Strategies for Robotic Sensing Using Acoustics

TLDR

Robotics is essential for modern manufacturing, and advancing productivity demands machine intelligence with real‑time adaptive sensing, for which acoustic and ultrasonic sensors can deliver range, object recognition, tactile force, and shape information that complements vision. The paper examines future system design requirements that influence the speed of machine adaptive behavior. The study reviews acoustic/ultrasonic and piezoelectric transduction systems, covering device physics, data acquisition, signal processing, and feature extraction. Current state of the art shows that substantial research is still needed.

Abstract

Robotics is an acknowledged critical component of present and future manufacturing methods. The competitive drive for still higher productivity is pressing robotics technology to provide machine intelligence capable of accurate, real-time adaptive behavior. Sensing is considered a major component of machine intelligence. For manufacturing, repair and maintenance, a sense of touch, proximity and depth perception, traditionally considered uniquely human activities, are among the key requirements for any machine to emulate - or surpass. Acoustic and ultrasonic based sensors are demonstrating a capacity to provide perceptual information for range, object recognition, tactile force and shape determination. They also complement machine vision. Numerous systems that make use of acousticshltrasonics and/or rely on the piezoelectric properties of transduction sensors will be examined: The device physics, data acquisition, signal processing, and feature extraction capabilities are always interrelated. Future requirements for systems design, which affects the speed of machine adaptive behavior, will also be discussed. The reader will note from the current state of the art that a great deal more research remains to be done.