Activity-Travel Pattern Monitoring Framework in Modern City Environments

Abstract

Understanding the various aspects of complicated modern cities has long been a pivotal issue in diverse application domains. Activity-Travel Pattern (ATP) is one of the most powerful tools to define and understand city residents’ everyday lives in domains such as transportation, urban planning and geography. ATP represents where people are located, how they move in urban areas, and which activities they do there in certain patterns. As modern cities encompasses more people, vehicles, spaces and services, understanding ATP becomes more useful yet challenging. Because of its practical importance, governmental organizations have periodically observed and studied ATPs in terms of space-time use of citizens over major cities. These observations have been performed through manual questionnaires and annotations, requiring huge amounts of time, labor and expense. Advances in mobile and ubiquitous computing technologies open up the opportunity for real-time monitoring of ATP. It will significantly enhance the scale and timeliness of the ATP observation compared to conventional manual methods. More important, it will incubate a number of new advanced ATP-based urban services such as predictive mobile advertising, dynamic urban facility relocation, and adaptive transportation scheduling. To support a number of future ATP-based services effectively, we develop a high-performance ATP monitoring framework, named ActraMon, which provides the infrastructural support of ATP monitoring. It helps service providers easily launch new services without building their own ATP monitoring systems. Basically, ActraMon provides a simple but effective computational model of ATP and a declarative query language facilitating the effective specification of various monitoring semantics. ActraMon also employs a novel shared staging architecture to process ATP monitoring queries in a highly efficient and scalable manner. The architecture of ActraMon achieves the high performance in two aspects....