This paper proposes a deep convolutional neural network for HAR using smartphone sensors.Experiments show that the proposed method derives relevant and more complex features.The method achieved an almost perfect classification on moving activities.It outperforms other state-of-the-art data mining techniques in HAR. Human activities are inherently translation invariant and hierarchical. Human activity recognition (HAR), a field that has garnered a lot of attention in recent years due to its high demand in various application domains, makes use of time-series sensor data to infer activities. In this paper, a deep convolutional neural network (convnet) is proposed to perform efficient and effective HAR using smartphone sensors by exploiting the inherent characteristics of activities and 1D time-series signals, at the same time providing a way to automatically and data-adaptively extract robust features from raw data. Experiments show that convnets indeed derive relevant and more complex features with every additional layer, although difference of feature complexity level decreases with every additional layer. A wider time span of temporal local correlation can be exploited (1?9-1?14) and a low pooling size (1?2-1?3) is shown to be beneficial. Convnets also achieved an almost perfect classification on moving activities, especially very similar ones which were previously perceived to be very difficult to classify. Lastly, convnets outperform other state-of-the-art data mining techniques in HAR for the benchmark dataset collected from 30 volunteer subjects, achieving an overall performance of 94.79% on the test set with raw sensor data, and 95.75% with additional information of temporal fast Fourier transform of the HAR data set.

Human activity recognition with smartphone sensors using deep learning neural networks

https://upcommons.upc.edu/bitstream/handle/2117/79600/online_har.pdf;sequence=1

Transition-Aware Human Activity Recognition Using Smartphones

Signal segmentation is a crucial stage in the activity recognition process; however, this has been rarely and vaguely characterized so far. Windowing approaches are normally used for segmentation, but no clear consensus exists on which window size should be preferably employed. In fact, most designs normally rely on figures used in previous works, but with no strict studies that support them. Intuitively, decreasing the window size allows for a faster activity detection, as well as reduced resources and energy needs. On the contrary, large data windows are normally considered for the recognition of complex activities. In this work, we present an extensive study to fairly characterize the windowing procedure, to determine its impact within the activity recognition process and to help clarify some of the habitual assumptions made during the recognition system design. To that end, some of the most widely used activity recognition procedures are evaluated for a wide range of window sizes and activities. From the evaluation, the interval 1-2 s proves to provide the best trade-off between recognition speed and accuracy. The study, specifically intended for on-body activity recognition systems, further provides designers with a set of guidelines devised to facilitate the system definition and configuration according to the particular application requirements and target activities.

/pdf/window-size-impact-in-human-activity-recognition-1ne72d19c7.pdf

Window Size Impact in Human Activity Recognition

A robust human activity recognition system using smartphone sensors and deep learning

https://www.gc.cuny.edu/CUNY_GC/media/Computer-Science/Student%20Presentations/Xing%20Su/Second_Exam_Slides_Xing_Su_6-12-2014.pdf

Activity recognition with smartphone sensors

As smartphone users have been increased, studies using mobile sensors on smartphone have been investigated in recent years. Activity recognition is one of the active research topics, which can be used for providing users the adaptive services with mobile devices. In this paper, an activity recognition system on a smartphone is proposed where the uncertain time-series acceleration signal is analyzed by using hierarchical hidden Markov models. In order to address the limitations on the memory storage and computational power of the mobile devices, the recognition models are designed hierarchy as actions and activities. We implemented the real-time activity recognition application on a smartphone with the Google android platform, and conducted experiments as well. Experimental results showed the feasibility of the proposed method.

Activity recognition using hierarchical hidden markov models on a smartphone with 3D accelerometer

http://sclab.yonsei.ac.kr/publications/Papers/IJ/2014_Activity%20recognition%20with%20android.pdf

Activity recognition with android phone using mixture-of-experts co-trained with labeled and unlabeled data

In most cases, the story or plot of popular role-playing games is constructed by professional designers as a main content. However, manual design of game content has a limitation in the quantitative aspect; it requires a large amount of time and effort. As game consumers want more diverse and rich contents, it is not easy to satisfy these needs with manual design, so procedural content generation is actively exploited to automatically generate game contents. In this paper, we propose a quest generation method using Petri net modules. A quest depending on the player's involvement or type determined by Bayesian network is generated by Petri net. Never Winter Night is used as a game platform to show the feasibility of the proposed method. In future works, we will collect players' playing history and evaluate the performance of Bayesian network inference for a player's type. Also, we will apply the proposed method to an open-source platform for a complete automatic quest generation system.

Context-Aware Petri Net for Dynamic Procedural Content Generation in Role-Playing Game

Life logs include people's experiences collected from various sources. It is used to support user's memory. There are many studies that collect and store life log for personal memory. In this paper, we collect log data from smart phone, derive contexts from the log, and then identify which is meaningful context by using a method based on KeyGraph. To evaluate the proposed method, we show an example of the meaningful places by using contexts and GPS logs collected from two users.

http://sclab.yonsei.ac.kr/publications/Papers/IC/IDEAL07-YSLee.pdf

Extracting meaningful contexts from mobile life log

Recently, activity recognition using built-in sensors in a mobile phone becomes an important issue. It can help us to provide context-aware services: health care, suitable content recommendation for a user's activity, and user adaptive interface. This paper proposes a layered hidden Markov model (HMM) to recognize both short-term activity and long-term activity in real time. The first layer of HMMs detects short, primitive activities with acceleration, magnetic field, and orientation data, while the second layer exploits the inference of the previous layer and other sensor values to recognize goal-oriented activities of longer time period. Experimental results demonstrate the superior performance of the proposed method over the alternatives in classifying long-term activities as well as short-term activities. The performance improvement is up to 10 % in the experiments, depending on the models compared.

Young-Seol Lee

Papers

Activity recognition using hierarchical hidden markov models on a smartphone with 3D accelerometer

Activity recognition with android phone using mixture-of-experts co-trained with labeled and unlabeled data

Context-Aware Petri Net for Dynamic Procedural Content Generation in Role-Playing Game

Extracting meaningful contexts from mobile life log

Layered hidden Markov models to recognize activity with built-in sensors on Android smartphone