A Reality Check for the Self-Driving Enthusiast ft. Sravan Puttagunta of Hyperspec AI

Unlike humans, autonomous vehicles cannot see, think, or act for themselves. Instead, they rely on high-definition maps to supply a pre-encoded context about the environment. Machines cannot improvise and adapt as quickly as humans or other biological systems, so safety concerns arise if a vehicle suddenly encounters an unplanned obstruction or outdated map information. That said, can we increase a car's reliance on real-time scene construction and reduce its dependency on prior maps and localization? If so, how might we go about it? 

 

To help us understand this further, we sat down with Sravan Puttagunta, Chief Executive Officer of Hyperspec AI. In this episode, we look at how autonomous vehicles can access more of the road network, why it's essential to consider what evolutionary biology has to say, and how to solve the challenges we encounter via a "first principles" approach. 

AutoVision News Website: https://www.autovision-news.com

Connect with us on LinkedIn: https://bit.ly/3OuDLyQ