ABSTRACT: Formal verification of complex robotics behaviors suffers from scalability problems and limited acceptance throughout the industry. This talk presents the ongoing work to bridge this gap using statistical model checking technique. It brings up the advantages and limitations of this technique along with some of the use-cases where it can very well fit the development process of more robust robotic systems. A modular approach to represent the system under test is discussed, with its contribution to bypassing the scalability problem, testing ‘black-box’ systems and increasing the fidelity by using physical simulation in the verification loop.