Mechanical tactile sensors for robotic systems using machine learning in fluids
Problem statement
Conventional measurement systems employ sensor arrays that are indirectly coupled with the actual mechanical setup requiring complex calibration procedures to establish accurate correlations between the sensor output and actual system parameters. This project investigates the development of a direct mechanical sensing approach that extracts quantitative information from intrinsic mechanical system dynamics, eliminating the need for electronic calibration and improving measurement reliability. The primary objectives include: 1) quantitative validation of mechanical measurement accuracy using conventional sensing methodologies, 2) development of a reproducible experimental framework for direct mechanical sensing, and 3) characterization of system performance parameters. The initial phase will focus on creating a fundamental setup consisting of a pipe section and pump configuration to serve as a proof-of-concept for direct mechanical sensing. The successful completion of this preliminary investigation will provide the foundation for scaling the technology to multi-node sensor networks, potentially offering significant advantages in measurement fidelity and system reliability for mechanical monitoring applications.
Team members
Eaton Yap – communicator
Tony Xiao – accountant
Liam Ross – facilitator
Robert Rafferty – admin
Client
James Pikul
UW – Mechanical Engineering