Rights statement: © ACM, 2019. This is the author's version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record was published in ACM Transactions on Spatial Algorithms and Systems (TSAS)
Accepted author manuscript, 9.13 MB, PDF document
Final published version
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - Crowd Replication
T2 - Sensing-Assisted Quantification of Human Behaviour in Public Spaces
AU - Hemminki, Samuli
AU - Kuribayashi, Keisuke
AU - Konomi, Shin'ichi
AU - Nurmi, Petteri Tapio
AU - Tarkoma, Sasu
N1 - © ACM, YYYY. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in PUBLICATION, {VOL#, ISS#, (DATE)} http://doi.acm.org/10.1145/nnnnnn.nnnnnn
PY - 2019/8/1
Y1 - 2019/8/1
N2 - A central challenge for public space design is to evaluate whether a given space promotes different types of activities. In this article, as our first contribution, we develop crowd replication as a novel sensor-assisted method for quantifying human behavior within public spaces. In crowd replication, a researcher is tasked with recording the behavior of people using a space while being instrumented with a mobile device that captures a sensor trace of the replicated movements and activities. Through mathematical modeling, behavioral indicators extracted from the replicated trajectories can be extrapolated to represent a larger target population. As our second contribution, we develop a novel highly accurate pedestrian sensing solution for reconstructing movement trajectories from sensor traces captured during the replication process. Our key insight is to tailor sensing to characteristics of the researcher performing replication, which allows reconstruction to operate robustly against variations in pace and other walking characteristics. We validate crowd replication through a case study carried out within a representative example of a metropolitan-scale public space. Our results show that crowd-replicated data closely mirrors human dynamics in public spaces and reduces overall data collection effort while producing high-quality indicators about behaviors and activities of people within the space. We also validate our pedestrian modeling approach through extensive benchmarks, demonstrating that our approach can reconstruct movement trajectories with high accuracy and robustness (median error below 1%). Finally, we demonstrate that our contributions enable capturing detailed indicators of liveliness, extent of social interaction, and other factors indicative of public space quality.
AB - A central challenge for public space design is to evaluate whether a given space promotes different types of activities. In this article, as our first contribution, we develop crowd replication as a novel sensor-assisted method for quantifying human behavior within public spaces. In crowd replication, a researcher is tasked with recording the behavior of people using a space while being instrumented with a mobile device that captures a sensor trace of the replicated movements and activities. Through mathematical modeling, behavioral indicators extracted from the replicated trajectories can be extrapolated to represent a larger target population. As our second contribution, we develop a novel highly accurate pedestrian sensing solution for reconstructing movement trajectories from sensor traces captured during the replication process. Our key insight is to tailor sensing to characteristics of the researcher performing replication, which allows reconstruction to operate robustly against variations in pace and other walking characteristics. We validate crowd replication through a case study carried out within a representative example of a metropolitan-scale public space. Our results show that crowd-replicated data closely mirrors human dynamics in public spaces and reduces overall data collection effort while producing high-quality indicators about behaviors and activities of people within the space. We also validate our pedestrian modeling approach through extensive benchmarks, demonstrating that our approach can reconstruct movement trajectories with high accuracy and robustness (median error below 1%). Finally, we demonstrate that our contributions enable capturing detailed indicators of liveliness, extent of social interaction, and other factors indicative of public space quality.
U2 - 10.1145/3317666
DO - 10.1145/3317666
M3 - Journal article
VL - 5
JO - ACM Transactions on Spatial Algorithms and Systems (TSAS)
JF - ACM Transactions on Spatial Algorithms and Systems (TSAS)
IS - 3
M1 - 15
ER -