Monocular SLAM System for MAVs Aided with Altitude and Range Measurements: a GPS-free Approach

Landmark-Based Scale Estimation and Correction of Visual Inertial Odometry for VTOL UAVs in a GPS-Denied Environment

With the rapid development of technology, unmanned aerial vehicles (UAVs) have become more popular and are applied in many areas. However, there are some environments where the Global Positioning System (GPS) is unavailable or has the problem of GPS signal outages, such as indoor and bridge inspections. Visual inertial odometry (VIO) is a popular research solution for non-GPS navigation. However, VIO has problems of scale errors and long-term drift. This study proposes a method to correct the position errors of VIO without the help of GPS information for vertical takeoff and landing (VTOL) UAVs. In the initial process, artificial landmarks are utilized to improve the positioning results of VIO by the known landmark information. The position of the UAV is estimated by VIO. Then, the accurate position is estimated by the extended Kalman filter (EKF) with the known landmark, which is used to obtain the scale correction using the least squares method. The Inertial Measurement Unit (IMU) data are used for integration in the time-update process. The EKF can be updated with two measurements. One is the visual odometry (VO) estimated directly by a landmark. The other is the VIO with scale correction. When the landmark is detected during takeoff phase, or the UAV is returning to the takeoff location during landing phase, the trajectory estimated by the landmark is used to update the scale correction. At the beginning of the experiments, preliminary verification was conducted on the ground. A self-developed UAV equipped with a visual-inertial sensor to collect data and a high-precision real time kinematic (RTK) to verify trajectory are applied to flight tests. The experimental results show that the method proposed in this research effectively solves the problems of scale and the long-term drift of VIO.

A Hybrid Visual-Based SLAM Architecture: Local Filter-Based SLAM with KeyFrame-Based Global Mapping

This work presents a hybrid visual-based SLAM architecture that aims to take advantage of the strengths of each of the two main methodologies currently available for implementing visual-based SLAM systems, while at the same time minimizing some of their drawbacks. The main idea is to implement a local SLAM process using a filter-based technique, and enable the tasks of building and maintaining a consistent global map of the environment, including the loop closure problem, to use the processes implemented using optimization-based techniques. Different variants of visual-based SLAM systems can be implemented using the proposed architecture. This work also presents the implementation case of a full monocular-based SLAM system for unmanned aerial vehicles that integrates additional sensory inputs. Experiments using real data obtained from the sensors of a quadrotor are presented to validate the feasibility of the proposed approach.

Camera-Based Tracking of Floating Objects using Fixed-wing UAVs

This article concerns tracking of floating objects using fixed-wing UAVs with a monocular thermal camera. Target tracking from an agile aerial vehicle is challenging because uncertainty in the UAV pose negatively affects the accuracy of the measurements obtained through thermal images. Consequently, the accuracy of the tracking estimates is degraded if navigation uncertainty is neglected. This is especially relevant for the estimated target covariance since inconsistency is a likely consequence. A tracking system based on the Schmidt-Kalman filter is proposed to mitigate navigation uncertainty. Images gathered with an uncertain UAV pose are weighted less than images captured with a reliable pose. The UAV pose is estimated independently in a multiplicative extended Kalman filter where the estimated covariance matrix is a measure of the uncertainty. The method is compared experimentally with two traditional alternatives based on the extended Kalman filter. The results show that the proposed method performs better with respect to consistency and accuracy.

Monocular Visual Inertial Direct SLAM with Robust Scale Estimation for Ground Robots/Vehicles

In this paper, we present a novel method for visual-inertial odometry for land vehicles. Our technique is robust to unintended, but unavoidable bumps, encountered when an off-road land vehicle traverses over potholes, speed-bumps or general change in terrain. In contrast to tightly-coupled methods for visual-inertial odometry, we split the joint visual and inertial residuals into two separate steps and perform the inertial optimization after the direct-visual alignment step. We utilize all visual and geometric information encoded in a keyframe by including the inverse-depth variances in our optimization objective, making our method a direct approach. The primary contribution of our work is the use of epipolar constraints, computed from a direct-image alignment, to correct pose prediction obtained by integrating IMU measurements, while simultaneously building a semi-dense map of the environment in real-time. Through experiments, both indoor and outdoor, we show that our method is robust to sudden spikes in inertial measurements while achieving better accuracy than the state-of-the art direct, tightly-coupled visual-inertial fusion method.

Computationally Efficient Cooperative Dynamic Range-Only SLAM Based on Sum of Gaussian Filter

A cooperative dynamic range-only simultaneous localization and mapping (CDRO-SLAM) algorithm based on the sum of Gaussian (SoG) filter was recently introduced. The main characteristics of the CDRO-SLAM are (i) the integration of inter-node ranges as well as usual direct robot-node ranges to improve the convergence rate and localization accuracy and (ii) the tracking of any moving nodes under dynamic environments by resetting and updating the SoG variables. In this paper, an efficient implementation of the CDRO-SLAM (eCDRO-SLAM) is proposed to mitigate the high computational burden of the CDRO-SLAM due to the inter-node measurements. Furthermore, a thorough computational analysis is presented, which reveals that the computational efficiency of the eCDRO-SLAM is significantly improved over the CDRO-SLAM. The performance of the proposed eCDRO-SLAM is compared with those of several conventional RO-SLAM algorithms and the results show that the proposed efficient algorithm has a faster convergence rate and a similar map estimation error regardless of the map size. Accordingly, the proposed eCDRO-SLAM can be utilized in various RO-SLAM applications.