Follow me

Wednesday, October 3, 2012

Tactile sensing for mechatronics – a state of art survey


M.H. Lee, H.R. Nicholls
Mechatronics, No. 9, 1999
Summary
                   In industrial applications, contact interaction are an important feature of physical manipulation systems, but research in the field of tactile sensors has undergone a drastic slowdown after the ’80, when is was supposed that it would had been a fundamental sensor for the following decades. A tactile sensor is therefore defined as a device that can measure a property of an object through the contact with it. The other 4 sensing modalities are basically all advanced today in technology and even computer vision has became cheaper.
The difficulties which have somehow stopped research in tactile sensors is due to the fact that in human being this sensing modality isn’t localized, it’s complicated to transduce and it’s difficult to imitate. In industry some basic forms of sensing, such as “spatial switches” are quite common and easily accessible, therefore they are not in the matter of this stat of art survey. In the 90’ the studies directed the transducing methods to the following technologies, not basically available: Resistance and Conductance, Capacitance, Piezoelectric and Pyroelectric, Magnetic, Magnetoelectric, Mechanical, Optical, Ultrasonic and strain gauges. Interesting researched have been performed in cutaneous sensor, which are basically divided in to extrinsic (mounted at or near the contact interface) and intrinsic sensing (which consist in derivation of contact data from force sensing within a mechanical structure), this study covers about the former one, which doesn’t deal with force/torque sensors. An important method for obtaining cutaneous sensor is by using array of integral sensing elements, which have been demonstrated to be capable of having a spatial resolution of 2-4 mm (Beebe). Gray and Fearing reported an 8 ⨯ 8 capacitive fabricated array of 1mm2 area.
One of the major problems regard inverse analysis, which is the issue of computing the changes on the surface from the sensed data gathered remotely through the elastic medium, since there isn’t a unique solution.
Artificial sensing fingers appear to be another interesting application for exploration and grasping, in this field at least two types of tactile sensors are considered: one for contact point localization and one for detecting more spatially diffuse dynamic events, such as contact slip.
Soft materials are becoming an interest matter for tactile sensing research and gels, followed by powders appear to be the best material in terms of impact and strain energy dissipation, conformability to surface and hysteresis effects. Also the fact that human tissue is composed by electrolytic materials have inspired researched such as Sawahata, Gong and Osada to use polyacrylamide, which, with similar mechanical properties, can capture the electrical change (piezoelectric effect). Tactile sensor can also reduce kinematic errors in stiffness control by locating precise contact point and tracking changes, being useful for dexterous multi-fingered hands (Howe).
Whiskers have also objects under study, in fact they appear to be fast, accurate and cheap, essentially being single point sensors. Son, Cutkosky and Howe demonstrated how intrinsic and extrinsic tactile sensors can be effective with less than 1 mm error contact location. Tactile sensors appear to have application also in haptic perception (integration of cutaneous surface sensing with information from position and movement variable of the manipulator), teleoperation (remote human operating a robot) and virtual reality (for which multi-sensor gloves or other actuators have been creator to provide tactile sense to the operator). Processing of the data my be with fuzzy logic, rule-based systems or model-based systems, but neural networks appear to be the fastest.
Key Concepts
Tactile Sensors
Key Results
Toyota ins an example of organization pushing workers to have “safe partnership” with robots, in order to achieve this either intrinsically safer equipment must be provided (Tobita et al.) or there must be comprehensive collision avoidance (Suita et al.), therefore tactile sensors would be a fundamental tool for human robot interaction ensuring reliability and safety conditions.

No comments: