لخّصلي

The task of human activity recognition using smartphone's built-in accelerometer has been well addressed in literature.In partic-ular, the existing works considered time segments of size 128 [7], 200 [2], 250 [8], 300 [9] and 512 [10], which corresponds to interval duration of 2.56-10 s. Smaller time intervals were used in [11], and while this work shows quite good performance, a very small private dataset obtained from 4 users and a limited range of activities makes its results incomparable to any existing solution.A different approach to feature extraction task is based on deep learning/CNNs, and several works have been conducted to adapt it to HAR problem.Though a number of papers proposed online HAR systems, they used recognition intervals that are generally quite long for online classification.When it comes to practical applications, one challenge that arises here is real-time classification of user activity.The first difference between the proposed solu-tions is how the input signals are treated.Furthermore, all mentioned systems were based on hand-designed features.

The task of human activity recognition using smartphone’s built-in accelerometer has been well addressed in literature. When it comes to practical applications, one challenge that arises here is real-time classification of user activity. Though a number of papers proposed online HAR systems, they used recognition intervals that are generally quite long for online classification. In partic-ular, the existing works considered time segments of size 128 [7], 200 [2], 250 [8], 300 [9] and 512 [10], which corresponds to interval duration of 2.56–10 s. Smaller time intervals were used in [11], and while this work shows quite good performance, a very small private dataset obtained from 4 users and a limited range of activities makes its results incomparable to any existing solution. Furthermore, all mentioned systems were based on hand-designed features.
A different approach to feature extraction task is based on deep learning/CNNs, and several works have been conducted to adapt it to HAR problem. The first difference between the proposed solu-tions is how the input signals are treated. In [12–14] the authors were focused on using multiple sensors and proposed to stack sig-nals from them row-by-row into one “sensor image” that is further passed to a Convolutional Neural Network. In [15], instead of deal-ing with a raw sensor image, a Discrete Fourier Transform was applied to this image and the obtained features were used for the classification. In [4,16,17], where human activity recognition was performed using accelerometer data from one device, the authors learned feature maps for x-, y- and z-accelerometer channels sep-arately that is similar to how an RGB image is typically processed by CNN. The architecture of CNNs also varied among the studies. In [4] one convolutional and two fully-connected layers were consid-ered, in [12,15,14] – two and one respectively. In [16,17,13] the authors have proposed even deeper architectures that consisted of three and four convolutional layers. Though deeper architec-tures are theoretically able to learn more abstract features, they often lead to data overfitting and therefore an appropriate balance should be maintained here. In this work we will show that appropri-ately tuned shallow CNN can yield an accurate classification while requiring less computational resources. Nowadays smartphones became an integral part of our daily lives and go with us everywhere, becoming a perfect tool for the analysis of human daily activities. For this reason we have chosen open WISDM [25] and UCI [26] datasets for training and performance evaluation of our model. These datasets contain accelerometer data from Android cell phones that was collected while users were performing a set of different activities, such as walking, jogging, stair climbing, sitting, lying and standing. Another advantage of these datasets is that they were already used in several research works. For WISDM dataset all previous works developed user-specific solutions, and only [5] considered a user-independent model and proposed using a combination of hand-crafted features and Random Forest or Dropout classifiers on top of them. UCI dataset has a version that is already split into training and test sets that contain data from different participants, therefore user-independent solution was dominant in this case. For UCI dataset, manual feature engineering was the prevailing approach [5,19–23,7], though several deep learning methods were also proposed [15–17,23,24]. In [23] Deep Boltzmann Machines were adapted for unsupervised feature extraction, and though they are not targeted on capturing local data structure their performance was superior to the other hand-crafted solutions. In [16] the authors used deep CNNs with three convolutional layers, but according to the experiments this caused significant data overfitting. A better performance was obtained with two-layered CNNs at the expense of using FFT features instead [15] or in addition [17] to the raw time series data. Another promising solution with low computa-tional cost [24] is based on Recurrent Neural Networks, though it is difficult to compare its accuracy to the previous solutions since a custom split of the dataset into training and test parts was used. The best results for UCI dataset were achieved using 561 hand-designed features proposed in [7] and various classifiers on top of them. Fur-ther experimental results obtained on WISDM and UCI datasets are presented in Table 1.