Geospatial assesment of a cost-effective IoT-based sensor for multi sensing detection in a campus setting

Abstract

The emergence of the Internet of Things (IoT) and the smart city concept has revolutionised automation frameworks, enabling the realisation of previously fictional scenarios. In line with this, this research aimed to develop an IoT-based sensor and integrate notification platforms for an early warning system during indoor fire hazards with a specific focus on the UKM campus building, using digital technology, including IoT sensors, actuators, and devices. The primary focus of this study is the application of sensor technology for detecting indoor fire hazards and incorporating a novel approach to assess the data quality produced by the sensor using geospatial analysis techniques. This research comprised seven phases: literature review, prototype development, database establishment, notification platform configuration, system integration, ground simulation, and data processing and geospatial analysis. To achieve this, a physical sensor consisting of temperature, humidity, smoke, and flame sensors was developed and subjected to simulations to assess its functionality. Five series of ground simulations with different time frames were conducted based on the probability of smoke ignition according to the building's attributes. Geospatial analyses, including inverse distance weighting interpolation, density calculation, standard deviational ellipse (SDE), and mathematical graph, were employed to represent and analyse the quality of the generated data. The raster representation indicated that the smoke distribution conformed to the simulation environment on the ground, with high correlation between regions related to the smoke ignition point and the presence of smoke. The analyses considered the wall as a barrier to ensure that the interpolation results closely represented the ground. The smoke density pattern throughout the simulation revealed that the areas related to sensor 5, sensor 6, sensor 7, sensor 9, and sensor 10 were the most intense spots, consistently intense from the beginning to the end. Additionally, the SDE analysis of the smoke data indicated a northwest-southeast distribution, consistent across the simulations despite being divided according to different timelines. Simulation 4 had the largest ellipse size, while simulation 1 represented the smallest. The findings of the study demonstrated that the proposed sensor prototype successfully detected early fire hazards in the UKM campus building. Moreover, the analysis of the sensor data revealed variations in sensitivity among different sensors, as indicated by the density estimation and standard deviation of the data. This research also contributes to the development of a smart campus framework, specifically focusing on the UKM campus building and its existing fire alarm system, which currently employs conventional methods. Overall, this research combines geospatial analysis and IoT-based sensor systems to develop a smart campus environment in UKM capable of early fire detection. The integration of geospatial techniques enhances the assessment of data quality and enables more effective monitoring and management of fire hazards within the UKM campus buildings.