Back to main page / Back to main page

Jump to the page in German language

Drones with Artificial Intelligence - Master Project - Study Field Intelligent Systems (SS2025)

The application possibilities of drones are diverse, but access is costly and technically complex. This is unfortunate from the perspective of universities, because the combination of drones and artificial intelligence opens up exciting scenarios for research and teaching.

Participants in the project will develop, investigate, and document affordable AI capable drones that are suitable for teaching and research.

• Task 1: Develop and build an affordable FPV drone (ideally <250g).
• Task 2: Integrate a concrete AI application (e.g., object detection).
• Task 3: Research options for autopilot and possibly integrate them.
• Task 4: Research options for delivery or payload and possibly integrate them.
• Task 5: Documentation and presentation of the results of tasks 1 to 4.

Groups of up to 4 people each develop and build a drone. Individual steps are:

• Research the legal framework.
• Collect functional requirements for hardware and software.
• Select hardware components.
• Frame, Flight Controller (FC), Electronic Speed Controller (ESC), camera, video transmitter, antenna, receiver, radio transmitter, FPV goggles, single board computer (e.g., Raspberry Pi), AI accelerator (e.g., Google Coral USB Accelerator), FPV goggles, etc.
• Selection criteria: acquisition cost, feature set, availability.
• The hardware can be provided by the university or procured independently by the participants.
• Select software components.
• Flight stack software (e.g., Betaflight, iNAV, ArduPilot, PX4), AI frameworks and libraries (e.g., TensorFlow Lite, YOLO), etc.
• Selection criteria: license, feature set, activity of the open source project.
• Assemble the drone, integrate hardware and software components, perform flight tests.
• Interest in hardware and system software is mandatory. Basic knowledge of Linux, microcontrollers, and handling electronic components is helpful.
• Documentation and presentation.
• Each team develops a complete and easy to understand online documentation (e.g., via GitHub Pages) and a presentation consisting of a poster and a live demonstration. There will be no classic slide presentations or project reports.
• The participants develop documentation and guides that enable students, researchers, and lecturers to rebuild the AI drones at low cost and use them for their own modules and research projects.

Schedule for the semester

Documents

Project description

Project description

Teams

Bananastalker

• Approach and results:
• Poster
• Documentation
• Project Page
• Hardware:
• Frame: FlyFishRC Volador II VX5
• Flight Controller (FC): SpeedyBee F405 AIO
• Video Transmitter (VTX): SpeedyBee TX800 VTX
• Camera: Caddx Ratel Pro Analog
• Receiver: Radiomaster RP1 ELRS 2.4GHz Nano
• Transmitter: Radiomaster Zorro ELRS
• Motors: XING-E Pro 2207 1800KV
• Video Headset: Skyzone Cobra X V4
• Single-board computer: Raspberry Pi Zero 2 W
• Camera for RPi: Arducam IMX219
• AI accelerator: Coral Google Edge TPU ML Accelerator
• Software:
• Betaflight

FlySpy

• Approach and results:
• Poster
• Documentation
• Project Page
• Hardware:
• Frame: FlyFishRC Volador II VX3
• Flight Controller (FC): SpeedyBee F405 AIO
• Video Transmitter (VTX): SpeedyBee TX800 VTX
• Camera: Caddx Ratel Pro Analog
• Receiver: Radiomaster RP1 ELRS 2.4GHz Nano
• Transmitter: Radiomaster Zorro ELRS
• Motors: XING-E Pro 1404 3850KV
• Video Headset: Skyzone Cobra S
• Single-board computer: Raspberry Pi Zero 2 W
• Camera for RPi: Arducam IMX219
• AI accelerator: Coral Google Edge TPU ML Accelerator
• Software:
• Betaflight

CloudChasers

• Approach and results:
• Poster
• Dokumentation
• Project Page
• Hardware:
• Frame: FlyFishRC Volador II VX3
• Flight Controller (FC): SpeedyBee F405 AIO
• Video Transmitter (VTX): SpeedyBee TX800 VTX
• Camera: Caddx Ratel Pro Analog
• Receiver: Radiomaster RP1 ELRS 2.4GHz Nano
• Transmitter: Radiomaster Zorro ELRS
• Motors: XING-E Pro 1404 3850KV
• Video Headset: Skyzone FPV Cobra
• Single-board computer: Raspberry Pi Zero 2 W
• Camera for RPi: Arducam IMX219
• AI accelerator: Coral Google Edge TPU ML Accelerator
• Software:
• Betaflight

Course evaluation

Contact

You can best reach me by e-mail: christianbaun@fb2.fra-uas.de