Deep Odometry Systems on Edge with EKF-LoRa Backend for Real-Time Indoor Positioning

Abstract

Ubiquitous positioning for pedestrians in adverse environments has been a long standing challenge. Despite dramatic progress made by Deep Learning, multi-sensor deep odometry systems still pose a high computational cost and suffer from cumulative drifting errors over time. Thanks to the increasing computational power of edge devices, we propose a novel ubiquitous positioning solution by integrating state-of-the-art deep odometry models on edge with an EKF (Extended Kalman Filter)-LoRa backend. We carefully select and compare three sensor modalities, i.e., an Inertial Measurement Unit (IMU), a millimetre-wave (mmWave) radar, and a thermal infrared camera, and implement their deep odometry inference engines to run in real-time. A pipeline for deploying deep odometry on edge platforms with different resource constraints is proposed. We design a LoRa link for positional data backhaul and project aggregated positions of deep odometry into the global frame. We find that a simple EKF backend is sufficient for generic odometry calibration with over 34% accuracy gains against any standalone deep odometry system. Extensive tests in different environments validate the efficiency and efficacy of our proposed positioning system.