Search

Effective early fire detection is crucial for preventing damage to people and buildings, especially in fire-prone historic structures. However, due to the infrequent occurrence of fire events throughout a building's lifespan, real-world data for training models are often sparse. In this study, we applied feature representation transfer and instance transfer in the context of early fire detection using multi-sensor nodes. The goal was to investigate whether training data from a small-scale setup (source domain) can be used to identify various incipient fire scenarios in their early stages within a full-scale test room (target domain). In a first step, we employed Linear Discriminant Analysis (LDA) to create a new feature space solely based on the source domain data and predicted four different fire types (smoldering wood, smoldering cotton, smoldering cable and candle fire) in the target domain with a classification rate up to 69% and a Cohen's Kappa of 0.58. Notably, lower classification performance was observed for sensor node positions close to the wall in the full-scale test room. In a second experiment, we applied the TrAdaBoost algorithm as a common instance transfer technique to adapt the model to the target domain, assuming that sparse information from the target domain is available. Boosting the data from 1% to 30% was utilized for individual sensor node positions in the target domain to adapt the model to the target domain. We found that additional boosting improved the classification performance (average classification rate of 73% and an average Cohen's Kappa of 0.63). However, it was noted that excessively boosting the data could lead to overfitting to a specific sensor node position in the target domain, resulting in a reduction in the overall classification performance. [ABSTRACT FROM AUTHOR]

Effective early fire detection is crucial for preventing damage to people and buildings, especially in fire-prone historic structures. However, due to the infrequent occurrence of fire events throughout a building’s lifespan, real-world data for training models are often sparse. In this study, we applied feature representation transfer and instance transfer in the context of early fire detection using multi-sensor nodes. The goal was to investigate whether training data from a small-scale setup (source domain) can be used to identify various incipient fire scenarios in their early stages within a full-scale test room (target domain). In a first step, we employed Linear Discriminant Analysis (LDA) to create a new feature space solely based on the source domain data and predicted four different fire types (smoldering wood, smoldering cotton, smoldering cable and candle fire) in the target domain with a classification rate up to 69% and a Cohen’s Kappa of 0.58. Notably, lower classification performance was observed for sensor node positions close to the wall in the full-scale test room. In a second experiment, we applied the TrAdaBoost algorithm as a common instance transfer technique to adapt the model to the target domain, assuming that sparse information from the target domain is available. Boosting the data from 1% to 30% was utilized for individual sensor node positions in the target domain to adapt the model to the target domain. We found that additional boosting improved the classification performance (average classification rate of 73% and an average Cohen’s Kappa of 0.63). However, it was noted that excessively boosting the data could lead to overfitting to a specific sensor node position in the target domain, resulting in a reduction in the overall classification performance.

Energy consumption in Wireless Sensor Networks (WSN) involving multiple sensor nodes is a crucial parameter in many applications like smart healthcare systems, home automation, environmental monitoring, and industrial use. Hence, an energy-efficient cluster-head (CH) selection strategy is imperative in a WSN to improve network performance. So to balance the harsh conditions in the network with fast changes in the energy dynamics, a novel energy-efficient adaptive fuzzy-based CH selection approach is projected. Extensive simulations exploited various real-time scenarios, such as varying the optimal position of the location of the base station and network energy. Additionally, the results showed an improved performance in the throughput (46%) and energy consumption (66%), which demonstrated the robustness and efficacy of the proposed model for the future designs of WSN applications. [ABSTRACT FROM AUTHOR]

The evaluation of the effects of air pollution on public health and human-wellbeing requires reliable data. Standard air quality monitoring stations provide accurate measurements of airborne pollutant levels, but, due to their sparse distribution, they cannot capture accurately the spatial variability of air pollutant concentrations within cities. Dedicated in-depth field campaigns have dense spatial coverage of the measurements but are held for relatively short time periods. Hence, their representativeness is limited. Moreover, the oftentimes integrated measurements represent time-averaged records. Recent advances in communication and sensor technologies enable the deployment of dense grids of Wireless Distributed Environmental Sensor Networks for air quality monitoring, yet their capability to capture urban-scale spatiotemporal pollutant patterns has not been thoroughly examined to date. Here, we summarize our studies on the practicalities of using data streams from sensor nodes for air quality measurement and the required methods to tune the results to different stakeholders and applications. We summarize the results from eight cities across Europe, five sensor technologies-three stationary (with one tested also while moving) and two personal sensor platforms, and eight ambient pollutants. Overall, few sensors showed an exceptional and consistent performance, which can shed light on the fine spatiotemporal urban variability of pollutant concentrations. Stationary sensor nodes were more reliable than personal nodes. In general, the sensor measurements tend to suffer from the interference of various environmental factors and require frequent calibrations. This calls for the development of suitable field calibration procedures, and several such in situ field calibrations are presented. [ABSTRACT FROM AUTHOR]

The evaluation of the effects of air pollution on public health and human-wellbeing requires reliable data. Standard air quality monitoring stations provide accurate measurements of airborne pollutant levels, but, due to their sparse distribution, they cannot capture accurately the spatial variability of air pollutant concentrations within cities. Dedicated in-depth field campaigns have dense spatial coverage of the measurements but are held for relatively short time periods. Hence, their representativeness is limited. Moreover, the oftentimes integrated measurements represent time-averaged records. Recent advances in communication and sensor technologies enable the deployment of dense grids of Wireless Distributed Environmental Sensor Networks for air quality monitoring, yet their capability to capture urban-scale spatiotemporal pollutant patterns has not been thoroughly examined to date. Here, we summarize our studies on the practicalities of using data streams from sensor nodes for air quality measurement and the required methods to tune the results to different stakeholders and applications. We summarize the results from eight cities across Europe, five sensor technologies-three stationary (with one tested also while moving) and two personal sensor platforms, and eight ambient pollutants. Overall, few sensors showed an exceptional and consistent performance, which can shed light on the fine spatiotemporal urban variability of pollutant concentrations. Stationary sensor nodes were more reliable than personal nodes. In general, the sensor measurements tend to suffer from the interference of various environmental factors and require frequent calibrations. This calls for the development of suitable field calibration procedures, and several such in situ field calibrations are presented.