Learning-Type Anchors-Driven Pose Estimation for the Autolanding Fixed-Wing UAVs

<p>We propose a learning-type anchors-driven real-time pose estimation method for the autolanding fixed-wing unmanned aerial vehicle (UAV). The proposed method enables online tracking of both position and attitude by the ground stereo vision system in the Global Navigation Satellite System denied environments. A pipeline of convolutional neural network (CNN)-based UAV anchors detection and anchors-driven UAV pose estimation are employed. To realize robust and accurate anchors detection, we design and implement a Block-CNN architecture to reduce the impact of the outliers. With the basis of the anchors, monocular and stereo vision-based filters are established to update the UAV position and attitude. To expand the training dataset without extra outdoor experiments, we develop a parallel system containing the outdoor and simulated systems with the same configuration. Simulated and outdoor experiments are performed to demonstrate the remarkable pose estimation accuracy improvement compared with the conventional Perspective-N-Points solution. In addition, the experiments also validate the feasibility of the proposed architecture and algorithm in terms of the accuracy and real-time capability requirements for fixed-wing autolanding UAVs.</p>

Learning-Type Anchors-Driven Pose Estimation for the Autolanding Fixed-Wing UAVs

<p>We propose a learning-type anchors-driven real-time pose estimation method for the autolanding fixed-wing unmanned aerial vehicle (UAV). The proposed method enables online tracking of both position and attitude by the ground stereo vision system in the Global Navigation Satellite System denied environments. A pipeline of convolutional neural network (CNN)-based UAV anchors detection and anchors-driven UAV pose estimation are employed. To realize robust and accurate anchors detection, we design and implement a Block-CNN architecture to reduce the impact of the outliers. With the basis of the anchors, monocular and stereo vision-based filters are established to update the UAV position and attitude. To expand the training dataset without extra outdoor experiments, we develop a parallel system containing the outdoor and simulated systems with the same configuration. Simulated and outdoor experiments are performed to demonstrate the remarkable pose estimation accuracy improvement compared with the conventional Perspective-N-Points solution. In addition, the experiments also validate the feasibility of the proposed architecture and algorithm in terms of the accuracy and real-time capability requirements for fixed-wing autolanding UAVs.</p>

A Study on the Characteristics of the Concentrated Solar Photovoltaic/thermal Integrated System

This paper mainly introduces a comprehensive solar energy utilization system with low-CPC and polycrystalline silicon solar photovoltaic components. Outdoor experiments have been made to test the electrical efficiency, heat efficiency and their changes of the system with fixed temperature of outlet water and different sunshine conditions. Besides, this paper also compares the effects of outlet water at various temperatures on the efficiency of the whole system as well as the temperatures and flow rate responses with different PID parameters. Through analysis of these experimental data, the most desirable temperature of outlet water and PID parameters have been obtained in CPC-PV/T Hybrid Thermal-electric System, thus providing references for relevant experimental research.

On-board cable attitude measurement and controller for outdoor aerial transportation

Deploying quadcopters for aerial transportation can be cost effective in impromptu material handling applications. However, such applications are limited mainly due to the requirement of onboard localization sensors and associated computation. The current work presents a human-controlled modality to successfully execute spontaneous outdoor flight of a quadcopter with a cable-suspended payload. Stable and smooth flights are achieved through an onboard integration of a custom-built sensor system and a controller to minimize payload oscillations. The feasibility of the proposed modality is demonstrated by conducting outdoor experiments and a case study in an unstructured environment.

Model-Based Approach for Cornering Stiffness and Yaw Moment of Inertia Estimation of a Scaled Electric Vehicle

This paper addresses the problem of estimating the tire cornering stiffness coecient and the yaw moment of inertia of a scaled car-like vehicle. The method merges measurements information of the vehicle lateral response along with its nonlinear planar model. Aiming effective and accurate results, we propose solving an optimization problem based on a representative dataset obtained experimentally using a persistently exciting input. The validation of the proposed method is shown by comparing the agreement between numerical simulations, evaluated with the estimated parameters, with experimental data. Three representative maneuvers are considered for this purpose, including indoor and outdoor experiments.