Service robots meant for human interaction need to simulate touch sensitivity, or contact force is to be known for control purposes. Measurements have thus to be quick, reliable and have a good frequency response in the required band. Moreover, minimal dimensions are required, as well as an evenly responsive contact surface. We chose a commercial contact force sensor which exhibits all of these features: minimal overall dimensions, uniform measuring area, customizable range and certified linearity, no drift and hysteresis or good frequency response features. We present here our static and dynamic calibration results along with peculiarities we came across. We wanted to assess if the sensors would exhibit linear behaviour, no drift nor hysteresis as stated by the manufacturer. The sensors should offer uniform sensitivity over the sensing area, so matching section weights can be used to load them for static calibration. Besides this simple method, we also used hydrostatic pressure in a sealed chamber, getting rid of errors uncertainties coming from loading over the edges or from point loading. This method also allows for dynamic testing, as impulse loading or different frequency pressure waves can be applied to the sensors. Calibration curves will be presented in different loading conditions, focusing our attention on loading velocity and frequency. Although sensor gain exhibits remarkable changes after repeated loading, its smooth and continuous output make us think it can still be put in a feedback chain.
Static and low frequency evaluation of a thin-film contact force sensor
MANUELLO BERTETTO, ANDREA;MEILI, SILVIA
2012-01-01
Abstract
Service robots meant for human interaction need to simulate touch sensitivity, or contact force is to be known for control purposes. Measurements have thus to be quick, reliable and have a good frequency response in the required band. Moreover, minimal dimensions are required, as well as an evenly responsive contact surface. We chose a commercial contact force sensor which exhibits all of these features: minimal overall dimensions, uniform measuring area, customizable range and certified linearity, no drift and hysteresis or good frequency response features. We present here our static and dynamic calibration results along with peculiarities we came across. We wanted to assess if the sensors would exhibit linear behaviour, no drift nor hysteresis as stated by the manufacturer. The sensors should offer uniform sensitivity over the sensing area, so matching section weights can be used to load them for static calibration. Besides this simple method, we also used hydrostatic pressure in a sealed chamber, getting rid of errors uncertainties coming from loading over the edges or from point loading. This method also allows for dynamic testing, as impulse loading or different frequency pressure waves can be applied to the sensors. Calibration curves will be presented in different loading conditions, focusing our attention on loading velocity and frequency. Although sensor gain exhibits remarkable changes after repeated loading, its smooth and continuous output make us think it can still be put in a feedback chain.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
