Detection of static and dynamic activities using uniaxial accelerometers

Abstract

Abstruct-Rehabilitation treatment may be improved by objective analysis of activities of daily living. For this reason, the feasibility of distinguishing several static and dynamic activities (standing, sitting, lying, walking, ascending stairs, descending stairs, cycling) using a small set of two or three uniaxial accelerometers mounted on the body was investigated. The accelerometer signals can be measured with a portable data acquisition system, which potentially makes it possible to perform online detection of static and dynamic activities in the home environment. However, the procedures described in this paper have yet to be evaluated in the home environment. Experiments were conducted on ten healthy subjects, with accelerometers mounted on several positions and orientations on the body, performing static and dynamic activities according to a fixed protocol. Specifically, accelerometers on the sternum and thigh were evaluated. These accelerometers were oriented in the sagittal plane, perpendicular to the long axis of the segment (tangential), or along this axis (radial). First, discrimination between the static or dynamic character of activities was investigated. This appeared to be feasible using an rms-detector applied on the signal of one sensor tangentially mounted on the thigh. Second, the distinction between static activities was investigated. Standing, sitting, lying supine, on a side and prone could be distinguished by observing the static signals of two accelerometers, one mounted tangentially on the thigh, and the second mounted radially on the sternum. Third, the distinction between the cyclical dynamic activities walking, stair ascent, stair descent and cycling was investigated. The discriminating potentials of several features of the accelerometer signals were assessed: the mean value, the standard deviation, the cycle time and the morphology. Signal morphology was expressed by the maximal cross-correlation coefficients with template signals for the different dynamic activities. The mean signal values and signal morphology of accelerometers mounted tangentially on the thigh and the sternum appeared to contribute to the discrimination of dynamic activities with varying detection performances. The standard deviation of the signal and the cycle time were primarily related to the speed of the dynamic activities, and did not contribute to the discrimination of the activities. Therefore, discrimination of dynamic activities on the basis of the combined evaluation of the mean signal value and signal morphology is proposed.

Mobility assessment in older people: new possibilities and challenges

“…Since the early nineties a number of papers have demonstrated the possibility of categorizing diVerent postures and activities from BFS signals (e.g., Bussmann et al 1995;Veltink et al 1996;Aminian et al 1999). These Wrst activity monitors were primarily based on the use of multiple accelerometers on trunk and leg segments.…”

Valoración de la capacidad funcional en el ámbito domiciliario y en la clínica: Nuevas posibilidades de aplicación de la acelerometría para la valoración de la marcha, equilibrio y potencia muscular en personas mayores

“…En el eje vertical las aceleraciones medidas son mucho más importantes que en los otros dos ejes, sin ser éstas despreciables 42,43 . La aceleración aumenta en magnitud de la cabeza a los tobillos.…”

Preprocessing techniques for context recognition from accelerometer data

“…Several researchers have used the mean to either directly or indirectly identify user posture (sitting, standing or lying) [11,19,22,23] and also to discriminate the type of activity as either dynamic or static [60]. Others have used the mean as input to classifiers like Neural Networks [51,59], Naive Bayes [27], Kohonen Self-Organizing Maps [29], Decision Trees [5], and even Fuzzy Inference [20].…”

“…(x,y), (x,z) and (y,z)). It then selects the pair of signals that exhibits the largest coefficients to distinguish between dynamic activities, as described in [60].…”

“…The angle between the accelerometer axis and the gravity pull can be determined from the mean of the accelerometer signal in all three axis [21,60].…”