Context-aware sensing of physiological signals

Mobile Applications in Evaluations of Knee Joint Kinematics: A Pilot Study

Because medical professionals lack the means to monitor exercises performed by patients in their home environment directly, there is a strong case for introducing technological solutions into this domain. They include methods that use wireless inertial sensors, which emit signals recorded and processed by special applications that work with mobile devices. This paper's aim is (a) to evaluate whether such sensors are suitable for qualitative and quantitative motion analysis, and (b) to determine the repeatability of results over a few recordings. Knee joint activity was analysed using a system of inertial sensors connected through a Wi-Fi network to mobile devices. The tested individuals did eight different activities, all of which engaged the knee joint. Each excercise was repeated three times. Study results did not reveal any statistically significant differences between the three measurements for most of the studied parameters. Furthermore, in almost every case, there were no statistically significant differences between the results of the right and left lower limb (p > 0.05). This study shows that easy use and repeatability of results combined with the feature of quantitative and qualitative analysis make the examined method useful for functional evaluations of the knee joint.

Context Relevant Prediction Model for COPD Domain Using Bayesian Belief Network

In the last three decades, researchers have examined extensively how context-aware systems can assist people, specifically those suffering from incurable diseases, to help them cope with their medical illness. Over the years, a huge number of studies on Chronic Obstructive Pulmonary Disease (COPD) have been published. However, how to derive relevant attributes and early detection of COPD exacerbations remains a challenge. In this research work, we will use an efficient algorithm to select relevant attributes where there is no proper approach in this domain. Such algorithm predicts exacerbations with high accuracy by adding discretization process, and organizes the pertinent attributes in priority order based on their impact to facilitate the emergency medical treatment. In this paper, we propose an extension of our existing Helper Context-Aware Engine System (HCES) for COPD. This project uses Bayesian network algorithm to depict the dependency between the COPD symptoms (attributes) in order to overcome the insufficiency and the independency hypothesis of naïve Bayesian. In addition, the dependency in Bayesian network is realized using TAN algorithm rather than consulting pneumologists. All these combined algorithms (discretization, selection, dependency, and the ordering of the relevant attributes) constitute an effective prediction model, comparing to effective ones. Moreover, an investigation and comparison of different scenarios of these algorithms are also done to verify which sequence of steps of prediction model gives more accurate results. Finally, we designed and validated a computer-aided support application to integrate different steps of this model. The findings of our system HCES has shown promising results using Area Under Receiver Operating Characteristic (AUC = 81.5%).

The promise of mHealth: daily activity monitoring and outcome assessments by wearable sensors

Mobile health tools that enable clinicians and researchers to monitor the type, quantity, and quality of everyday activities of patients and trial participants have long been needed to improve daily care, design more clinically meaningful randomized trials of interventions, and establish cost-effective, evidence-based practices. Inexpensive, unobtrusive wireless sensors, including accelerometers, gyroscopes, and pressure-sensitive textiles, combined with Internet-based communications and machine-learning algorithms trained to recognize upper- and lower-extremity movements, have begun to fulfill this need. Continuous data from ankle triaxial accelerometers, for example, can be transmitted from the home and community via WiFi or a smartphone to a remote data analysis server. Reports can include the walking speed and duration of every bout of ambulation, spatiotemporal symmetries between the legs, and the type, duration, and energy used during exercise. For daily care, this readily accessible flow of real-world information allows clinicians to monitor the amount and quality of exercise for risk factor management and compliance in the practice of skills. Feedback may motivate better self-management as well as serve home-based rehabilitation efforts. Monitoring patients with chronic diseases and after hospitalization or the start of new medications for a decline in daily activity may help detect medical complications before rehospitalization becomes necessary. For clinical trials, repeated laboratory-quality assessments of key activities in the community, rather than by clinic testing, self-report, and ordinal scales, may reduce the cost and burden of travel, improve recruitment and retention, and capture more reliable, valid, and responsive ratio-scaled outcome measures that are not mere surrogates for changes in daily impairment, disability, and functioning.

Energy-efficient context classification with dynamic sensor control

Energy efficiency has been a longstanding design challenge for wearable sensor systems. It is especially crucial in continuous subject state monitoring due to the ongoing need for compact sizes and better sensors. This paper presents an energy-efficient classification algorithm, based on partially observable Markov decision process (POMDP). In every time step, POMDP dynamically selects sensors for classification via a sensor selection policy. The sensor selection problem is formalized as an optimization problem, where the objective is to minimize misclassification cost given some energy budget. State transitions are modeled as a hidden Markov model (HMM), and the corresponding sensor selection policy is represented using a finite-state controller (FSC). To evaluate this framework, sensor data were collected from multiple subjects in their free-living conditions. Relative accuracies and energy reductions from the proposed method are compared against naïve Bayes (always-on) and simple random strategies to validate the relative performance of the algorithm. When the objective is to maintain the same classification accuracy, significant energy reduction is achieved.

Reliability and validity of bilateral ankle accelerometer algorithms for activity recognition and walking speed after stroke

BACKGROUND AND PURPOSE

Outcome measures of mobility for large stroke trials are limited to timed walks for short distances in a laboratory, step counters and ordinal scales of disability and quality of life. Continuous monitoring and outcome measurements of the type and quantity of activity in the community would provide direct data about daily performance, including compliance with exercise and skills practice during routine care and clinical trials.

METHODS

Twelve adults with impaired ambulation from hemiparetic stroke and 6 healthy controls wore triaxial accelerometers on their ankles. Walking speed for repeated outdoor walks was determined by machine-learning algorithms and compared to a stopwatch calculation of speed for distances not known to the algorithm. The reliability of recognizing walking, exercise, and cycling by the algorithms was compared to activity logs.

RESULTS

A high correlation was found between stopwatch-measured outdoor walking speed and algorithm-calculated speed (Pearson coefficient, 0.98; P=0.001) and for repeated measures of algorithm-derived walking speed (P=0.01). Bouts of walking >5 steps, variations in walking speed, cycling, stair climbing, and leg exercises were correctly identified during a day in the community. Compared to healthy subjects, those with stroke were, as expected, more sedentary and slower, and their gait revealed high paretic-to-unaffected leg swing ratios.

CONCLUSIONS

Test-retest reliability and concurrent and construct validity are high for activity pattern-recognition Bayesian algorithms developed from inertial sensors. This ratio scale data can provide real-world monitoring and outcome measurements of lower extremity activities and walking speed for stroke and rehabilitation studies.

Episodic sampling: towards energy-efficient patient monitoring with wearable sensors

Energy efficiency presents a critical design challenge in wireless, wearable sensor technology, mainly because of the associated diagnostic objectives required in each monitoring application. In order to maximize the operating lifetime during real-life monitoring and maintain sufficient classification accuracy, the wearable sensors require hardware support that allows dynamic power control on the sensors and wireless interfaces as well as monitoring algorithms to control these components intelligently. This paper introduces a context-aware sensing technique known as episodic sampling - a method of performing context classification only at specific time instances. Based on Additive-Increase/Multiplicative-Decrease (AIMD), episodic sampling demonstrates an energy reduction of 85 percent with a loss of only 5 percent in classification accuracy in our experiment.