Analysis of Properties of an Active Linear Gesture Sensor

Abstract

Basic gesture sensors can play a significant role as input units in mobile smart devices. However, they have to handle a wide variety of gestures while preserving the advantages of basic sensors. In this paper a user-determined approach to the design of a sparse optical gesture sensor is proposed. The statistical research on a study group of individuals includes the measurement of user-related parameters like the speed of a performed swipe (dynamic gesture) and the morphology of fingers. The obtained results, as well as other a priori requirements for an optical gesture sensor were further used in the design process. Several properties were examined using simulations or experimental verification. It was shown that the designed optical gesture sensor provides accurate localization of fingers, and recognizes a set of static and dynamic hand gestures using a relatively low level of power consumption.

The Passive Operating Mode of the Linear Optical Gesture Sensor

“…In the active mode, the LEDs of the sensor pulsate synchronously with the frequency of 40Hz [27]. In this mode 4 IR LEDs are the source of the light which reflects from a hand performing a gesture nearby the sensor.…”

“…For example, in the active mode, when the value of the standard deviation of the 8 values in DF exceeds a given, experimentally established threshold value (T sd ), the analyses of the stored pattern are triggered. For the active mode, the T sd value was previously established as 0.1V [27]. However, the relevant threshold value for the passive mode may be different.…”

“…Other interesting parameters that should be evaluated from the data frames are: the longitudinal component of the localization of a hand in relation to the sensor and the code of the recognized pose of a hand. The localization of the hand can be estimated by calculating the center of gravity (COG) of the light pattern represented by the DF values [27]. The second parameter, the code representing the hand pose, is the output from the selected hand pose classifier.…”

“…It was utilized in order to observe the impact of varied ambient light conditions on the parameters of the DF of the stable sensor-hand finger unit. The width of artificial fingers (32 mm) was taken upon the study performed on the group of volunteers [27]. A 2FJ obstacle was mounted in front of the center of the sensor, 20 mm from the face of the PDs (around the half of the assumed operating distance [27]).…”

“…The width of artificial fingers (32 mm) was taken upon the study performed on the group of volunteers [27]. A 2FJ obstacle was mounted in front of the center of the sensor, 20 mm from the face of the PDs (around the half of the assumed operating distance [27]).…”