VINS-Fusion Evaluation

Introduction

1 The overview of VINS-Fusion

VINS-Fusion is a sensor fusion framework to fuse local states with global sensors, which achieves locally accurate and globally drift-free pose estimation and achieves accurate self-localization for autonomous applications (drones, cars, and AR/VR). Local estimations, produced by existing VO/VIO approaches, are fused with global sensors in a pose graph optimization. Within the graph optimization, local estimations are aligned into a global coordinate. Meanwhile, the accumulated drifts are eliminated. VINS-Fusion is an extension of VINS-Mono, which supports multiple visual-inertial sensor types (mono camera + IMU, stereo cameras + IMU, even stereo cameras only). code link. paper link.

Image description — Fig. 1. Workflow diagram of VINS-Fusion

2 VINS-Fusion Compilation Process

2.1 Requirement

(1)This package requires Ubuntu 64-bit 16.04 or 18.04. ROS Kinetic or Melodic.
(2)This package requires Ceres.

2.2 Build VINS-Fusion

Clone the repository and catkin_make:


            cd ~/catkin_ws/src

            git clone https://github.com/HKUST-Aerial-Robotics/VINS-Fusion.git

            cd ../

            catkin_make

            source ~/catkin_ws/devel/setup.bash

if you fail in this step, try to find another computer with clean system or reinstall Ubuntu and ROS

Evaluation

Platforms	Sequences	Length(m)	ATE
A Handheld Device	Escalator	77.460	0.746
	MCR normal dark	76.499	0.377
	MCR aggressive 6dof light	100.871	2.373
	Parkway loop night	461.049	25.894
	Forest	130.937	0.933
A UGV	Elevator	39.336	X
	Indoor loop	270.674	0.767
	MCR hdr	193.918	7.459
	Street day	2064.475	51.820
	Parkway loop night	461.051	12.964
A QR	Underground	98.312	0.659
	MCR hdr	85.08	0.573
	Forest	108.037	0.301
A UAV	MCR loop light	104.989	0.041
A Car	Urban night loop	1807.884	33.386