College of Engineering Unit:
The goal of this device is to prevent bird-window collisions. When the device detects a bird flying towards a window, the device alerts the bird to allow the bird to change its flight path away.
To accomplish this goal, the system utilizes two sensor methods and two deterrent methods. To identify birds in danger, a 60 GHz doppler radar motion detector is used to detect when birds are flying towards a window and a machine learning powered camera is used once motion is detected to ascertain whether or not the detected motion was that of a bird. To alert the bird to danger, prior to the collision, the system uses a speaker playing a chickadee warning call to notify the bird of impending danger and a strobe light to indicate where that danger lies.
The system was largely successful in detecting the motion of birds and similar small objects and the chickadee warning call was shown to be effective in convincing birds that danger was present. However, the implementation of the machine learning identification of birds and the use of a strobe light to indicate where the danger lied both were shown to have a potential for future improvement.
Project Website(s):
| Attachment | Size |
|---|---|
 active_bird_defense_executive_summary.pdf | 197.01 KB |
| ece_44x_group_21_project_document_0.pdf | 13.01 MB |
| Project Poster.pdf | 3.14 MB |