Incorporating SLAM and mobile sensing for indoor CO2 monitoring and source position estimation

Abstract Indoor Air Quality (IAQ) monitoring is becoming an increasingly important topic since it critically affects occupants’ health, comfort, and safety, etc. Conventionally, monitoring sensors are distributed at stationary positions for measurement acquisitions, which cannot easily become wireless causing a high cost of infrastructure installation and maintenance. Furthermore, stationary sensing often incurs the problems of non-detection zones and low-granularity monitoring, especially in complex indoor scenarios where environmental factors are dynamic and non-convex functions. To address these critical issues, this paper proposes a mobile IAQ sensing using an automated robot equipped with one LIDAR and one CO2 sensor, to enable the prompt detection and positioning of contaminant sources (a CO2 contaminant source). Both stationary sensing (with 9 CO2 sensors) and mobile sensing (a robot with one CO2 sensor) are evaluated in real-world experiments carried out in a typical laboratory room. The spatiotemporal analysis demonstrates that the automated mobile sensing is capable of efficient and agile IAQ survey. The heat maps of the CO2 concentration illustrate that mobile sensing observes better accuracy and granularity of the pollutant detections, as mobile sensing estimates the position of the CO2 source about 1.83 meters away from the ground truth position, in contrast with stationary sensing of 3.1 meters. The proposed autonomous and mobile IAQ monitoring consumes much less infrastructure (one sensor instead of nine stationary sensors), implementation complexity, and amount of data for communication and processing.