Development of an RPAS payload for indoor mapping and extended uses
Tutor / director / evaluatorHornero Ocaña, Gemma
Document typeMaster thesis
Rights accessRestricted access - confidentiality agreement
Nowadays, while outdoor navigation and mapping services using RPAs have been quite common, navigation and mapping technologies for indoor environment application have not been widely put into practice. There exists a gap between the indoor navigation and mapping services currently provided in the market and the ever-growing requirements of its applications, especially when it comes to life-saving tasks in emergency and disaster management. The main goal of this master project is to provide a suitable solution for indoor navigation and mapping services in the post-disaster scenarios to help rescue teams in their task. The project aims to develop a low-cost system, which can serve as a payload on a small RPAS for data acquisition while performing online near real-time mapping mission based on visual-inertial Simultaneous Localization and Mapping (SLAM) in a post-disaster indoor scenario. The project mainly includes four phases. First, the use case for disaster management and the corresponding user requirements have been defined. Secondly, the hardware and software architecture has been designed based on the system requirements after the feasibility study. The payload prototype consists of an Odroid XU4 single board computer, GNSS/IMU module, and an Intelrealsense D435 RGB-D camera. Thirdly, the detail data-processing approaches have been implemented and tested. The onboard platform runs on UBUNTU 16.04. ROS. Extend Kalman Filter (EKF) is used for data fusion of the IMU and visual odometry generated from the RGB-D camera. Back-end from the RTABMAP ROS package is used for loop detection and minimize global inconsistencies based on pose graph optimization methods. Post-processing procedures have been proposed to refine the mapping results and generate 2D floorplans. Finally, verification and validation work has been done in several indoor scenarios to evaluate the performance of the system. The onboard system has been able to map an indoor office room and a challenging floor layout. Due to hardware and time limitation, planned flight tests were not able to be carried out and have been left for future work. The contributes of the work are twofold. It offers a complete solution for low-cost, fast indoor mapping with RPAS in post-disaster scenarios, and supports the viability based on the state-of-the-art sensor and algorithms currently available. On the other hand, it shows the existing gap between the current technologies and the requirement for application in real disaster scenarios as several challenges remain to be overcome in the future.
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder