LEVIO: Lightweight Embedded Visual Inertial Odometry for Resource-Constrained Devices

2026-02-01

IEEE Sensors Journal

AI Summary

This work proposes a lightweight, full-featured visual-inertial odometry (VIO) pipeline tailored for resource-constrained platforms such as micro aerial vehicles and smart glasses. By streamlining core algorithms—specifically ORB feature tracking and bundle adjustment—and co-optimizing software with hardware, the system achieves a memory-efficient, highly parallelized architecture compatible with ultra-low-power SoCs like RISC-V. Experimental validation on public VIO benchmarks demonstrates that the proposed approach maintains competitive localization accuracy while achieving real-time performance of 20 FPS on a RISC-V platform with power consumption below 100 mW, thereby significantly improving energy efficiency.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Machine Learning: Hardware-aware ML
• Intelligent Robots: Localization, Mapping, and Navigation

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Virtualization and resource management in Web systems and infrastructures
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Accurate, infrastructure-less sensor systems for motion tracking are essential for mobile robotics and augmented reality (AR) applications. The most popular stateof- the-art visual-inertial odometry (VIO) systems, however, are too computationally demanding for resource-constrained hardware, such as micro-drones and smart glasses. This work presents LEVIO, a fully featured VIO pipeline optimized for ultralow-power compute platforms, allowing six-degreesof- freedom (DoFs) real-time sensing. LEVIO incorporates established VIO components such as oriented FAST and rotated BRIEF (ORB) feature tracking and bundle adjustment (BA), while emphasizing a computationally efficient architecture with parallelization and low memory usage to suit embedded microcontrollers and low-power system on chip (SoCs). This article proposes and details the algorithmic design choices and the hardware-software co-optimization approach, and presents real-time performance on resource-constrained hardware. LEVIO is validated on a parallelprocessing ultralow-power RISC-V SoC, achieving 20 FPS while consuming less than 100mW, and benchmarked against public VIO datasets, offering a compelling balance between efficiency and accuracy. To facilitate reproducibility and adoption, the complete implementation is released as open-source.

Problem

• Visual-Inertial Odometry
• Resource-Constrained Devices
• Embedded Systems
• Low-Power Computing
• Real-Time Motion Tracking

Innovation

• Visual-Inertial Odometry
• Embedded Systems
• Low-Power Computing
• Hardware-Software Co-optimization
• RISC-V