Mining Emerging Patterns for recognizing activities of multiple users in pervasive computing

Abstract

Understanding and recognizing human activities from sensor readings is an important task in pervasive computing. Existing work on activity recognition mainly focuses on recognizing activities for a single user in a smart home environment. However, in real life, there are often multiple inhabitants live in such an environment. Recognizing activities of not only a single user, but also multiple users is essential to the development of practical context-aware applications in pervasive computing. In this paper, we investigate the fundamental problem of recognizing activities for multiple users from sensor readings in a home environment, and propose a novel pattern mining approach to recognize both single-user and multi-user activities in a unified solution. We exploit Emerging Patterna type of knowledge pattern that describes significant changes between classes of data -for constructing our activity models, and propose an Emerging Pattern based Multi-user Activity Recognizer (epMAR) to recognize both single-user and multiuser activities. We conduct our empirical studies by collecting real-world activity traces done by two volunteers over a period of two weeks in a smart home environment, and analyze the performance in detail with respect to various activity cases in a multi-user scenario. Our experimental results demonstrate that our epMAR recognizer achieves an average accuracy of 89.72% for all the activity cases.

Discovery of topological relations for spatial Activity Recognition

“…Other researchers are trying to reach a step beyond by developing solution to the recognition of activities realized by multiple users [26,28] or accomplished simultaneously (concurrent or interleaved ADLs). They do not achieve the same level of performance we describe in this paper and are far from solving the various issues caused by these challenges.…”

Activity recognition in smart homes with self verification of assignments

“…We apply leave one day out cross validation, a stringent test using single day data for testing and the remaining for training and the process is repeated for all days. In Kasteren dataset, ARSH-SV attains high F1-score in all the seven activities compared to no time EDN, NB and J48-DT and for five activities compared to AR-SPM, while it shows comparatively less F1-score in the dinner activity compared to temporal EDN and AR-SPM, due to the similarity with breakfast and the availability of less (9) instances for training. In Kyoto7, ARSH-SV achieves better F1-score in most of the activities while comparable to AR-SPM in a few.…”

“…The obtained sensor data is partitioned into multiple segments in order to map them to the activity descriptions known as activity segmentation, where a segment is a consecutive sequence of time instants during which an activity is performed [6]. Activity segmentation is performed using different techniques, sliding windows [7], relative weighting of objects in adjacent activities [8] or pattern mining [9], just to name a few. Segmented activity instances are classified in activity classes using different learning models such as Hidden Markov Model (HMM) [10], Conditional Random Fields (CRF) [11], Naive Bayes (NB) [12], Support Vector Machine (SVM) [13], Artificial Neural Network (ANN) [14,15], and Decision Tree (DT) [16].…”

An Experiment in Hierarchical Recognition of Group Activities Using Wearable Sensors

“…The group activity is not necessarily the same as the sum of the activities of the individuals in it [4]. This implies that the activity or context of a group is a function of the activity or context of each individual in the group, in the same way that the context or activity of a single user is a function of sensor measurements.…”