This thesis documents a sensor networking project with an interest in internal environment monitoring in relation to thermal comfort. As part of this project sensor nodes were designed, built and deployed. Data was collected from the nodes via a wired Ethernet network and was stored in a database. The network remains operational several years after its initial deployment. The collected data was analyzed in conjunction with data from a local meteorological station and the building's smart fiscal energy meters. The analysis suggests the possibility of automated thermal comfort classification using data from a sensor network.

The ways that people use libraries have changed drastically over the past few decades. Proliferation of computers and the internet have led to the purpose of libraries expanding from being only places where information is stored, to spaces where people teach, learn, create, and collaborate. Due to this, the ways that people occupy the space in a library have also changed. To keep up with these changes and improve patron experience, institutions collect data to determine how their spaces are being used. This thesis involves the development a system that collects, stores, and analyzes data relevant to occupancy to learn how a space is being utilized. Data is collected from a temperature and humidity sensor, passive Infrared sensor, and an Infrared thermal sensor array to observe people as they occupy and move through a space. Algorithms were developed to analyze the collected sensor data to determine how many people are occupying a space or the directions that people are moving through a space. The algorithms demonstrate the ability to track multiple people moving through a space as well as count the number of people in a space with an RMSE of roughly 0.39 people.

Accelerometers deployed within smart structures produce a wealth of vibration data that can be analyzed to infer information about the types of acceleration events that are occurring within the structure. In the case of monitored smart buildings, some of these acceleration events are linked to occupant behavior, such as walking, operating machinery, closing doors, etc. The identification and classification of such events has many potential applications within a smart structure or city. Understanding occupant patterns could be beneficial for operations, retail, or HVAC management, as it could be used to monitor occupancy flow with a relatively sparse sensor network. It may also have detrimental implications in terms of cybersecurity, where such information could be mined for malicious practices if unauthorized access to the data was obtained. This work presents methods for the detection and classification of vibration events in an experimental smart building, Goodwin Hall at Virginia Tech. Goodwin Hall's 200+ accelerometer network is used to gather acceleration data, from which vibration events are automatically detected and clustered. The presence of a vibration event is detected from a raw acceleration signal with an adaptive RMS threshold method. A feature vector is then created for each extracted event as areas under regions of the FFT of the event's acceleration signal. The feature vectors are then mapped into a low-dimensional space using principal component analysis, where they are clustered with various unsupervised algorithms. These processes have shown to be successful when gathering vibration events from a single-sensor setup, but pose challenges when expanded to a multi-sensor network. Because of this, expanded applications such as a semi-supervised classifier for events detected anywhere in the building are currently still under development. This semi-supervised process, combined with the known location of each sensor would allow inferences to be drawn about the frequency of different activity types in regions of the building not captured in the labeled data. Future work intends to address these multi-sensor challenges with adjustments to the algorithm process. / Master of Science / All objects experience vibrations when they are disturbed by some force. In the case of this work, the object is complex, a classroom building, but the principle still stands. When the building is disturbed by a force it will vibrate, even if the force is small, such as a person walking down a hallway or closing a door. The vibrations caused by these 'events' are unique to the type of event, that is, footstep vibrations will be different from door vibrations. These vibrations are observed with accelerometers, and the corresponding signal is used to determine what type of event caused the vibration. First, an event is automatically detected within the signal and separated from it. Second, characteristics unique to the signal are identified, a process known as 'feature extraction.' Finally, those features are used to distinguish the event from others and to identify what had caused it based on previous experimental data. The ability to detect these events and classify them introduces many interesting applications, including any that would stem from occupant detection, including improved security or operations, retail, or HVAC management. The methods here may also be applicable to other applications, such as monitoring bridges and machinery, or for developing cutting-edge smartphone applications with the accelerometer that is built in.

Occupancy monitoring is an important research problem with a broad range of applications in security, surveillance, and resource management in smart building environments. As a result, it has immediate solutions to solving some of society's most pressing issues. For example, HVAC and lighting systems in the US consume approximately 45-50% of the total energy a building uses. Smart buildings can reduce wasted energy by incorporating networkable occupancy sensors to obtain real-time occupancy data for the facilities. Therefore, occupancy monitoring systems can enable significant cost savings and carbon reduction. In addition, workplaces have quickly adapted and implemented COVID-19 safety measures by preventing overcrowding using real-time information on people density. While there are many sensors, RGB cameras have proven to be the most accurate. However, cameras create privacy concerns. Hence, our research aims to design an efficient occupancy monitoring system with minimal privacy invasion. We conducted a systematic study on sensor characterization using various low-resolution infrared sensors and proposed a unified processing algorithms pipeline for occupancy estimation. This research also investigates low-resolution thermal imaging sensors with a chessboard reading pattern, focusing on algorithm design issues and proposing solutions when detecting moving objects. Our proposed approach achieves about 99% accuracy in occupancy estimation, enabling many practical smart building applications. We also added additional sensors to our system using sensor fusion technology to boost its functionality and demonstrated the system's feasibility by deploying it in different real-time application scenarios.