Self-organization of collective escape in pigeon flocks

Bird flocks under predation demonstrate complex patterns of collective escape. These patterns may emerge by self-organization from local interactions among group-members. Computational models have been shown to be valuable for identifying what behavioral rules may govern such interactions among individuals during collective motion. However, our knowledge of such rules for collective escape is limited by the lack of quantitative data on bird flocks under predation in the field. In the present study, we analyze the first GPS trajectories of pigeons in airborne flocks attacked by a robotic falcon in order to build a species-specific model of collective escape. We use our model to examine a recently identified distance-dependent pattern of collective behavior: the closer the prey is to the predator, the higher the frequency with which flock members turn away from it. We first extract from the empirical data of pigeon flocks the characteristics of their shape and internal structure (bearing angle and distance to nearest neighbors). Combining these with information on their coordination from the literature, we build an agent-based model adjusted to pigeons’ collective escape. We show that the pattern of turning away from the predator with increased frequency when the predator is closer arises without prey prioritizing escape when the predator is near. Instead, it emerges through self-organization from a behavioral rule to avoid the predator independently of their distance to it. During this self-organization process, we show how flock members increase their consensus over which direction to escape and turn collectively as the predator gets closer. Our results suggest that coordination among flock members, combined with simple escape rules, reduces the cognitive costs of tracking the predator while flocking. Such escape rules that are independent of the distance to the predator can now be investigated in other species. Our study showcases the important role of computational models in the interpretation of empirical findings of collective behavior.

Trip Purpose Imputation Using GPS Trajectories with Machine Learning

We studied trip purpose imputation using data mining and machine learning techniques based on a dataset of GPS-based trajectories gathered in Switzerland. With a large number of labeled activities in eight categories, we explored location information using hierarchical clustering and achieved a classification accuracy of 86.7% using a random forest approach as a baseline. The contribution of this study is summarized below. Firstly, using information from GPS trajectories exclusively without personal information shows a negligible decrease in accuracy (0.9%), which indicates the good performance of our data mining steps and the wide applicability of our imputation scheme in case of limited information availability. Secondly, the dependence of model performance on the geographical location, the number of participants, and the duration of the survey is investigated to provide a reference when comparing classification accuracy. Furthermore, we show the ensemble filter to be an excellent tool in this research field not only because of the increased accuracy (93.6%), especially for minority classes, but also the reduced uncertainties in blindly trusting the labeling of activities by participants, which is vulnerable to class noise due to the large survey response burden. Finally, the trip purpose derivation accuracy across participants reaches 74.8%, which is significant and suggests the possibility of effectively applying a model trained on GPS trajectories of a small subset of citizens to a larger GPS trajectory sample.

Impact Assessing of Traffic Lights via GPS Vehicle Trajectories

The adaptability of traffic lights in the control of vehicle traffic heavily affects the trafficability of vehicles and the travel efficiency of traffic participants in busy urban areas. Existing studies mainly have focused on the presence of traffic lights, but rarely evaluate the impact of traffic lights by analyzing traffic data, thus there is no solution for practicably and precisely self-regulating traffic lights. To address these issues, we propose a low-cost and fast traffic signal detection and impact assessment framework, which detects traffic lights from GPS trajectories and intersection features in a supervised way, and analyzes the impact range and time of traffic lights from intersection track data segments. The experimental results show that our approach gains the best AUC value of 0.95 under the ROC standard classification and indicates that the impact pattern of traffic lights at intersections is high related to the travel rule of traffic participants.

Similarity Matrix-Based Anomaly Detection for Clinical Intervention

The use of digital phenotyping methods in clinical care has allowed for improved investigation of spatiotemporal behaviors of patients. Moreover, detecting abnormalities in mobile sensor data patterns can be instrumental in identifying potential changes in symptomology. We propose a method that temporally aligns sensor data in order to achieve interpretable measures of similarity. These computed measures can then be used for anomaly detection, baseline routine computation, and trajectory clustering. In addition, we apply this method on a study of 695 college participants, as well as on a patient with worsening anxiety and depression. With varying temporal constraints, we find mild correlations between changes in routine and clinical scores. Furthermore, in our experiment on an individual with elevated depression and anxiety, we are able to cluster GPS trajectories, allowing for improved understanding and visualization of routines with respect to symptomology. In the future, we aim to apply this method on individuals that undergo data collection for longer periods of time, thus allowing for a better understanding of long-term routines and signals for clinical intervention.

Assessing social anxiety using gps trajectories and point-of-interest data

Mental health problems are highly prevalent and appear to be increasing in frequency and severity among the college student population. The upsurge in mobile and wearable wireless technologies capable of intense, longitudinal tracking of individuals, provide valuable opportunities to examine temporal patterns and dynamic interactions of key variables in mental health research. In this paper, we present a feasibility study leveraging non-invasive mobile sensing technology to passively assess college students' social anxiety, one of the most common disorders in the college student population. We have first developed a smartphone application to continuously track GPS locations of college students, then we built an analytic infrastructure to collect the GPS trajectories and finally we analyzed student behaviors (e.g. studying or staying at home) using Point-Of-Interest (POI). The whole framework supports intense, longitudinal, dynamic tracking of college students to evaluate how their anxiety and behaviors change in the college campus environment. The collected data provides critical information about how students' social anxiety levels and their mobility patterns are correlated. Our primary analysis based on 18 college students demonstrated that social anxiety level is significantly correlated with places students' visited and location transitions.

Automatic generation of road network data from taxi GPS trajectories

Keeping road network databases up-to-date is crucial to Geographical Information System (GIS) applications such as road networking. The vector road centerlines extracted from field surveys and satellite images are expensive and labor intensive with long updating processes. The GPS data crowd-sourced by public transportation users, provides an expanding source for enhancing road maps because of its rich spatial-temporal coverage and reasonable level of accuracy. The overall objective of this project is to implement an optimized methodology, which generates road centerline from GPS data obtained from taxis in Beijing without using any reference plans. Since the dataset used in this project has longer time intervals between trajectories compared to previous studies, the extracted road network on straight road segments are more accurate than the extracted road network on highway ramps in this project.