Daily Living Movement Recognition for Pedestrian Dead Reckoning Applications

Abstract

Nowadays, activity recognition is a central topic in numerous applications such as patient and sport activity monitoring, surveillance, and navigation. By focusing on the latter, in particular Pedestrian Dead Reckoning navigation systems, activity recognition is generally exploited to get landmarks on the map of the buildings in order to permit the calibration of the navigation routines. The present work aims to provide a contribution to the definition of a more effective movement recognition for Pedestrian Dead Reckoning applications. The signal acquired by a belt-mounted triaxial accelerometer is considered as the input to the movement segmentation procedure which exploits Continuous Wavelet Transform to detect and segment cyclic movements such as walking. Furthermore, the segmented movements are provided to a supervised learning classifier in order to distinguish between activities such as walking and walking downstairs and upstairs. In particular, four supervised learning classification families are tested: decision tree, Support Vector Machine,k-nearest neighbour, and Ensemble Learner. Finally, the accuracy of the considered classification models is evaluated and the relative confusion matrices are presented.

Probabilistic Context-Aware Step Length Estimation for Pedestrian Dead Reckoning

“…The proposed WC-SLE algorithm is compared with the highest context-based SLE (HC-SLE) algorithm, which only considers the traveled distance corresponding to the context with the highest probability. The SLE routines are part of a PDR framework, which also defines the procedures of step detection and context classification, as shown in Fig.2 The step detection algorithm aims to identify the step time boundaries by performing the continuous wavelet transform (CWT) analysis of the specific force signal [30], [31]. The context classification algorithm uses the RVM method to classify a fixed portion of data, which generally may include several steps, into several contexts, each with an associated probabilities.…”

“…Since the step detection is the primary stage in a PDR process, either false or missed step detections can strongly affect the estimation of the traveled distance. In this work, the step detection algorithm exploits the continuous wavelet transform (CWT) analysis [30] and follows the block diagram illustrated in Fig.7. The specific force is used from multiple epoch over a 50% overlapped sliding time window of 1-second duration.…”

Road Surface Anomaly Assessment Using Low-Cost Accelerometers: A Machine Learning Approach

“…Therefore, although effective, their use is limited along minor roads (e.g., urban/local rural roads), which represent the most common type of road [ 22 ]. Many researchers have exploited acceleration signals for motion detection and classification in pedestrian contexts [ 23 , 24 , 25 ], whereas few studies have analyzed the results obtained from several vehicles equipped with accelerometers for data acquisition and specific algorithms for data processing. From this perspective, in recent years the reliability of accelerometers, GPS and even smartphone sensors has been exploited to address this issue.…”