We propose a walking distance estimation method based on an adaptive step-length estimator at various walking speeds using a smartphone. First, we apply a fast Fourier transform (FFT)-based smoother on the acceleration data collected by the smartphone to remove the interference signals. Then, we analyze these data using a set of step-detection rules in order to detect walking steps. Using an adaptive estimator, which is based on a model of average step speed, we accurately obtain the walking step length. To evaluate the accuracy of the proposed method, we examine the distance estimation for four different distances and three speed levels. The experimental results show that the proposed method significantly outperforms conventional estimation methods in terms of accuracy.

This paper proposes a novel method of estimating walking distance based on a precise counting of walking strides using insole sensors. We use an inertial triaxial accelerometer and eight pressure sensors installed in the insole of a shoe to record walkers’ movement data. The data is then transmitted to a smartphone to filter out noise and determine stance and swing phases. Based on phase information, we count the number of strides traveled and estimate the movement distance. To evaluate the accuracy of the proposed method, we created two walking databases on seven healthy participants and tested the proposed method. The first database, which is called the short distance database, consists of collected data from all seven healthy subjects walking on a 16 m distance. The second one, named the long distance database, is constructed from walking data of three healthy subjects who have participated in the short database for an 89 m distance. The experimental results show that the proposed method performs walking distance estimation accurately with the mean error rates of 4.8% and 3.1% for the short and long distance databases, respectively. Moreover, the maximum difference of the swing phase determination with respect to time is 0.08 s and 0.06 s for starting and stopping points of swing phases, respectively. Therefore, the stride counting method provides a highly precise result when subjects walk.

Human Activity Recognition (HAR) is a challenging task in the field of human-related signal processing. Owing to the development of wearable sensing technology, an emerging research approach in HAR is to identify user-performed tasks by using data collected from wearable sensors. In this paper, we propose a novel system for monitoring and recognizing daily living activities using an off-the-shelf smart band and two smart shoes. The system aims at providing a useful tool for solving problems regarding body part placement, fusion of multimodal sensors and feature selection for a specific set of activities. The system collects inertial and plantar pressure data at wrist and foot to analyze and then, extract, select important features for recognition. We construct and compare two predictive models of classifying activities from the reduced feature set. A comparison of the classification for each wearable device and a fusion scheme is provided to identify the best body part for activity recognition: either the wrist or the feet. This comparison also demonstrated the effective HAR performance of the proposed system.

We propose a data refinement and channel selection method for vapor classification in a portable e-nose system. For the robust e-nose system in a real environment, we propose to reduce the noise in the data measured by sensor arrays and distinguish the important part in the data by the use of feature feedback. Experimental results on different volatile organic compounds data show that the proposed data refinement method gives good clustering for different classes and improves the classification performance. Also, we design a new sensor array that consists only of the useful channels. For this purpose, each channel is evaluated by measuring its discriminative power based on the feature mask used in the data refinement. Through the experimental results, we show that the new sensor array improves both the classification rates and the efficiency in computation and data storage.

In this letter, we propose a novel method for classifying daily wrist activities by using a smart band. Triaxial acceleration data are collected by built-in sensors of the smart band during experiments regarding five activities, i.e., texting, calling, placing a hand in a pocket, carrying a suitcase, and swinging a hand. We analyze patterns in the sensor signals during these activities based on three types of features, i.e., norm, norm-variance, and frequency-domain features. After extracting the significant features, a multi-class support vector machine (SVM) algorithm is applied to classify these activities. We obtained recognition error rates of approximately 2.7% by applying the proposed method to the experimental dataset.