Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Embedded gimbal system for land-based tracking of unmanned aerial vehicles

Manesis Athanasios

Full record

---

Summary

Embedded systems are one of the most important and rapidly developing fields in engineering. They find application from small household appliances to critical airplane safety systems that have hundreds of them to perform functions such as in-flight entertainment systems, temperature control, speed control, flight management systems, flight data recorders, engine control and many more. With the use of artificial intelligence and machine learning algorithms these systems become more intelligent, efficient and have much more autonomy than in the past. In this thesis we studied and implemented such an embedded system that has the ability to detect and track UAVs (drone) through optical recognition. Detection and monitoring is visible via a monitor which is connected to the embedded system and has three different modes, one for horizontal scanning, one for vertical scanning and one for tracking. The embedded system consists of a rotation mechanism (PTZ) in which a camera is mounted and an RPi-4 which through the camera recognizes the drone using a machine learning model for object recognition and generates the appropriate angles to the gimbal to turn in that way so the drone stays as close to the center of the camera as possible. Communication between the RPi and the PTZ device is achieved via RS485 using the Pelco-P and Pelco-D protocols that the rotation mechanism supports. To convert the angles into the appropriate Pelco-P and Pelco-D frames, a python library was developed so that by using the corresponding functions, the appropriate messages can be produced without anyone needing to know exactly how these frames are created.

Available Files

Services