Integrated Use of Agent Based Modelling (ABM) with Wearable Sensors for Personal Exposure Assessment

Abstract

Considering the substantial hurdles involved in collecting individual data for whole populations, this study introduces a first approach of simulating human movement and interaction behaviour, using Agent Based Modelling (ABM). A city scale ABM was developed for the urban area of Thessaloniki, Greece. The model feeds into population-based exposure assessment, basing its estimations onto emerging behaviours of all the heterogeneous entities (agents) that compose the city-system. Population statistics, road and buildings networks data were transformed into human, road and building agents respectively. Time-use survey outputs and personal sensor collected data were associated with human agent rules, aiming to model representative to real-world routines. In addition, time-geography of exposure data, derived from a personal sensors campaign on 100 households, was used to inform the model. As a prevalence of an agent-specific decision-making, virtual individuals of different sociodemographic backgrounds express different spatiotemporal behaviours and their trajectories are coupled with spatially resolved pollution levels. Personal exposure was evaluated by assigning PM2.5 concentrations to human agents based on coordinates, type of location and intensity of encountered activities. Study results indicated that PM2.5 inhalation adjusted exposure between housemates can differ by 32% whereas exposure between two neighbours can vary by as much as 77%, due to the prevalence of different behaviours. This approach permits the cost-effective construction of time-activity diaries and daily exposure profiles, considering different microenvironments and socioeconomic characteristics. The proposed method leads to a refined exposure assessment model, addressing effectively vulnerable subgroups of population. It can serve as a tool to evaluate probable impacts of public health policies prior to implementation, reducing the time and cost required to identify effective measures.