The Innovation Center (IZ) observes trends in technology and society and serves as an incubator and accelerator for the explorative development of modern business models for the entire IABG Group. The Innovation Center sees itself as the spearhead for industrializing the latest research and putting it to beneficial use for customers. We experiment not only with technologies and methods but also with organizational forms and business models. We are guided by the question of how an increasingly globalized and urban society can be developed in a secure and sustainable manner, whose organizations and systems are already networked at Internet speed.
The popularity of unmanned aerial vehicles (UAVs) has laid the groundwork for interest in cooperating UAVs. In particular, swarms of UAVs are the subject of many research efforts [1], including in applications related to surveillance and tracking. Swarms are typically groups of UAVs that are working together to achieve a common goal. As a multi-robot system, one of the ways that a UAV swarm can be characterized is by the way it is organized: when a single leader is giving instructions to all UAVs in the swarm, then the swarm is considered strongly centralized. One of the general motivations for using swarms is to have a system that is robust to failures; if one UAV unexpectedly fails, the others will continue to operate and be capable of completing the swarm’s task, although possibly with degraded performance.
This thesis topic will aim to (1) develop a swarm route planning system for a strongly centralized swarm, and (2) test the new route planning system in situations where one UAV stops functioning during the mission. Working with TUM Chair of Software & Systems Engineering, you will develop a swarm route planning system for a specific object tracking mission. The goal of the route planner will be to allocate goal positions (waypoints) to all the UAVs at all time-points such that the UAV swarm fulfills the mission of single/multi object tracking. IABG-developed simulation tools based on Airsim will be used to design and implement the system. To test the system, you will design simulation-based experiments that show the performance and robustness of your implementation.

Tätigkeit:

• Use existing literature to build your understanding of strongly centralized swarms and route planners for UAVs. Concretely describe the mission scenario including the research questions you want to answer regarding the performance and robustness of your implementation.
• Based on your findings in the literature, design and implement a route planner for the swarm (using e.g. search or optimization approaches).
• Using the existing Airsim-based simulation environment, measure and evaluate your implementation’s performance and robustness compared to an equivalent route-planning approach regarding performance (based on the quality of object tracking) and robustness (based on the change in performance when one UAV fails).
• Write the thesis document and prepare the presentation.

Voraussetzungen:

• Student in Computer Science, Mathematics, Robotics, Aerospace, Mechanical Engineering, or a related field
• Strong foundation in mathematics, particularly in optimization and probability theory
• Proficiency in programming languages such as Python and C++
• Familiarity with (swarm) path/route planning approaches and basic (multirotor) flight mechanics
• Familiarity with robotics simulation tools such as ROS (Robot Operating System), or Unreal Engine or AirSim
• Excellent communication skills in English

Ziele:
As a master’s student in our team, you will have the opportunity to work independently and promote the developed solution as a product to interested stakeholders

Please refer the Job description for details