Back to Home

Seminar Talk

Vision-based Autonomy Systems for Drones

SpeakerDr. Kalpana Seshadrinathan
AffiliationHead of Deep Learning at Skydio Inc., San Jose, CA, USA.
Date23 June 2026
Time11:30 AM
VenueEE Seminar Hall

Abstract

Skydio is the leading US drone company and the world leader in autonomous flight. Skydio leverages breakthrough AI to create the world’s most intelligent flying machines for use by consumers enterprises and government customers. Skydio drones are used to inspect the critical infrastructure that our society depends on provide situational awareness to public safety operators during active situations or emergency response conduct reconnaissance missions during strategic military operations to name just a few use cases. At the core of our products is a vision-based autonomy system which pushes the state of the art in deep learning computer vision planning and controls with a particular focus on real-world robustness. Skydio drones are built from the ground up to take the burden of safe flight off of the pilot based on our AI-driven engine Skydio Autonomy™. Each drone carries six fisheye navigation cameras and high-resolution user facing visual and radiometric thermal cameras along with powerful onboard compute capabilities to enable the drone to see and understand its surroundings plan a path through them avoid obstacles in any direction and track subjects seamlessly. Drones encounter extreme visual scenarios not typically considered by academia nor encountered by cars ground robots or AR applications that they must deftly navigate through. In this talk I will discuss Skydio's technology stack with a particular focus on the Skydio Autonomy engine. In particular I will discuss Skydio's vision based obstacle avoidance system for 24/7 operations in day and night the Skydio Shadow system for tracking and following subjects and other autonomous capabilities that bring the skills of an expert pilot to our drones. I will conclude with a discussion of ongoing challenges and new directions of research we are pursuing to usher in the next chapter of intelligent and autonomous aerial robotics.