Extracting point of interest and classifying environment for low sampling crowd sensing smartphone sensor data

Abstract

Abstract-The advancement of smartphones with various type of sensors enabled us to harness diverse information with crowd sensing mobile application. However, traditional approaches have suffered drawbacks such as high battery consumption as a trade off to obtain high accuracy data using high sampling rate. To mitigate the battery consumption, we proposed low sampling point of interest (POI) extraction framework, which is built upon validation based stay points detection (VSPD) and sensor fusion based environment classification (SFEC). We studied various of clustering algorithm and showed that density based spatial clustering of application with noise (DBSCAN) algorithms produce most accurate result among existing methods. The SFEC model is utilized for classifying the indoor or outdoor environment of the POI clustered earlier by VSPD. Real world data are collected, benchmarked using existing clustering method to denote effectiveness of low sampling rate model in high noise spatial temporal data.

Extracting Global Shipping Networks from Massive Historical Automatic Identification System Sensor Data: A Bottom-Up Approach

“…It performs relatively well to detect clusters with arbitrary shapes. It has proven to be computationally efficient for large datasets and also requires little inputs from domain knowledge [24,25]. Palma et al [21] use DBSCAN to detect stops by clustering low speed points into a single trajectory.…”

Column-Wise Element Selection for Computationally Efficient Nonnegative Coupled Matrix Tensor Factorization

“…The first smart city data (D1) is collected from the activities of Singapores elderly people living in the Bukit Panjang region. D1 was gathered using smartphone sensors that consist of 37 users home location, 1295 locations visited by them and the time [41]. We represent D1 in the tensor model as 37 × 1295 × 7 along with user-user physical distance matrix representation 37 × 37 to identify temporal patterns over 7 days of the week.…”

Identifying Indoor Points of Interest via Mobile Crowdsensing: An Experimental Study

“…According to the work done in [10], different clustering algorithms are experimentally evaluated, and DBSCAN [20] has been chosen as the preferred clustering technique due to its capabilities in forming arbitrary shape clusters. In this paper we propose a modified DBSCAN technique to cluster the raw Wi-Fi RSS measurements for a particular device in a given environment.…”

“…We exploit Louvain method for community detection [21] to obtain insights about popular POI among users. Let the number of POI in a given indoor environment is h, the number of pair-wise cosine similarities (I) is calculated as shown in the Equation 10.…”

“…Previous work has done to detect the surrounding environment type (e.g. Indoor/Outdoor) using smartphone based sensors [9], [10]. Understanding the lifestyle of older adults is conducted in [11] through smartphone application data.…”