Virtual Booth

Bullying types, such as verbal, physical and social in schools mostly happen in private indoor locations, such as toilets. Due to privacy concerns, the use of closed-circuit television systems (CCTV) is prohibited in such areas. Since conventional approaches relying on manual reporting are often ineffective due to fear of retaliation, there is a critical need for an automated detection system. Audio such as shouting, crying and loud impacts are strong indicators of violent and aggressive behaviour and can be used for automated detection in environments where visual monitoring is unsuitable. This project addresses the challenge of detecting such incidents in privacy-sensitive areas by developing an AI-powered smart anti-bullying monitoring system based on audio sensing. The objective of this study is to design and implement a ceiling-mounted, privacy-preserving system, which is capable of detecting potential bullying incidents and alerting relevant authorities in a timely manner. The proposed system collects audio data using a digital microphone integrated with ESP32-S3 microcontroller and transmits the recorded sound samples to a cloud-based artificial intelligence model for analysis. Based on the AI classification result, notifications are automatically sent to teachers and parents via the messaging platform, Telegram when bullying incidents occur. Local system feedback is provided through an OLED display and LED indicators, while a manual panic button enables immediate emergency activation. A web-based dashboard is also developed to monitor system activity, including detected threats, normal audio events, system status and deployment location. Preliminary evaluation shows that the system can reliably capture audio and deliver alerts when bullying situations are triggered. The proposed solution provides a scalable and cost-effective approach to improve student safety in privacy environments.