Distributed Variable-Baseline Stereo SLAM from two UAVs

Abstract

Visual-Inertial Odometry (VIO) has been widely used and researched to control and aid the automation of navigation of robots especially in the absence of absolute position measurements, such as GPS. However, when the observable landmarks in the scene lie far away, as in highaltitude flights for example, the fidelity of the metric scale estimate in VIO greatly degrades. Aiming to tackle this issue, in this work, we utilize the virtual stereo setup formed by two Unmanned Aerial Vehicles (UAVs), equipped with one camera and one Inertial Measurement Unit (IMU) each, exploiting their view overlap and relative distance measurements between them using onboard Ultra-Wideband (UWB) modules to enable collaborative VIO. In particular, we propose a decentralized collaborative estimation scheme, where each agent holds its own local map, achieving a low pose estimation latency, while ensuring consistency of each agents' estimates via consensus-based optimization. Following a thorough evaluation in photorealistic simulations, we demonstrate the effectiveness of the approach at high-altitude flights of up to 160m, going significantly beyond the capabilities of state-of-the-art VIO methods. Finally, we show the advantage of actively adjusting the baseline on-the-fly over a fixed, target baseline, resulting in a significant reduction of the estimation error.Videohttps://youtu.be/SdL4Jb-BQ28

Variable Baseline and Flexible Configuration Stereo Vision Using Two Aerial Robots

“…If the issues are resolved, such markers can be used again in place of mocap. Other choices considered are inside–out tracking methods such as dual SLAM introduced in [ 19 ] and the collaborative localization of stereo cameras on UAVs described in [ 20 ].…”

VR-SLAM: A Visual-Range Simultaneous Localization and Mapping System using Monocular Camera and Ultra-wideband Sensors

“…, where 18,18), ([15, 16, 17, 4], [15,16,17,5], [1, 1, 1, −1], 18, 18), [6,9,12,18], [1, 1, 1, −1], 18, 18), [7,10,13,18], [1, 1, 1, −1], 18, 18), [8,11,14,18], [1, 1, 1, −1], 18, 18), P 6 = sparse([4], [18], [1], 18, 18), (42) P 7 = sparse( [18], [18], [1], 18,18),…”

“…where P 8 = sparse([6, 9, 12, 5], [6,9,12,18], [1, 1, 1, −1], 18, 18), P 9 = sparse([7, 10, 13, 5], [7,10,13,18], [1, 1, 1, −1], 18, 18), P 10 = sparse( [8,11,14,5], [8,11,14,18], [1, 1, 1, −1], 18, 18), P 11 = sparse([6, 9, 12], [7,10,13], [1, 1, 1], 18, 18), P 12 = sparse( [6,9,12], [8,11,14], [1, 1, 1], 18, 18), (43) P 13 = sparse( [7,10,13], [8,11,14], [1, 1, 1], 18, 18), p 8 = 0, p 9 = 0, p 10 = 0, p 11 = 0, p 12 = 0, p 13 = 0.…”

“…In this context, UWB sensor provides several key advantages [3]: different from VO, UWB distance measurements are drift-free and unaffected by visual conditions; unlike GPS, UWB can be used both indoors and outdoors; compared to LiDAR, UWB sensor is smaller, lighter, more affordable and simpler to install on a robot platform. UWB can also be employed in multi-robot cases as a communication network [4], [5], as a variable baseline between two robots [6], or as inter-robot constraints for pose graph optimization [7] or formation control [8]. On the other hand, UWB sensor requires good line-of-sight (LoS) for accurate measurement [9], is unable to provide any perceptual information about the environment, and the performance of range-based localization methods relies on having enough UWB anchors installed and their configuration does not fall into a degenerate case [2].…”

Multiple UAVs for synchronous-shared tasks and longterm autonomy

“…Collaborative multi-robot mapping techniques have been developed for registering images from different robots. [10,11]. They either adhere to static scene assumptions, or the robots try to relocalize in a pre-built map of the environments.…”