Aytar - Autonomous Search & Rescue Drone

Aytar is an autonomous drone project developed to assist in search and rescue operations by detecting human voices and emergency signals in real-time. The project advanced to the semi-finals in the Teknofest competition, achieving a preliminary score of 90/100.

Project Goal

The primary objective was to create a system capable of identifying victims under debris or in difficult terrain by analyzing sound signals, specifically focusing on emergency keywords and human speech patterns.

Technical Implementation

The software architecture focused on audio processing and real-time decision making:

• Audio Analysis: Utilized MFCC (Mel-Frequency Cepstral Coefficients) analysis to extract features from sound signals, distinguishing human speech from background noise.
• Multi-Language Support: Implemented speech recognition algorithms supporting Turkish, English, French, and Spanish to detect emergency keywords in international contexts.
• Real-Time Processing: Developed algorithms to process live microphone input and automatically trigger notification systems upon detecting distress signals.

Tech Stack

• Core Language: Python
• Audio Processing: SpeechRecognition Library, MFCC Algorithms, Google Speech API
• Computer Vision: OpenCV (integrated for visual navigation support)
• Methodology: Natural Language Processing (NLP) for keyword extraction and classification