Sandeep Singh Sandha

Amazon (United States), Abacus Health Solutions

Papers

Total Citations

H-Index

About

Sandeep Singh Sandha is a leading researcher at the intersection of embedded artificial intelligence and inertial navigation, specializing in enabling precise localization on the world’s most resource-constrained devices. His work fundamentally addresses the challenge of accurate positioning in GPS-denied environments, particularly for Internet-of-Things (IoT) platforms and precision agriculture. Sandha’s major contributions include the development of ultra-lightweight neural-Kalman filters that fuse deep learning with classical state estimation. His pioneering paper, "Neural-Kalman GNSS/INS Navigation for Precision Agriculture" (2023, 15 citations), introduces a robust neural-inertial sequence learning approach that tracks agricultural robots using only ultra-intermittent GNSS updates, achieving high-resolution navigation with minimal computational overhead. Complementing this, his work "Inertial Navigation on Extremely Resource-Constrained Platforms" (2023, 5 citations) systematically explores methods, opportunities, and challenges for deploying inertial navigation on low-power, small-footprint IoT devices, moving beyond traditional physics-based heuristics. By bridging the gap between deep learning and classical Kalman filtering, Sandha is creating practical, deployable navigation solutions that operate reliably where conventional systems fail, making him a key innovator in edge-AI localization.