Sensing and Classifying Roadway Obstacles in Smart Cities: The Street Bump System

IEEE Access

卷:4

Chris Osgood¹  Ioannis CH. Paschalidis²  Theodora S. Brisimi²  Christos G. Cassandras²  Yue Zhang² 

[1] City of Boston, Boston, MA, USA;

[2] Department of Electrical and Computer Engineering, Division of Systems Engineering, Center for Information and Systems Engineering, Boston University, Boston, MA, USA;

关键词: Classification;    anomaly detection;    machine learning;    smart cities;   

DOI  :  10.1109/ACCESS.2016.2529562

来源: DOAJ

【 摘 要 】

We develop an infrastructure-free approach for anomaly detection and identification based on data collected through a smartphone application (Street Bump). The approach is capable of effectively classifying roadway obstacles into predefined categories using machine learning algorithms, as well as prioritizing actionable ones in need of immediate attention based on a proposed anomaly index. We explore some novel variants of classification algorithms that combine clustering with classification and introduce appropriate regularization in order to concentrate on a sparse set of most relevant features, which has the effect of reducing overfitting. Furthermore, the anomaly index we introduce combines novel metrics of obstacle irregularity computed based on the data captured by the Street Bump smartphone application. Results on an actual data set provided by the City of Boston illustrate the feasibility and the effectiveness of our system in practice.

【 授权许可】

Unknown