Wearable Sensors and Systems Terms of Use Detailed Terms

Abstract
AI

Wearable technology is becoming increasingly significant in both research and clinical settings due to its potential for long-term individual monitoring in home and community contexts. This paper discusses the evolution of wearable sensors and systems, particularly their applications in monitoring patients with conditions like Parkinson's disease. The integration of wireless technology and e-textiles is highlighted as a key aspect of current developments, emphasizing the need for hybrid systems to realize the full potential of wearable technology in clinical health monitoring.

Figures (9)

Fig. 2. Schematic representation of a wearable system that allows one to collect movement and physiological data (i.e, heart rate and respiratory rate). Movement data are gathered using the four nodes equipped with accelerome- ters that are strapped around the ankles and wrists. Physio- logical data are collected via a chest strap.

Fig. 1. Schematic representation of a system for patients’ monitoring in the home and community settings. A subject is shown while exercising at the gym (e.g., undergoing balance therapy). Exercise compliance, exercise performance, and the associ- ated physiological responses (i.e., heart rate and respiratory rate) are monitored via wearable sensors. A cell phone serves as a data logger and gateway for communication with a remote location via a cell phone network and/or the Internet.

Fig. 3. Examples of e-textile technologies developed over the past ten years. (a) A system developed by Wade and Asada (22) relying upon special buttons that carry sensor technology to record physiological and movement data. (6) A system developed by De Rossi’s research team (23) for monitoring the movements of the shoulder and elbow via recordings of the voltage drop on conductive elastomers that are printed on the garment. (Figures used with permission.)

Fig. 4. An important application of wearable technology consists of monitoring individuals while they exercise. The focus here is on wellness/fitness monitoring, with the poten- tial development of methodologies to improve exercise compliance. Wearable sensors allow one to monitor move- ment, respiratory rate, and heart rate. ECG: electrocardio- graphic recordings; SeO2: oxygen saturation. Paralleling the progress in wearable technology, applica- tions gradually shifted their focus toward medical problems that require enhanced reliability compared with systems designed for wellness applications. Monitoring patients with Parkinson’s disease to improve clinical management of symp- toms is an example of one such type of application. Currently, clinical visits are inadequate to sample the severity of parkin- sonian symptoms, because symptoms vary in response to a medication dosage with a time constant of hours, a time inter- val that does not lend itself to direct patient observation by

Fig. 5. Home-monitoring applications would often benefit from technology for remote examination of patients. The screen cap- ture presented shows the graphical interface of an application recently developed by Matt Welsh’s research team at Harvard University to monitor patients with Parkinson’s disease in the home. The software application provides clinicians with access to wearable sensor data and measures the severity of parkinsonian symptoms (45). Wearable technology has the potential for addressing these problems by providing a means of gathering objective measures of the severity of symptoms over a period of time sufficient, for instance, to reliably assess the effectiveness of medication adjustments. Seminal work by Ghika et al. [3] and Spieker et al. [4] exploring the use of sensor technology to capture the severity of parkinsonian symptoms was followed by the work by Keijsers et al. [41]-[43] aimed at assessing the effectiveness of medications in attenuating the severity of symptoms using wearable sensors. More recent research has been focused on integrating and further developing these tech- niques into complete wearable systems for home monitoring of patients with Parkinson’s disease [44]-[46]. We anticipate that home monitoring of patients with Parkinson’s disease will be integrated in the near future with remote assessment tools leveraging videoconferencing and remote access to sensor data to facilitate clinical evaluation of the severity of parkinso- nian symptoms. Figure 5 shows a software application recently developed by Matt Welsh’s research team and my research team (supported by the Michael J. Fox Foundation) as part of a joint effort toward the development of Web-based applications devoted to the collection of data from patients with Parkinson’s

Fig. 6. Rehabilitation robotics is combined with wearable technology for the purpose of enhancing functions that are not pro- vided by robotic systems (51). Features provided by commercially available robotic systems like the one shown in this figure (Armeo by Hocoma AG) can be augmented via the use of wearable sensors. In the example presented, a sensorized glove provides the system with the ability to implement exercises targeting the recovery of hand function. This capability would not be available if the robot were to be used alone. The aforementioned approach is expected to benefit sub- jects undergoing physical therapy to recover arm and hand functions. The combination of wearable technology (i.e., the sensorized glove) and robotics allows one to improve the qual- ity of the intervention. The robot alone does not lend itself to the implementation of therapeutic exercises that focus on hand function, an aspect of physical therapy that is known to be of paramount importance when one aims at achieving recovery of the subject’s functional capability. The sensorized glove is therefore a key factor in improving the clinical intervention in the presented application scenario. Patients that are candidates for the use of these technologies include individuals who have suffered a stroke, a traumatic brain injury, or experienced other neurological problems leading to impairments and func- tional limitations of the upper limbs. It is important to note that traditional physical therapy tech- niques could theoretically lead to similar results to the ones expected from robotic therapy (although recent research sug- gests that robotic therapy leveraging interactive games leads to better results than therapeutic interventions simply based on delivering a high number of repetitions of specific move- ments [53]). However, the intensity of the exercise that has been shown to benefit patients when robotics is relied upon cannot be achieved in the current health-care system model by means of traditional interventions based on manual ther- apy administered by clinical personnel in a one-to-one ratio with patients (i.e., with one therapist working with a single patient at a time). This is because the number of physical ther- apy sessions that are reimbursed by insurance companies is imited. Also, it must be observed that a higher number of movement repetitions can be achieved within a single session using robotics compared with traditional therapeutic inter- ventions. In this context, wearable technology provides a means to enhance available rehabilitation robotic platforms

Fig. 7. It is envisioned that combining (a) wearable technol- ogy and (6) home robots will result in novel home-monitor- ing applications. Wearable sensors will Communicate wirelessly with home robots (image courtesy of iRobot) that will in turn resoond to alarm messages using onboard capa- bilities (j.e., image processing) to interact with the patient and assess the severity of the situation. Figure 8 would also provide connectivity with interactive gaming systems like the Nintendo Wii. Physical and occupa- tional therapists have demonstrated a growing interest in the use of off-the-shelf interactive gaming systems as a tool that complements traditional clinical interventions. The use of off- the-shelf interactive gaming systems is very attractive in the context of implementing home interventions, but commer- cially available systems lack the ability of monitoring move- ment patterns in a way that is satisfactory from a rehabilitation intervention standpoint. While using interactive gaming sys- tems, subjects must be encouraged to use appropriate motor control strategies (rather than compensatory mechanisms). In the scenario shown in Figure 8, the system would provide appropriate feedback during the performance of home exer- cises, based on the analysis of data recorded using wearable sensors. The sensors would be used to monitor movement pat- terns, and a home robot would be relied upon to convey feed- back to the individual. Finally, wearable sensors and home

Fig. 8. Complex systems under development will soon provide enhanced monitoring capability, the ability to facilitate clinical interventions, and features that are suitable for detecting emergency situations, assess needs (e.g., via gathering images and other information using a home robot), and alert a remote clinical center when necessary. It is worth noting that the combination of home robots and wearable technology is somewhat complementary to installing sensing components in living environments [61]. Although some applications might be better served by sensing compo- nents installed in the home [62], [63], the use of home robots has a significant potential for decreasing costs and mitigating the level of obtrusiveness of the monitoring system. It is known that robots are often perceived by people as pets. Therefore, one would expect that a home robot would be more easily accepted than a set of Web cameras positioned in all the rooms of the home. Home robots are also easier to control, and they provide the assurance that privacy is not violated. For instance, the camera positioned on the robot can be easily flipped so that the lens does not face the subject, thus reassur- ing individuals that the privacy is not violated even by robotics would be combined to achieve prompt detection of falls in the home environment. A key factor in minimizing the severity of fall-related injuries is to promptly detect the fall event and alert clinical personnel. During the past decade, a number of devices for fall detection have been developed by researchers [56]-[60], and fall-detection devices have been introduced on the market. These systems are typically based on body-worn units (e.g., pendants and wrist straps) equipped with an accelerometer. The units are programmed to detect alls based on the analysis of accelerometer data and to send an alarm message to a caregiver. Unfortunately, the potentia benefit of these systems is limited by poor compliance, because subjects are overwhelmed by the large number o alse detections of falls (1.e., false positives) that mark existing systems. This is somehow inevitable because fall-detection systems have to be extremely sensitive to the occurrence of a all. To achieve high sensitivity, low specificity (i.e., high rate of false detections) has to be tolerated. In the system shown in Figure 8, a home robot is combined with the use of a body- worn unit to minimize the number of false positives. The body-worn unit sends a message to the robot when the unit detects a fall. The robot responds by using a combination of video processing and human-robot interaction techniques to assess whether the subject actually fell. If the robot determines that the subject fell or if it cannot determine whether the