cuVSLAM: CUDA accelerated visual odometry and mapping

TL;DR

cuVSLAM is a CUDA-accelerated visual SLAM system supporting multiple sensors and configurations, optimized for real-time edge computing, demonstrating superior benchmark performance.

Contribution

The paper introduces cuVSLAM, a novel CUDA-optimized visual SLAM framework supporting diverse sensor setups and configurations for real-time robotic applications.

Findings

Achieves real-time performance on NVIDIA Jetson devices.

Outperforms existing SLAM methods on benchmark datasets.

Supports up to 32 cameras in arbitrary configurations.

Abstract

Accurate and robust pose estimation is a key requirement for any autonomous robot. We present cuVSLAM, a state-of-the-art solution for visual simultaneous localization and mapping, which can operate with a variety of visual-inertial sensor suites, including multiple RGB and depth cameras, and inertial measurement units. cuVSLAM supports operation with as few as one RGB camera to as many as 32 cameras, in arbitrary geometric configurations, thus supporting a wide range of robotic setups. cuVSLAM is specifically optimized using CUDA to deploy in real-time applications with minimal computational overhead on edge-computing devices such as the NVIDIA Jetson. We present the design and implementation of cuVSLAM, example use cases, and empirical results on several state-of-the-art benchmarks demonstrating the best-in-class performance of cuVSLAM.