Sensor-based Activity Research Articles

"Are You Sure?": Lapses in Self-Reported Activities Among Healthy Older Adults Reporting Online.

Accurate retrospective reporting of activities and symptoms has been shown to be problematic for older adults, yet standard clinical care relies on self-reports to aid in assessment and management. Our aim was to examine the relationship between self-report and sensor-based measures of activity. We administered an online activity survey to participants in our ongoing longitudinal study of in-home ubiquitous monitoring. We found a wide range of accuracies when comparing self-report with time-stamped sensor-based data. Of the 95 participants who completed the 2-hr activity log, nearly one quarter did not complete the task in a way that could potentially be compared with sensor data. Where comparisons were possible, agreement between self-reported and sensor-based activity was achieved by a minority of participants. The findings suggest that capture of real-time events with unobtrusive activity monitoring may be a more reliable approach to describing behavioral patterns and meaningful changes in older adults.

Sensor-based activity recognition using extended belief rule-based inference methodology.

The recently developed extended belief rule-based inference methodology (RIMER+) recognizes the need of modeling different types of information and uncertainty that usually coexist in real environments. A home setting with sensors located in different rooms and on different appliances can be considered as a particularly relevant example of such an environment, which brings a range of challenges for sensor-based activity recognition. Although RIMER+ has been designed as a generic decision model that could be applied in a wide range of situations, this paper discusses how this methodology can be adapted to recognize human activities using binary sensors within smart environments. The evaluation of RIMER+ against other state-of-the-art classifiers in terms of accuracy, efficiency and applicability was found to be significantly relevant, specially in situations of input data incompleteness, and it demonstrates the potential of this methodology and underpins the basis to develop further research on the topic.

Activity Graph Feature Selection for Activity Pattern Classification

Sensor-based activity recognition is attracting growing attention in many applications. Several studies have been performed to analyze activity patterns from an activity database gathered by activity recognition. Activity pattern classification is a technique that predicts class labels of people such as individual identification, nationalities, and jobs. For this classification problem, it is important to mine discriminative features reflecting the intrinsic patterns of each individual. In this paper, we propose a framework that can classify activity patterns effectively. We extensively analyze activity models from a classification viewpoint. Based on the analysis, we represent activities as activity graphs by combining every combination of daily activity sequences in meaningful periods. Frequent patterns over these activity graphs can be used as discriminative features, since they reflect people's intrinsic lifestyles. Experiments show that the proposed method achieves high classification accuracy compared with existing graph classification techniques.

Assessing physical activity in inpatient rehabilitation—sensor-based validation of the PAIR

Increased physical activity is positively associated with better health in community-dwelling older persons. It is unclear whether physical activity also influences success of inpatient rehabilitation. For the assessment of physical activity in inpatient rehabilitation the Physical Activity in Inpatient Rehabilitation Assessment (PAIR), a short questionnaire based on five questions, was developed and preliminary validated. In this study, the PAIR was validated against a sensor-based physical activity measurement. Seventy functionally impaired and cognitively mostly intact patients of a German geriatric inpatient rehabilitation clinic who had undergone hip surgery (n = 62 women, median age = 83 years) participated. Physical activity was measured using the PAIR and a sensor-based activity monitor (Physilog ®; BioAGM, CH). Assessments were conducted at admission (T1) and 2 weeks later (T2) during the rehabilitation process. To assess concurrent and predictive validity, Spearman correlations and linear regression models were calculated using sensor-based walking activity and uptime activity (walking and standing time) as dependent variables. Criterion-related concurrent validity using physical activity sensors was weak to moderate. Correlations were slightly higher at T2 (r = 0.45-0.53) than at T1 (r = 0.44-0.46). The objectively measured variance of physical activity, explained by the PAIR, ranged from 25 to 43 %. PAIR activity scores and sensor-based walking or total activity increased in a dose-dependent manner, confirming the scoring system of the PAIR. The application time was usually less than 2 min. The validity of the PAIR is weak to moderate when compared to a sensor-based activity monitor and comparable to existing physical activity assessments for community-dwelling older adults.

Sensor-based surgical activity recognition in unconstrained environments

Introduction: Automatic surgical activity recognition in the operating room (OR) is mandatory to enable assistive surgical systems to manage the information presented to the surgical team. Therefore the purpose of our study was to develop and evaluate an activity recognition model. Material and methods: The system was conceived as a hierarchical recognition model which separated the recognition task into activity aspects. The concept used radio frequency identification (RFID) for instrument recognition and accelerometers to infer the performed surgical action. Activity recognition was done by combining intermediate results of the aspect recognition. A basic scheme of signal feature generation, clustering and sequence learning was replicated in all recognition subsystems. Hidden Markov models (HMM) were used to generate probability distributions over aspects and activities. Simulated functional endoscopic sinus surgeries (FESS) were used to evaluate the system. Results and discussion: The system was able to detect surgical activities with an accuracy of 95%. Instrument recognition performed best with 99% accuracy. Action recognition showed lower accuracies with 81% due to the high variability of surgical motions. All stages of the recognition scheme were evaluated. The model allows distinguishing several surgical activities in an unconstrained surgical environment. Future improvements could push activity recognition even further.

Recommendations for Standardizing Validation Procedures Assessing Physical Activity of Older Persons by Monitoring Body Postures and Movements

Physical activity is an important determinant of health and well-being in older persons and contributes to their social participation and quality of life. Hence, assessment tools are needed to study this physical activity in free-living conditions. Wearable motion sensing technology is used to assess physical activity. However, there is a lack of harmonisation of validation protocols and applied statistics, which make it hard to compare available and future studies. Therefore, the aim of this paper is to formulate recommendations for assessing the validity of sensor-based activity monitoring in older persons with focus on the measurement of body postures and movements. Validation studies of body-worn devices providing parameters on body postures and movements were identified and summarized and an extensive inter-active process between authors resulted in recommendations about: information on the assessed persons, the technical system, and the analysis of relevant parameters of physical activity, based on a standardized and semi-structured protocol. The recommended protocols can be regarded as a first attempt to standardize validity studies in the area of monitoring physical activity.

Sensor-based home activity monitoring of geriatric fracture patients in rehabilitation – The GAL NATARS study

O 013 Sensor-based home activity monitoring of geriatric fracture patients in rehabilitation –The GAL NATARS study M. Marscholleka,b,c,d,e,f,g,h,i, M. Apela,b,c,d,e,f,g,h,i, P. Bentea,b,c,d,e,f,g,h,i, L. Dasenbrocka,b,c,d,e,f,g,h,i, K. Elbersa,b,c,d,e,f,g,h,i, R. Hauxa,b,c,d,e,f,g,h,i, A. Heina,b,c,d,e,f,g,h,i, G. Kolba,b,c,d,e,f,g,h,i, H. Kunemunda,b,c,d,e,f,g,h,i, C. Lammel-Polchaua,b,c,d,e,f,g,h,i, M. Meisa,b,c,d,e,f,g,h,i, H. Meyer zu Schwabedissena,b,c,d,e,f,g,h,i, H. Remmersa,b,c,d,e,f,g,h,i, E.E. Steena,b,c,d,e,f,g,h,i, J. Wanga,b,c,d,e,f,g,h,i, K.H. Wolf a,b,c,d,e,f,g,h,i, J. Bauera,b,c,d,e,f,g,h,i a Hannover Medical School, Hannover, Germany b Klinikum Oldenburg, Oldenburg. Germany c Std. Klinikum Braunschweig, Braunschweig, Germany d St. Bonifatius Hospital Lingen, Lingen, Germany e TU Braunschweig, Braunschweig, Germany f OFFIS, Oldenburg, Germany g Universitat Vechta, Vechta, Germany h Horzentrum Oldenburg, Oldenburg, Germany i Universitat Osnabruck, Osnabruck, Germany

Is my configuration any good: checking usability in an interactive sensor-based activity monitor

We investigate formal analysis of two aspects of usability in a deployed interactive, configurable and context-aware system: an event-driven, sensor-based homecare activity monitor system. The system was not designed from formal requirements or specification: we model the system as it is in the context of an agile development process. Our aim was to determine if formal modelling and analysis can contribute to improving usability, and if so, which style of modelling is most suitable. The purpose of the analysis is to inform configurers about how to interact with the system, so the system is more usable for participants, and to guide future developments. We consider redundancies in configuration rules defined by carers and participants and the interaction modality of the output messages.Two approaches to modelling are considered: a deep embedding in which devices, sensors and rules are represented explicitly by data structures in the modelling language and non-determinism is employed to model all possible device and sensor states, and a shallow embedding in which the rules and device and sensor states are represented directly in propositional logic. The former requires a conventional machine and a model-checker for analysis, whereas the latter is implemented using a SAT solver directly on the activity monitor hardware. We draw conclusions about the role of formal models and reasoning in deployed systems and the need for clear semantics and ontologies for interaction modalities.

The Opportunity challenge: A benchmark database for on-body sensor-based activity recognition

There is a growing interest on using ambient and wearable sensors for human activity recognition, fostered by several application domains and wider availability of sensing technologies. This has triggered increasing attention on the development of robust machine learning techniques that exploits multimodal sensor setups. However, unlike other applications, there are no established benchmarking problems for this field. As a matter of fact, methods are usually tested on custom datasets acquired in very specific experimental setups. Furthermore, data is seldom shared between different groups. Our goal is to address this issue by introducing a versatile human activity dataset recorded in a sensor-rich environment. This database was the basis of an open challenge on activity recognition. We report here the outcome of this challenge, as well as baseline performance using different classification techniques. We expect this benchmarking database will motivate other researchers to replicate and outperform the presented results, thus contributing to further advances in the state-of-the-art of activity recognition methods.

Sensor-Based Activity Recognition

Research on sensor-based activity recognition has, recently, made significant progress and is attracting growing attention in a number of disciplines and application domains. However, there is a lack of high-level overview on this topic that can inform related communities of the research state of the art. In this paper, we present a comprehensive survey to examine the development and current status of various aspects of sensor-based activity recognition. We first discuss the general rationale and distinctions of vision-based and sensor-based activity recognition. Then, we review the major approaches and methods associated with sensor-based activity monitoring, modeling, and recognition from which strengths and weaknesses of those approaches are highlighted. We make a primary distinction in this paper between data-driven and knowledge-driven approaches, and use this distinction to structure our survey. We also discuss some promising directions for future research.

Active tool for human activity data collection

Sensor-based human activity data (HAD) collection, used for engineering human activity recognition (HAR) systems, is a cumbersome process. Its success depends on two factors: correct activity labels and correct sensor activation sequence. This is very hard to achieve, as is seen in many previous studies. Therefore, a system for automatically observing HAD collection is introduced, making it a lot easier and less noisy. At the heart of the system sits a novel data mining tool that uses Web activity data to achieve the above goal. The effectiveness of the proposed system is shown using a set of two experiments.

Review of Sensor-based Activity Recognition Systems

Activity recognition (AR) has become a hot research topic due to its strength in providing personalized sup port for many diverse applications such as healthcare and security. Due to its importance, a considerable amount of AR systems have been developed. In general, these systems utilize diverse sensors to obtain the activity related information, which are then used by machine learning techniques to infer human’s ongoing activity. According to the types of sensors used, existing AR systems can be roughly divided into two catego ries: 1. Video sensor based AR. It remotely observes human activity using video sensors; 2. Physical sensor based AR (PSAR). It attaches physical sensors to the body of human or objects (appliances) to infer human activity. Based on the location of attached sensors, PSAR consists of two subcategories: Wearable sensor based AR and object usage based AR. In this work, different types of AR are reviewed. We think this review is significant because no existing review papers include all the different types of AR as a whole. For each type of AR, its main techniques, characteristics, strengths and limitations are discussed and summarized. We also make a comparative analysis for them. Finally the main research challenges in AR are pointed out.

Unobtrusive Measurement of Indoor Energy Expenditure Using an Infrared Sensor-Based Activity Monitoring System

The purpose of this study was to develop an unobtrusive energy expenditure (EE) measurement system using an infrared (IR) sensor-based activity monitoring system to measure indoor activities and to estimate individual quantitative EE. IR-sensor activation counts were measured with a Bluetooth-based monitoring system and the standard EE was calculated using an established regression equation. Ten male subjects participated in the experiment and three different EE measurement systems (gas analyzer, accelerometer, IR sensor) were used simultaneously in order to determine the regression equation and evaluate the performance. As a standard measurement, oxygen consumption was simultaneously measured by a portable metabolic system (Metamax 3X, Cortex, Germany). A single room experiment was performed to develop a regression model of the standard EE measurement from the proposed IR sensor-based measurement system. In addition, correlation and regression analyses were done to compare the performance of the IR system with that of the Actigraph system. We determined that our proposed IR-based EE measurement system shows a similar correlation to the Actigraph system with the standard measurement system.

Plan-based complex event detection across distributed sources

Complex Event Detection (CED) is emerging as a key capability for many monitoring applications such as intrusion detection, sensor-based activity & phenomena tracking, and network monitoring. Existing CED solutions commonly assume centralized availability and processing of all relevant events, and thus incur significant overhead in distributed settings. In this paper, we present and evaluate communication efficient techniques that can efficiently perform CED across distributed event sources. Our techniques are plan-based: we generate multi-step event acquisition and processing plans that leverage temporal relationships among events and event occurrence statistics to minimize event transmission costs, while meeting application-specific latency expectations. We present an optimal but exponential-time dynamic programming algorithm and two polynomial-time heuristic algorithms, as well as their extensions for detecting multiple complex events with common sub-expressions. We characterize the behavior and performance of our solutions via extensive experimentation on synthetic and real-world data sets using our prototype implementation.

An object-oriented environment for robot system architectures

The object-oriented Robot Independent Programming Environment (RIPE), which is being used for the rapid design and implementation of a variety of applications, is described. A system architecture based on hierarchies of distributed multiprocessors provides the computing platform for a layered programming structure that models the application as a set of software objects. These objects are designed to support model-based automated planning and programming, real-time sensor-based activity, error handling, and robust communication. The object-oriented paradigm provides mechanisms such as inheritance and polymorphism which allow the implementation of the system to satisfy the goals of software reusability, extensibility, reliability, and portability. Designing a hierarchy of generic parent classes and device-specific subclasses which inherit the same interface allows a Robot Independent Programming Language (RIPL) to be realized. Workcell tasks demonstrating robot cask handling operations for nuclear waste facilities, which are successfully implemented using this object-oriented software environment, are discussed. >