Accurate building occupancy estimation with inhomogeneous Markov chain

• المؤلف الرئيسي: Lee, Jun Hao
• مؤلفون آخرون: Soh Yeng Chai
• التنسيق: Final Year Project
• اللغة: English
• منشور في: Nanyang Technological University 2022
• الموضوعات: Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/157505

The energy consumption of a high-rise structure increases as the urban population grows. The knowledge of a building's occupancy is critical since it has a significant impact on the building's energy usage. To avoid the situation worsening, it is critical to create an occupancy model to address the issue of building energy efficiency. While there are several methods for estimating a building's occupancy, each has its own set of disadvantages. As a result, a less invasive and more accurate method for assessing interior occupancy is critical. This work studies the use of an inhomogeneous Markov chain to anticipate occupancy in a multi-occupant single zone (MOSZ) situation. This experiment's MOSZ scenario is restricted to an NTU Hive lecture room. The number of occupants in the room is counted by PIR sensors and the counts will be served as the states of the Markov chain. Based on the room's actual occupancy measurements, MATLAB simulations are conducted to forecast occupancy. The model's performance is measured using the mean occupancy, the initial arrival time, the continuous occupation time, the number of high occurrences, and the transitions between vacant and occupied states. The normalized root mean square error (NRSME) is used to assess the model's performance. The result of initial arrival and continuous occupation duration are positive, but the others are not that good. These difficulties may be addressed with a larger dataset, and with corrections to the MATLAB code.