In this work, algorithms are developed and evaluated to de- tect physical activities from data acquired using five small biaxial ac- celerometers worn simultaneously on different parts of the body. Ac- celeration data was collected from 20 subjects without researcher su- pervision or observation. Subjects were asked to perform a sequence of everyday tasks but not told specifically where or how to do them. Mean, energy, frequency-domain entropy, and correlation of acceleration data was calculated and several classifiers using these features were tested. De- cision tree classifiers showed the best performance recognizing everyday activities with an overall accuracy rate of 84%. The results show that although some activities are recognized well with subject-independent training data, others appear to require subject-specific training data. The results suggest that multiple accelerometers aid in recognition because conjunctions in acceleration feature values can effectively discriminate many activities. With just two biaxial accelerometers - thigh and wrist - the recognition performance dropped only slightly. This is the first work to investigate performance of recognition algorithms with multiple, wire-free accelerometers on 20 activities using datasets annotated by the subjects themselves.

/pdf/activity-recognition-from-user-annotated-acceleration-data-lh3b5ajkau.pdf

Activity recognition from user-annotated acceleration data

Mobile devices are becoming increasingly sophisticated and the latest generation of smart cell phones now incorporates many diverse and powerful sensors These sensors include GPS sensors, vision sensors (ie, cameras), audio sensors (ie, microphones), light sensors, temperature sensors, direction sensors (ie, magnetic compasses), and acceleration sensors (ie, accelerometers) The availability of these sensors in mass-marketed communication devices creates exciting new opportunities for data mining and data mining applications In this paper we describe and evaluate a system that uses phone-based accelerometers to perform activity recognition, a task which involves identifying the physical activity a user is performing To implement our system we collected labeled accelerometer data from twenty-nine users as they performed daily activities such as walking, jogging, climbing stairs, sitting, and standing, and then aggregated this time series data into examples that summarize the user activity over 10- second intervals We then used the resulting training data to induce a predictive model for activity recognition This work is significant because the activity recognition model permits us to gain useful knowledge about the habits of millions of users passively---just by having them carry cell phones in their pockets Our work has a wide range of applications, including automatic customization of the mobile device's behavior based upon a user's activity (eg, sending calls directly to voicemail if a user is jogging) and generating a daily/weekly activity profile to determine if a user (perhaps an obese child) is performing a healthy amount of exercise

/pdf/activity-recognition-using-cell-phone-accelerometers-3k6h9fhibm.pdf

Activity recognition using cell phone accelerometers

http://gruberpeplab.com/3131/Kreibig_2010.pdf

Autonomic nervous system activity in emotion: A review

Providing accurate and opportune information on people's activities and behaviors is one of the most important tasks in pervasive computing. Innumerable applications can be visualized, for instance, in medical, security, entertainment, and tactical scenarios. Despite human activity recognition (HAR) being an active field for more than a decade, there are still key aspects that, if addressed, would constitute a significant turn in the way people interact with mobile devices. This paper surveys the state of the art in HAR based on wearable sensors. A general architecture is first presented along with a description of the main components of any HAR system. We also propose a two-level taxonomy in accordance to the learning approach (either supervised or semi-supervised) and the response time (either offline or online). Then, the principal issues and challenges are discussed, as well as the main solutions to each one of them. Twenty eight systems are qualitatively evaluated in terms of recognition performance, energy consumption, obtrusiveness, and flexibility, among others. Finally, we present some open problems and ideas that, due to their high relevance, should be addressed in future research.

A Survey on Human Activity Recognition using Wearable Sensors

Activity recognition fits within the bigger framework of context awareness. In this paper, we report on our efforts to recognize user activity from accelerometer data. Activity recognition is formulated as a classification problem. Performance of base-level classifiers and meta-level classifiers is compared. Plurality Voting is found to perform consistently well across different settings.

/pdf/activity-recognition-from-accelerometer-data-3lvh3isrg0.pdf

Activity recognition from accelerometer data

Detection of posture and motion by accelerometry : a validation study in ambulatory monitoring

Basic motion patterns and posture can be distinguished by multichannel accelerometry, as recently shown. A refinement of this method appeared to be desirable to further increase its effectiveness, especially to distinguish walking and climbing stairs, and body rotation during sleep. Recordings were made of 31 subjects, according to a standard protocol comprising 13 motions and postures. This recording was repeated three times with appropriate permutation. Five uni-axial sensors and three sites of placement (sternum with three axes, right and left thigh) were selected. A hierarchical classification strategy used a standard protocol (i.e., individual reference patterns) to distinguish subtypes of moving behaviors and posture. The analysis method of the actometer signals reliably detected 13 different postural and activity conditions (only 3.2% misclassifications). A minimum set of sensors can be found for a given application; for example, a two-sensor configuration would clearly suffice to differentiate between four basic classes (sitting, standing, lying, moving) in ambulatory monitoring.

/pdf/motion-pattern-and-posture-correctly-assessed-by-calibrated-3izfph06bx.pdf

Motion pattern and posture: correctly assessed by calibrated accelerometers.

There are few studies on the development of oculomotor functions during childhood. B. Fischer, M. Biscaldi, and S. Gezeck (1997) reported improvement of antisaccade task performance between ages 6 and 16 years. The present study is a replication and extension of those results. In three age groups (6-7, 10-11, 18-26 years), saccades during pro- and antisaccade tasks with 200-ms gap and overlap and during a fixation task were measured. Adults exhibited faster saccades and less prosaccades during the antisaccade tasks than 10-11-year-old children; these two groups had faster saccades during all tasks and less prosaccades during the anti- and the fixation task than 6-7-year-old subjects. Both children groups made more express saccades than adults. Results suggest different degrees of age-related improvement for different saccadic parameters, the effects being greatest for prosaccade inhibition during the antisaccade task and in line with the assumed protracted development of prefrontal functions.

Development of prosaccade and antisaccade task performance in participants aged 6 to 26 years.

Modern assessment of posture and motion involves the use of wide bandwidth piezoresistive accelerometers. The direct current (DC) component allows for assessment of slow motion and change in position referring to the gravitational axis: the alternating current (AC) component, calibrated in g. represents acceleration along the sensitive axis of the device. A method study was designed to evaluate the division in DC and AC components, reliabilities, discrimination between conditions, and detection of types of physical activities. Recordings were made in 26 student participants for eight conditions: sitting, standing, lying supine, sitting and typing on a PC keyboard, walking, climbing stairs, walking downstairs, and cycling. This procedure was repeated in reversed order. A classification of physical activities according to the eight conditions (first trial) and based on four parameters, that is DC components trunk, thigh, and lower leg and AC component trunk, was correct in almost 100% of patterns, when applied to the second trial.

Friedrich Foerster

Papers

Detection of posture and motion by accelerometry : a validation study in ambulatory monitoring

Motion pattern and posture: correctly assessed by calibrated accelerometers.

Development of prosaccade and antisaccade task performance in participants aged 6 to 26 years.

Assessment of posture and motion by multichannel piezoresistive accelerometer recordings

The law of initial value: a rare exception.