Session F5b: PANEL: Navigation and Spatial Intelligence in Embodied AI

RETURN TO SESSION LIST

Date: Friday, September 18, 2026
Time: 10:35 a.m. - 12:15 p.m.
Location: Windsong 3-4

Leading AI and navigation scientists to address the critical challenges facing navigation in embodied agents such as drones, humanoids, and quadrupeds. Effective autonomy is currently constrained by two key factors: the operation in perceptually-degraded environments (e.g., indoor, subterranean, dense canopy, low-light) and the strict size, weight, power, and cost (SWaP-C) limitations of onboard sensors and computation. We will dissect the most pressing open problems in navigation, mapping, and perception. The discussion will highlight recent advancements in: Vision-and-Language Navigation (VLN), AI-driven 3D Reconstruction, Reinforcement Learning, and Visual and Navigation Foundation Models. Furthermore, the panel will examine the evolving role of GNSS, its integration with other sensing modalities, and the strategies for achieving robust localization in signal-denied environments.

	Dr. Mainak Biswas, Qualcomm
	Dr. Grace Gao, Stanford University
	Dr. Li-Ta Hsu, The Hong Kong Polytechnic University

12:15 p.m. – 1:30 p.m., Awards Luncheon, Lunch served until 12:30 p.m.; late arrivals will not be served

RETURN TO SESSION LIST