Modern Information Retrieval
Using GPS to learn significant locations and predict movement across multiple users
Personal and Ubiquitous Computing
Modeling Multiple Time Series for Anomaly Detection
ICDM '05 Proceedings of the Fifth IEEE International Conference on Data Mining
Global distance-based segmentation of trajectories
Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining
Learning and inferring transportation routines
Artificial Intelligence
Efficient anomaly monitoring over moving object trajectory streams
Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining
Motion-Alert: automatic anomaly detection in massive moving objects
ISI'06 Proceedings of the 4th IEEE international conference on Intelligence and Security Informatics
EDBT'06 Proceedings of the 10th international conference on Advances in Database Technology
Predestination: inferring destinations from partial trajectories
UbiComp'06 Proceedings of the 8th international conference on Ubiquitous Computing
Hi-index | 0.00 |
Persons with mental impairments tend to be viewed as unemployable and systematically excluded from labor markets. However, this assumption has been challenged recently after the development of community rehabilitation, and supported employment services in particular. With sufficient and appropriate support on the job, many people with mental illness are capable of participating in the world of work to various levels, which not only provides them with financial support but also opportunity for social integration. Social integration includes community-based living, recreation and leisure pursuits, use of community services, and independent movement in and around the community through the use of public transportation. Coupled with this increased independence and integration is risk. With repeated training continued with daily practice, the individuals usually have no problems of getting lost or disoriented. However, there are occasions that individuals do forget how to travel to and from work. For example, part-timers with fewer shifts have more chances of running into transportation problems because they forget the routes. For places with many distractions, few landmarks that can help remain oriented, or surroundings that look similar, the situations can become worse. To decrease the risk for victimization of individuals with disabilities as they increasingly participate in their communities and seek social inclusion, mobile technology is used with a focus on increased autonomous functioning. Because individuals with disabilities are frequently dependent on others for support across environments, strategies and skills must be introduced that directly lead to access of those supports. Care providers are at the center of the support they have. Unfortunately, their care providers are often overloaded by too much work that results from pre-service training, coaching, escorting them to workplaces, and so on. To relieve their care providers from labor-intensive aids with traveling back and forth from work, we propose a PDA that is carried by the individual who has cognitive impairments. The PDA enables individuals to respond to unexpected situations such as being lost by effectively using the handheld device to alert themselves or call for assistance from the support system. In this article, we build a real-time anomaly detection system and conduct field experiments in communitybased settings for individuals with cognitive impairments. Deviation detection considers trajectories as input and tries to identify anomaly in following normal routes such as taking public transportation from the workplace to home or vice versa. Trajectory clustering is one of the methods of identifying normal routes from abnormal ones [9, 11]. However, trajectory clustering is performed only when trajectories of trips are completed. This makes the clustering algorithms less useful in identifying anomaly and discovering route deviation in the real time. Predicting personal destinations is the main principle behind most previous work in pervasive computing on modeling and predicting transportation routines. Many studies in this area share the trait that candidate destinations are extracted from GPS histories, i.e. places that subjects have actually visited [2, 8, 10]. In contrast, our work in anomaly detection is more focused on extracting anomaly at times users are trying to follow transportation routines, not necessarily inferring personal destinations.