Markerless Indoor Localization of a Drone without GNSS

Mon, 15 Jun 2026 00:00:00 +0000

Description

The drone used at the AerodrOHM currently determines its position via an external, marker-based tracking system. This provides very accurate poses but requires a permanently installed, instrumented flight space. The goal of this thesis is a markerless, GNSS-free indoor localization based solely on onboard sensors, so that the drone can fly independently of the external tracking. The work builds on the existing drone and the established ROS system architecture.

Work Packages

Familiarization with the existing drone platform and the marker-based tracking system at the AerodrOHM
Selection and integration of suitable onboard sensors (camera, IMU, optionally optical-flow or depth sensor)
Implementation of a markerless, GNSS-free localization method (e.g. visual-inertial odometry or visual SLAM)
Integration into the existing ROS system architecture and flight control
Evaluation of accuracy and robustness against the marker-based tracking as ground truth
Flight experiments and documentation of the results at the AerodrOHM

Requirements

Programming skills (Python and/or C++)
Ideally some experience with ROS / ROS 2
Basic knowledge of computer vision and/or state estimation (e.g. Kalman filter)
Interest in drones, sensor integration and localization
Independent and diligent way of working

This topic can be completed as a project or bachelor’s thesis subject to agreement.

Supervision

Role	Name	E-Mail
Supervisor	Prof. Dr. Christian Pfitzner	christian.pfitzner@th-nuernberg.de

Human-Robot Interaction on the UR12e – Tic-Tac-Toe as an Interactive Demo

Fri, 12 Jun 2026 00:00:00 +0000

Motivation

Collaborative robot arms such as the UR12e make robotics immediately tangible — especially when people can interact with them directly. A game like Tic-Tac-Toe is ideal for this: everyone knows the rules, the outcome is easy to follow, and the robot demonstrates perception, decision-making and precise motion in a single, illustrative sequence.

At the TTZ Nürnberger Land, this is to become the basis for a demonstrator that makes robotics graspable for a broad audience — at family events, open-house days and for visitor groups. At the same time, the project serves as a platform to put concepts of safe human-robot interaction into practice.

Work Packages

Perception

Camera-based detection of the board and the placed game pieces
Calibration between camera and robot coordinate frames (hand-eye calibration)

Game Logic

Implementation of the Tic-Tac-Toe logic (e.g. minimax) for move computation
Detection of game end, win, loss and draw

Robot Control

Control of the UR12e via its existing interfaces (ROS or URScript)
Grasping and safely placing the game pieces
Functional safety in collaborative operation (force/speed limiting)

Interaction & Demo

Intuitive game flow and user guidance understandable for non-experts
Robustness testing and preparation as a presentable demo
Optional outlook: conceptual extension to chess

Requirements

Programming skills in Python and/or C++
Ideally some prior exposure to robotics
Interest in computer vision, game logic and human-robot interaction
Enthusiasm for hands-on work with real robot hardware

This topic is offered as a project or bachelor’s thesis and, given a good fit, can serve as a foundation for follow-up work (e.g. extension to chess).

Supervision

Role	Name	E-Mail
Supervisor	Prof. Dr. Christian Pfitzner	christian.pfitzner@th-nuernberg.de
Co-Supervisor	M. Sc. Patrick Fußy	patrick.fussy@th-nuernberg.de

Location: TTZ Nürnberger Land / TH Nürnberg Georg Simon Ohm

Computer Vision | Mobile Robotics Research Group — Prof. Dr. Christian Pfitzner

Markerless Indoor Localization of a Drone without GNSS

Description

Work Packages

Requirements

Supervision

Human-Robot Interaction on the UR12e – Tic-Tac-Toe as an Interactive Demo

Motivation

Work Packages

Requirements

Supervision