Human Tactile Sensing of Robots

Human dexterity is a marvelous thing: people can grasp a wide variety of sizes and shapes, perform complex tasks, and switch between grasp in response to changing task requirements. This is due to our sophisticated control part capabilities. In large measure this control capability is founded force and tactile sensing, especially the ability to sense conditions at the finger object contact.

For the last two decades robotics researchers have worked to create a touch artificial sense to give robots some of the same manipulation capabilities that humans posses. While vision has received the most attention in robot sensing research, touch is vital for many tasks. Dextrous manipulation requires control of motions and forces at the contact between the fingers and the environment, which can only be accomplished through touch. Tactile sensing is able to provide information about mechanical properties such as friction, compliance and mass. Knowledge of these parameters is essential if robots are to reliably handle unknown objects in unstructured environment.

Although touch sensing is the dextrous basis manipulation, early work in tactile sensing research focused on the creation of sensor device and object recognition algorithms. Particular attention has been devoted to skin-like array sensors that built a simple tactile array sensor and demonstrated recognition of flat objects such as washers. The multi-fingered hands creation increased interest in tactile sensing for manipulation, beginning with preliminary work on incorporating tactile information in manipulation. In the last few years studies on the use of sensing tactile in real time control of manipulation have begun to appear. Tactile sensors provided information that guided the execution of the tasks, including edge tracking, automatic grasping, and rolling manipulation. These experimental studies have begun to explain the ways that sensing tactile enhances manipulation capabilities and many questions remain unanswered. Currently we lack a comprehensive theory that defines sensing requirements for various manipulation tasks.

No comments:

Post a Comment