What is it about?
The paper is the first to showcase the deployability of neural network-based inertial navigation. We use a physics-aware approach to constrain drift and position update errors, while using advances in TinyML to port lightweight localization models on embedded devices.
Featured Image
Photo by henry perks on Unsplash
Why is it important?
Useful for terrestrial and marine "search and rescue" missions, underwater sensor networks, oceanic biodiversity and marine health tracking, wildlife monitoring, deep-space small satellite localization, and localizing micro unmanned vehicles and robots, where compute device payload and resource availability are limited, and cannot have continuous access to GPS or network infrastructure.
Perspectives
We are the first to showcase the use of neural networks in real-time to perform inertial localization.
Swapnil Sayan Saha
University of California Los Angeles
Read the Original
This page is a summary of: TinyOdom, Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies, July 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3534594.
You can read the full text:
Contributors
The following have contributed to this page
