Reach Out and Touch Some Robot
The news media were excited to report an advance in materials science last week that could pave the way for touchy-feely robots (see BBC News, News @ Nature, LiveScience and National Geographic News, for instance). Two scientists produced a thin film with touch resolution comparable to that of a human finger, an order of magnitude better than previous attempts. The work was published in Science.1
While this is indisputably a remarkable achievement, few of the popular articles drew detailed comparisons and contrasts with the capabilities of human skin. Richard Crowder, in the same issue of Science,2 started his analysis of the paper with a touch of realism:
Today’s state-of-the-art dexterous robotic hands cannot achieve tasks that most 6-year-old children can do without thinking, such as tie a shoelace or build a house of cards. The improvement of the manipulative capabilities of robotic hands requires advances in a wide range of technologies, including mechanics, actuators, sensors, and artificial intelligence. Many robots—such as NASA’s Robonaunt—have the dexterity required to perform some of the tasks that we take for granted, but replication of the full manipulative capabilities of the human hand is still years away.
Calling the development of tactile sensors “one of the most difficult problems in robotics,” went on to describe how difficult it is to measure slippage while gripping an object. Though he had praise for the touch-sensitive film invented by Maheshwari and Saraf, he continued to underscore the problems facing robot designers emulating a human hand.
To appreciate the advantages of having a single sensor for both slip and texture, it is worth considering how it can be integrated into robotic hands that need to restrain and manipulate a wide range of objects under a wide range of conditions. To achieve a satisfactory grasp, optimal force control is required. Any movement of a robotic hand may result in the grasped object slipping and possibly being dropped; hence, the sensors on the hand have to register any slippage and adjust the applied forces to bring the object back to rest….
In short, he wrote, “The problem of defining the required grasp force is crucial and can be posed as an optimization problem.” The new tactile sensory material is thus only part of a multi-faceted problem that will require mastery of many disciplines. “The next stage in the development of this sensor is to look at its robustness and performance characteristics over time,” Crowder ended. “Once these issues are resolved, the sensor can be integrated into a dexterous hand, hopefully leading to an improvement in dexterity.” Fortunately, robotics designers will have their human sense of touch to aid them in this endeavor.
1Maheshwari and Saraf, “High-Resolution Thin-Film Device to Sense Texture by Touch,” Science, 9 June 2006: Vol. 312. no. 5779, pp. 1501 – 1504, DOI: 10.1126/science.1126216.
2Richard Crowder, “Toward Robots That Can Sense Texture by Touch,” Science, 9 June 2006: Vol. 312. no. 5779, pp. 1478 – 1479, DOI: 10.1126/science.1129110.
Would you want a robotic hand encircling your neck, with the inventor’s assurance it will only apply the right force to give you a gentle massage? What if the software, it turns out, needs a patch?
The researchers can be rightly satisfied of their achievement, but use this story as an occasion to consider the wonder of human touch. Articles like this one on Apologetics Press remind us of the exquisite engineering behind this sense that brings such pleasure and awareness to life. More than a mere force-sensitive material, human skin is integrated with a powerful processing and feedback apparatus (the central nervous system) with force actuators (muscles) and automatic temperature controls. In addition, skin is self-repairing, nourished by a circulating network of nutrients, armed with defenses, and able to breathe without leaking.
Inventors can make telescopes stronger than eyes, and machines stronger than muscles, and sensors more precise than nerves. But let no man boast till he can optimize all capabilities of a human being in a 150-lb package (more or less) and make it all reproduce itself through a cell the size of a pinhead. Even then, the inventor will not have breathed into his creation the breath of life to make it a living soul. He will only have underscored the degree of intelligent design it takes to design and optimize integrated, multifunctional systems.
We hope robotics science will continue to advance – not only toward the improvement of our lives, but toward illustrating the principle that integrated, optimized systems do not arise from unguided processes.