Vision‐based Localization and Robot‐centric Mapping in Riverine Environments

Abstract

This paper presents a vision-based localization and mapping algorithm developed for an unmanned aerial vehicle (UAV) that can operate in a riverine environment. Our algorithm estimates the three-dimensional positions of point features along a river and the pose of the UAV. By detecting features surrounding a river and the corresponding reflections on the water's surface, we can exploit multiple-view geometry to enhance the observability of the estimation system. We use a robot-centric mapping framework to further improve the observability of the estimation system while reducing the computational burden. We analyze the performance of the proposed algorithm with numerical simulations and demonstrate its effectiveness through experiments with data from Crystal Lake Park in Urbana, Illinois. We also draw a comparison to existing approaches. Our experimental platform is equipped with a lightweight monocular camera, an inertial measurement unit, a magnetometer, an altimeter, and an onboard computer. To our knowledge, this is the first result that exploits the reflections of features in a riverine environment for localization and mapping. C 2015 Wiley Periodicals, Inc.

A Survey on Aerial Swarm Robotics

“…Instead, the robots must rely on on-board sensors, such as cameras, lidars, and inertial measurement units (IMUs). Cameras and lidars are exteroceptive sensors, relying on external features to provide incremental pose estimates [219]. On the other hand, IMUs are interoceptive sensors, providing high-frequency velocity and attitude feedback for the purpose of real-time control.…”

Omnidirectional-vision-based estimation for containment detection of a robotic mower

“…In our previous work [19], we used a decoupled set of nonlinear observers [20] with depth parametrization, opposed to the inversedistance with initial view measurements used in this paper, and showed simulation results, along with limited experimental results of only the boundary estimation and landmark mapping. In this paper we employ an omnidirectional camera as our primary sensor to solve 6-DOF localization and 3D mapping by using a robot-centric framework with initial view measurements, which is introduced in our prior work [4]. We apply the results to detect the containment of our robotic mower.…”

“…Note that we parametrize the landmarks with a unit vector and an inverse-distance from the robot instead of a normalized pixel coordinates and a depth along the optical axis, which we used in [4]. We are able to have a continuous parametrization of the landmarks that are acquired through an omnidirectional camera.…”

“…The location of the mower during autonomous mowing can be estimated with the map composed of landmarks given by (4). The state vector is composed of the estimate of the locationp w b ∈ R 3 and the orientation quaternionsq w b ∈ R 3 of the mower with respect to the world frame, and the velocityv b of the mower.…”

“…Recent advances in vision-based estimation technologies are enabling autonomous robots to execute tasks in various environments [1]- [4]. The goal in our research is to expand the application scope of vision-based estimation by developing a containment detection system for robotic mowing.…”