cs.CVFeb 24, 2026

Optimizing Occupancy Sensor Placement in Smart Environments

Authors: Hao Lu, Richard J. Radke

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Understanding the locations of occupants in a commercial built environment is critical for realizing energy savings by delivering lighting, heating, and cooling only where it is needed. The key to achieving this goal is being able to recognize zone occupancy in real time, without impeding occupants' activities or compromising privacy. While low-resolution, privacy-preserving time-of-flight (ToF) sensor networks have demonstrated good performance in zone counting, the performance depends on careful sensor placement. To address this issue, we propose an automatic sensor placement method that determines optimal sensor layouts for a given number of sensors, and can predict the counting accuracy of such a layout. In particular, given the geometric constraints of an office environment, we simulate a large number of occupant trajectories. We then formulate the sensor placement problem as an integer linear programming (ILP) problem and solve it with the branch and bound method. We demonstrate the effectiveness of the proposed method based on simulations of several different office environments.

Figures & tables

Figure 1: (a) An example floor plan of an office suite. (b) The color-labeled floor plan of the office suite. There are 6 labels: walls (black), obstacles (grey), doorways (brown), areas of interest (red), zone boundaries (green), and walkable regions (white).
Figure 2: Shortest paths between the same two endpoints in cases that without random obstacles (a) and with random obstacles (b).
Figure 3: Shortest paths between the same two endpoints without the doorway penalty (blue) and with the doorway penalty (yellow).
Figure 4: Shortest paths between the same two endpoints without the wall penalty (blue) and with the wall penalty (yellow).
Figure 5: A heatmap of 3000 randomly generated occupant trajectories in an office suite.
Figure 6: Sensor placement (blue boxes) based on the strategy of covering the densest cells.
Figure 7: Sensor placement (blue boxes) based on the strategy of maximally covering the intersection points between occupant trajectories and doorways.
Figure 8: Shortcomings of the strategy of maximally covering the intersection points between trajectories and doorways, in a scenario with two zone boundaries. (a) Two sensors; both zones are independently covered. (b) and (c) One sensor; one zone is not sensed. (d) A better placement strategy with one sensor to cover both zones. Note that I put all the caption info into the bottom caption; I prefer this and it’s also easier syntax.
Figure 9: Zone boundary dilation and path filtering. (a) Example layout with one zone boundary. (b) The zone boundary is dilated by ff meters. (c) Paths intersecting the dilated zone boundary. (d) Only path segments that transition the zone boundary are retained.
Figure 10: Sensor placement (blue boxes) with our proposed method.
Figure 11: Several of the floor plans used to test the proposed method.
Figure 12: Unity 3D simulation automatically created from the floor plan in Figure 11 d.
Figure 13: Sensor placement using proposed method with 7 sensor in Office 4.
Table I: Coverage rate and CCR for different layout of office environments.
Figure 14: The variance of classification correct rate against parameter f (m).
Figure 15: Choosing the number of sensors in Office 2 by trading off performance and cost.
Figure 16: Comparing CCR performance when the users move exactly according to the input model (blue lines), to different areas of interest (red lines), or totally random locations (green lines).

Explore similar work

CardsList

  1. Experimental study on surveillance video-based indoor occupancy measurement with occupant-centric control

    Mar 27, 2026Irfan Qaisar, Kailai Sun, Qingshan Jia +1Video SurveillanceBuilding Energy Management

    Center for Intelligent and Networked Systems, Department of Automation, BNRist, Tsinghua University, Beijing, 100084, China · Urban Mobility Lab, Massachusetts Institute of Technology, Cambridge, MA 02139, USA · Singapore-MIT Alliance for Research and Technology Centre, Massachusetts Institute of Technology, 138602, Singapore

  2. FLM-Occ: Feed-forward Likelihood Maximization for Efficient Indoor Occupancy Prediction

    Jun 19, 2026Guangcheng Chen, Lihuang Fang, Huaqi Tao +3Superquadric Representation3D Occupancy Prediction

    Southern University of Science and Technology · Shenzhen Key Laboratory of Robotics and Computer Vision · Guangdong Laboratory of Artificial Intelligence and Digital Economy (Shenzhen)

  3. Data-driven Optimal Sensor Placement for Aerosol Monitoring in the UK

    Oct 8, 2026Talha Ansar, Waqar Muhammad Ashraf, Alejandro Coca-Castro +3Bayesian Experimental Design

    University of Engineering and Technology, Lahore, Pakistan · The Alan Turing Institute, London, UK · University of Cambridge, Cambridge, UK