Project information

• Developed a fault-tolerance approach for tactile sensors in robotic systems using a genetic algorithm to detect and compensate for faulty taxels within a matrix of sensing elements.
• Designed a piezoresistive tactile sensor in a 4x4 grid, attached to a robotic finger, and tested with graded textures to simulate human touch sensations.
• Injected faults (e.g., drift, hysteresis) using Gaussian noise and developed a sensor fault detection method leveraging correlation estimation and statistical measures.
• Implemented a genetic algorithm with custom fitness functions using Discrete Fourier Transform parameters to correct faulty readings, with crossover and mutation for adaptive fault compensation.
• Project Report: Link