To path indicator Subpages of “Teaching“ To navigation To quick access To footer with other services
To content
FLW is the Chair of Material Handling and Warehousing
Department of Mechanical Engineering
© FLW, TU Dortmund University

Thesis List

Below you will find the currently advertised thesis topics

To apply, please read the description of the procedure in the FAQ and the document templates.
Only one application per student will be considered at a time.
(The application deadline is open for each topic and ends when sufficient applications have been received.)

Smart warehouses increasingly depend on Autonomous Mobile Robot (AMR) fleets, yet safe and predictable routing remains difficult under dynamic layouts, mixed traffic, and uncertainty in perception and communication. This thesis conducts a systematic literature review of Perception-Aware Route Planning in intralogistics using i.e., the PRISMA methodology. The goal is to structure the state of the art across warehouse planning, collective perception, uncertainty modelling, and communication-aware decision making, and to identify research gaps that prevent end-to-end integration toward PARP. Core tasks include defining research questions and inclusion criteria, designing database queries, screening and quality assessment, and extracting evidence into a concept matrix. The thesis should compare with existing state of the art approaches in Warehouse such as shortest-path routing, time-window and congestion-aware routing, multi-robot traffic management, and WMS-driven task allocation, against PARP concepts that integrate perception uncertainty, time-critical communication constraints, and safety risk into planning. Expected outcomes are a PRISMA flow, a taxonomy, an evidence-based gap analysis, and a research agenda aligned with 6R logistics principle (time, cost, quality, safety).
Requirements: Background in logistics or industrial engineering; ability to work with scientific databases and structured reviews; basic familiarity with AMRs, warehouse operations, and performance metrics.

(Published: February 9, 2026)

Collective perception can improve safety and efficiency for multi-robot navigation, but only if planning accounts for uncertainty and time-critical communication limits. This thesis implements a Perception-Aware Route Planning (PARP) pipeline and validates it in simulation and or on a physical robot fleet. The student will design a planning interface that converts perception uncertainty and communication quality indicators into risk-aware cost terms for routing and will integrate it into a ROS2 navigation stack. Core tasks include combining existing GNN-based perception fusion algorithm, defining Planner Usable Confidence (PUC) metrics at track and route-segment level, coupling PUC into a global planner cost function, and evaluating behavior under controlled stressors such as occlusions, packet loss, delay, and density changes. The thesis should report safety and predictability metrics, for example collision rate, near-miss indicators, replan frequency, travel time variance, and throughput proxies.
Requirements: Strong programming skills (Python/C++), robotics fundamentals (localization, mapping, planning), and experience with ROS/ROS2. Basic understanding of wireless communication concepts (5G,6G, Network Topology) is recommended. Familiarity with Gazebo/Isaac Sim or real robot experimentation is a plus.

(Published: February 9, 2026)

Event cameras offer high temporal resolution and robustness to motion blur, making them suitable for high-speed robotic perception. However, 2D object segmentation with event data remains challenging due to ego-motion, background clutter, and the lack of stable image representations.

In this thesis, an event-based SLAM system is used to estimate the camera pose over time. The focus of the work is on exploiting the estimated camera pose to improve 2D event-based segmentation. The pose information is used to compensate camera motion during event accumulation, suppress background events, and stabilize object boundaries in the image plane.

The student will implement and evaluate pose-assisted event representations and compare segmentation performance with pose-agnostic baselines. The impact of camera pose accuracy on segmentation quality will be analyzed quantitatively. As an optional extension, the thesis will discuss how improved 2D segmentation and camera pose can support multi-view 3D bounding box estimation.

Requirements:

  • Solid programming skills in Python
  • Basic knowledge of computer vision
  • Familiarity with robotics concepts (coordinate frames, camera models)
  • Interest in event-based vision
  • Experience with ROS
  • Basic understanding of SLAM concepts

(Published: February 9, 2026)

How to use the Application Form

  1. Enter your personal details

    • Fill in your First Name, Last Name, and Email Address (use your university email)

  2. Select the type of work

    • Choose one option: Project Work, Bachelor Thesis, or Master Thesis.

  3. Specify your thesis topic

    • Make sure that the title of the thesis matches the type of work!

    • In the field “Which thesis would you like to apply for?”, copy and paste the title or topic of your thesis. If the thesis is not offert in the list above your application will not be considered. If you have an external Thesis topic, in collaboration with a company fill in the proposed thesis topic. Ensure that the topic covers a scientific question/goal.

  4. Upload required documents

    • Transcript of Records: Upload your academic transcript (PDF)

    • Letter of Motivation: Upload your motivation letter (PDF)

    • (optional) Upload your  CV (PDF)

  5. Submit your application

    • Review all fields and make sure everything is complete

    • Click the green “Send” button to submit your form

  6. Confirmation 

    • You should now be forwarded to another conformation side if not please contact our technical support team via e-mail 

* Required field

Thesis Application Form

(If an external thesis or a thesis in cooperation with the Fraunhofer Institute is to be written, an Exposé is required)