Fast and Accurate Tactile Sensor System for a Human-Interactive Robot

With the advent of the aging society, the demand for nursing care for the elderly is becoming much larger. The application of robotics to helping on-site caregivers is consequently one of the most important new areas of robotics research. Such humaninteractive robots, which share humans’ environments and interact with them, should be covered with soft areal tactile sensors for safety, communication, and dextrous manipulation. Tactile sensors have interested many researchers and various types of tactile sensors have been proposed so far. Many tactile sensors have been developed on the basis of microelectro-mechanical system (MEMS) technology (for example, (Suzuki, 1993; Souza & Wise, 1997)). They have a high-density and narrow covering area realized by applying MEMS technology, and as a result, are not suitable for covering a large area of a robot’s surface. Some tactile sensors suitable for use on robot fingers or grippers have also been developed (Nakamura & Shinoda, 2001; Yamada et al., 2002; Shimojo et al., 2004). Many of them have the ability to detect tangential stress and can be used in grasping force control. Their main target is robot fingers, and consequently they were not designed to cover a large area. There are also commercially available tactile sensors such as those offered by Tekscan (Tekscan, 2008) based on pressure-sensitive ink or rubber, and KINOTEXTM tactile sensors (Reimer & Danisch, 1999) utilizing the change in the intensity of light scattered by the covering urethane foam when deformed. However, they are not sufficiently accurate because of strong hysteresis and creep characteristics. The idea of covering a large area of a robot’s surface with soft tactile skinlike sensors is attracting researchers (Lumelsky et al., 2001). Some human-interactive robots for which a large area of their surface is covered with soft tactile sensors have actually been developed (Inaba et al. 1996; Tajima et al. 2002; Kanda et al. 2002; Mitsunaga et al. 2006; Ohmura et al., 2006; Ohmura & Kuniyoshi, 2007). However, the tactile sensors are not suitable for humaninteractive robots, particularly when physical labor using tactile sensation is required. For example, one tactile sensor in (Tajima et al. 2002) has only 3 values as its output, and another tactile sensor in (Tajima et al. 2002) is gel-type and cannot be used over a long period because of the evaporation of the contained water. The tactile sensor in (Mitsunaga et al. 2006) has only 56 elements in total. Flexible fabric-based tactile sensors using an electrically conductive fabric have also been proposed for covering a robot (Inaba et al. 1996), but the O pe n A cc es s D at ab as e w w w .in te ch w eb .o rg