AN ADAPTIVE SELF-SCANNING TACTILE SENSOR WITH COMPUTATIONAL CAPABILITIES

TACTILE SENSORS ARE A CLASS OF ROBOTIC SENSORS CAPABLE OF MEASURING CONTACT INTERACTION BETWEEN A ROBOTIC GRIPPER AND A GRASPED OBJECT. THE USE OF TACTILE SENSORS WILL GREATLY ENHANCE THE USE OF UNMANNED VEHICLES OR ROBOTICS IN BOTH THE MILITARY AND CIVILIAN SECTORS BY MAKING ROBOTS MORE ABLE TO DEAL WITH UNCERTAINTY IN THE ENVIRONMENT. CURRENTLY, THE USE OF TACTILE SENSORS IN REALTIME CONTROL APPLICATIONS IS LIMITED BY THE LOW UPDATE RATES AVAILABLE FROM CURRENT SENSORS. THIS PROJECT SEEKS TO OVERCOME THIS RESTRICTION BY CREATING A COMPUTATIONAL TACTILE SENSOR (CTS) THAT COMBINES SPECIAL-PURPOSE VLSI CIRCUITRY WITH AN ULTRASONIC TACTILE SENSOR TO ACHIEVE HIGH DATA RATES. THE RESULTING CTS LOWERS BANDWIDTH AND WIRE REQUIREMENTS BY PERFORMING LOCAL SELF-SCANNING OF THE TACTILE ARRAY AND PREPROCESSING OF THE TACTILE DATA. THE DATA RATES ACHIEVED BY THE CTS WILL BE SUFFICIENT FOR APPLICATIONS THAT USE TACTILE SENSORS IN REAL-TIME CONTROL APPLICATIONS. THE TECHNOLOGY AND ALGORITHMS DEVELOPED IN THIS RESEARCH EFFORT WILL ALSO BE APPLICABLE, IN GENERAL, TO ALL ARRAY-BASED TACTILE SENSOR TECHNOLOGIES, GIVING THE SAME REDUCTIONS IN BANDWIDTH AND WIRE COUNT, AND ALLOWING OPERATION IN REAL-TIME APPLICATIONS.