Thesis supervisor: Szabolcs Rózsa
Location of studies (in Hungarian): BME Általános- és Felsőgeodézia Tanszék Abbreviation of location of studies: BMEAF
Description of the research topic:
There is a significant demand for surveying and mapping tasks in areas, which is not accessible by humans (e.g. nuclear power plants, public utility pipes, damping supports of reinforced concrete slabs). These areas can be surveyed and mapped by autonomous robotic platforms equipped with surveying instruments. The navigation of such platforms relies on different positioning techniques in indoor and outdoor environments. In outdoor environments the inertial navigation systems (gyroscopes, accelerometers and magnetometers – INS) and the odometry supports the widely used global navigation satellite systems (GNSS) to provide accurate positioning and orientation observations. However, in indoor environments the GNSS needs to be substituted by other positioning techniques. One solution can be the application of ultra-wide band (UWB) sensors.
Apart from the application of positioning and orientation sensors, the autonomous positioning and mapping systems rely on simultaneous localization and mapping (SLAM) algorithms, which help to assist the positioning and orientation based on the fitting of 2D/3D point clouds measured by LiDAR sensors.
It is trivial that a reliable positioning and orientation solution can be created by the fusion of various sensors. Since the accuracy of these sensors is affected by the environment, the optimal position and orientation can be achieved by a data processing system relying on artificial intelligence algorithms.
The aim of the doctoral research is to further develop the recently realised robotic platform that already has a low level of autonomous navigation capabilities in indoor environments, and to create a robotic platform that is capable to carry out automatic mapping. The platform needs to achieve an optimal position and orientation determination using various sensors to be used in indoor and/or outdoor environments, such as GNSS, INS, UWB, odometer, camera, LiDAR, ultrasonic distance meter.
In order to achieve these goals, the PhD candidate must study and examine the accuracy and reliability of the various positioning sensors, with a special focus on the recently developed UWB sensors and the various low-cost sensors commonly used on small and lightweight platforms.
The candidate develops an optimal estimation methodology for the simultaneous localisation and mapping in dynamically changing environments using the 2D and 3D mapping sensors of the platform. In this part of the work various filtering and estimation techniques must be tested and evaluated, like the extended Kalman filter (EKF), the unscented Kalman filter (UKF), particle filtering and sparse extended information filter (SEIF).
The research also aims at the development of the data processing methodology of the fused mapping and positioning sensors, that takes into consideration the environment-dependent accuracy of the various sensors using self-learning algorithms based on artificial intelligence theory. Thus the platform is foreseen to be able to achieve highly accurate positioning and orientation determination in both indoor and outdoor environments.
The data processing algorithms will be developed in open software environment, and will be tested on real observations taken by the platform. The validation of the methodology will be done by independent traditional land surveying techniques.
The overall result of the doctoral research will be an autonomously navigating robotic platform that is capable to do mapping, where the traditional surveying and mapping techniques cannot be applied.
Number of students who can be accepted: 1
Deadline for application: 2022-12-20
2024. IV. 17. ODT ülés Az ODT következő ülésére 2024. június 14-én, pénteken 10.00 órakor kerül sor a Semmelweis Egyetem Szenátusi termében (Bp. Üllői út 26. I. emelet).
Login
