Real-time safety alerting system for construction sites

Construction sites represent complex, dynamic, environments where ensuring safety is crucial yet challenging. Traditional safety measures often fall short due to their reliance on manual monitoring and intervention, which are often time-consuming and prone to errors. To address these challenges, this study proposes a Real-Time Safety Alerting System for Construction Sites, by employing Bluetooth Low Energy (BLE) devices for indoor localization coupled with a customized Android application for real-time monitoring. Leveraging the FIND3 framework, significant customization was introduced to create a robust system capable of tracking workers and equipment within a construction site with the objective of alerting potential safety hazards. The core components of the system include an Android application, server-side components, and an interactive front-end website. The Android app, installed on workers' phones, interfaces with BLE devices deployed across a construction site, facilitating precise indoor positioning using a fingerprinting algorithm. Server-side components, implemented using Go, Python, and Docker, provide administrative control, data management, and real-time monitoring. An interactive website displays the location of workers and equipment on a floor plan, alongside real-time safety alerts. The system architecture further incorporates construction site zoning, dividing the floor plan into different zones with specific safety requirements. Machine learning algorithms such as k-NN, Random Forest, and SVM are employed on the server-side to analyze location data for hazard detection and safety management. Real-time alarms and notifications are relayed to both workers and employers through the Android app and website, enhancing the overall safety management of the construction site. Through rigorous testing and evaluation on an emulated construction site, this system demonstrates a promising approach to bolster construction site safety, providing a foundation for further enhancements and real-world deployment. This thesis delineates the design, implementation, and evaluation of this system on a proof-of-concept system, shedding light on the potential of integrating modern wireless communication, indoor localization, and machine learning technologies to improve construction site safety management.