Center for Machine Learning and Intelligent Systems
About  Citation Policy  Donate a Data Set  Contact


Repository Web            Google
View ALL Data Sets

× Check out the beta version of the new UCI Machine Learning Repository we are currently testing! Contact us if you have any issues, questions, or concerns. Click here to try out the new site.

Room Occupancy Estimation Data Set
Download: Data Folder, Data Set Description

Abstract: Data set for estimating the precise number of occupants in a room using multiple non-intrusive environmental sensors like temperature, light, sound, CO2 and PIR.

Data Set Characteristics:  

Multivariate, Time-Series

Number of Instances:

10129

Area:

Computer

Attribute Characteristics:

Real

Number of Attributes:

16

Date Donated

2021-01-16

Associated Tasks:

Classification

Missing Values?

N/A

Number of Web Hits:

7128


Source:

Adarsh Pal Singh (IIIT Hyderabad, India): adarshpal.singh '@' alumni.iiit.ac.in
Dr. Sachin Chaudhari (IIIT Hyderabad, India): sachin.c '@' iiit.ac.in


Data Set Information:

The experimental testbed for occupancy estimation was deployed in a 6m × 4.6m room. The setup consisted of 7 sensor nodes and one edge node in a star configuration with the sensor nodes transmitting data to the edge every 30s using wireless transceivers. No HVAC systems were in use while the dataset was being collected.

Five different types of non-intrusive sensors were used in this experiment: temperature, light, sound, CO2 and digital passive infrared (PIR). The CO2, sound and PIR sensors needed manual calibration. For the CO2 sensor, zero-point calibration was manually done before its first use by keeping it in a clean environment for over 20 minutes and then pulling the calibration pin (HD pin) low for over 7s. The sound sensor is essentially a microphone with a variable-gain analog amplifier attached to it. Therefore, the output of this sensor is analog which is read by the microcontroller’s ADC in volts. The potentiometer tied to the gain of the amplifier was adjusted to ensure the highest sensitivity. The PIR sensor has two trimpots: one to tweak the sensitivity and the other to tweak the time for which the output stays high after detecting motion. Both of these were adjusted to the highest values. Sensor nodes S1-S4 consisted of temperature, light and sound sensors, S5 had a CO2 sensor and S6 and S7 had one PIR sensor each that were deployed on the ceiling ledges at an angle that maximized the sensor’s field of view for motion detection.

The data was collected for a period of 4 days in a controlled manner with the occupancy in the room varying between 0 and 3 people. The ground truth of the occupancy count in the room was noted manually.

Please refer to our publications for more details.


Attribute Information:

Date: YYYY/MM/DD
Time: HH:MM:SS
Temperature: In degree Celsius
Light: In Lux
Sound: In Volts (amplifier output read by ADC)
CO2: In PPM
CO2 Slope: Slope of CO2 values taken in a sliding window
PIR: Binary value conveying motion detection
Room_Occupancy_Count: Ground Truth


Relevant Papers:

1. Adarsh Pal Singh, Vivek Jain, Sachin Chaudhari, Frank Alexander Kraemer, Stefan Werner and Vishal Garg, “Machine Learning-Based Occupancy Estimation Using Multivariate Sensor Nodes,” in 2018 IEEE Globecom Workshops (GC Wkshps), 2018.
2. Adarsh Pal Singh, 'Machine Learning for IoT Applications: Sensor Data Analytics and Data Reduction Techniques', Masters Thesis, [Web Link], 2020.



Citation Request:

If you use this dataset in your research, please cite the following paper:
Adarsh Pal Singh, Vivek Jain, Sachin Chaudhari, Frank Alexander Kraemer, Stefan Werner and Vishal Garg, “Machine Learning-Based Occupancy Estimation Using Multivariate Sensor Nodes,” in 2018 IEEE Globecom Workshops (GC Wkshps), 2018.


Supported By:

 In Collaboration With:

About  ||  Citation Policy  ||  Donation Policy  ||  Contact  ||  CML