“Occupant presence and behavior in buildings is considered a key element towards building intelligent and pervasive environments. Yet, practical applications of energy intelligent buildings typically suffer from high sensor unreliability,” highlights a recent paper that proposes a fundamentally new way of inferring occupancy in smart buildings.
The study has been able to demonstrate an additional 30% energy savings by propagating senor uncertainty and advocates the use of probabilistic data over traditional discrete classification outputs. This research has the potential to disrupt the popular approach to occupancy sensing and bring about new levels of energy efficiency.
The paper – Propagating sensor uncertainty to better infer office occupancy in smart building control – was authored by Charikleia Papatsimpa and Jean-Paul Linnartz of the Department of Electrical Engineering at the Eindhoven University of Technology, in the Netherlands. They propose a layered probabilistic framework for occupancy-based control in intelligent buildings. In this cascade of layers, where each layer addresses different aspects of the occupancy detection problem in a probabilistic manner rather than in a hard rule engine.
Over the past decade, numerous systems have been proposed for smart building control, for example:
- Energy Efficient building environment control strategies using real-time occupancy measurements by Erickson et al. (2009) developed a system that tracks user movement in building spaces using a camera network solution called SCOPES, where a framework, based on closest distance Markov chains is used to capture the dynamics of user occupancy, in order to save energy in various areas, namely HVAC and lighting.
- An Indoor Localization Algorithm for Lighting Control using RFID by Zhen et al. (2009) promotes the use of sensor badges for building occupants possible to achieve relatively accurate localization based on support vector machines (SVM).
Such systems, however, “are based on advanced sensor modalities, such as cameras, which are expensive and wearable devices, which may generate privacy concerns,” say Papatsimpa and Linnartz of the Eindhoven team. “The domain of building control can tolerate a loss in accuracy in favor of installation cost and privacy, thus, simple, binary sensors are preferred [in our approach], that is easy to retrofit in existing buildings and comply with the existing privacy regulations.”
Others have experimented with the use of radar sensing technology, which offers a promising approach to overcome the limitations of the more common occupancy sensors. In fact, we have seen an expansion in the use of radar sensors for occupancy and other applications in recent years:
- Noninvasive microwave measurement of respiration by J.C. Lin of Wayne State University in Detroit, conducted one of the first demonstrations of radar for human sensing but with a range far too low for most building occupancy applications.
- Evaluation of a Doppler radar sensor system for vital signs detection and activity monitoring in a radio-frequency shielded room by Kuutti et al. of Aalto University School of Electrical Engineering in Finland, presented a more advanced Doppler radar-based method for the detection of vital signs, presence and activity. However, an aluminum coating on surfaces was required for accurate detection.
- Real time indoor presence detection with a novel radar on a chip by Deiana et al. went so far as to identify the activity of occupants in a room, it could even pick up small movements such as typing to provide additional information to the system. The approach failed, however, to pick up on an occupant sitting completely still.
“In all cases, although microwave radar technology has gained increased popularity for activity detection, its feasibility and implementation in the context of smart building and control applications remains limited,” the Eindhoven researchers concluded. However, considering that most sensor-based approaches suffer from high sensor unreliability, “radar sensing technology offers a promising solution that still remains not fully explored,” they added as they set about to find a solution.
In order to reduce the measurement uncertainty, the researchers needed to apply stochastic models like Bayesian networks, Gaussian mixture models and support vector machines that have been previously implemented to improve occupancy detection, or they could approximate office occupancy as a Markov process that moves between a number of states. The challenge, therefore, was to determine the current state of that process from the sequence of observations made by these “imperfect” sensors using a concept is known as a Hidden Markov Model, which has been extensively explored for other applications.
The layered representation they developed allows for the separation of the architecture into separate building blocks, where, cleanly separated algorithms may run on different devices/entities. Their innovative solution combines the advantages of a hidden Markov model that naturally captures the temporal structure of office occupancy with Bayesian modeling, which captures the physical properties of the sensor network. The inferential results of each layer are then used as the input to the consequent layer.
“Our results showed significant energy saving potential in a typical open-plan office environment without sacrificing user comfort,” Papatsimpa and Linnartz said. “According to our simulation analysis presented in, the suggested solution was able to achieve energy savings of up to 30%, compared to baseline manual control, without too much energy wasted during un-occupied periods (false positives). Savings are expected to be higher when considering the energy wasted during non-occupied hours in commercial buildings.”
Their ongoing work, which includes efforts to fuse information from multiple sensor sources within the building in order to further improve the detection performance, could have a real impact on future occupancy sensing technology. In fact, their layered representation in combination with a communication protocol offers a flexible framework for combining information from heterogeneous sensor networks and ensures better coherence. This is one approach to watch in the the occupancy sensing space.