Alireza Abbasimoshaei

This paper presents a method for an interactive surface recognition system using a robot and a vibrotactile sensor. The robot was equipped with an artificial finger made of metal to which a three-axis accelerometer was attached. The robot interacted with 10 different surfaces by performing scratching movements on each of them. Random Forest is used as the surface classification algorithm by extracting statistical features of the vibrations detected by the accelerometer. By using only one accelerometer as the main measurement feedback, the results show that it has acceptable classification accuracy.

Hand paralysis, caused by conditions such as spinal cord injuries, strokes, and arthritis, significantly hinders daily activities. Wearable exo-gloves and telerehabilitation offer effective hand training solutions to aid the recovery process. This study presents the development of lightweight wearable exo-gloves designed for finger telerehabilitation. The prototype uses NiTi shape memory alloy (SMA) actuators to control five fingers. Specialized end effectors target the metacarpophalangeal (MCP), proximal interphalangeal (PIP), and distal interphalangeal (DIP) joints, mimicking human finger tendon actions. A variable structure controller, managed through a web-based Human–Machine Interface (HMI), allows remote adjustments. Thermal behavior, dynamics, and overall performance were modeled in MATLAB Simulink, with experimental validation confirming the model’s efficacy. The phase transformation characteristics of NiTi shape memory wire were studied using the Souza–Auricchio model within COMSOL Multiphysics 6.2 software. Comparing the simulation to trial data showed an average error of 2.76°. The range of motion for the MCP, PIP, and DIP joints was 21°, 65°, and 60.3°, respectively. Additionally, a minimum torque of 0.2 Nm at each finger joint was observed, which is sufficient to overcome resistance and meet the torque requirements. Results demonstrate that integrating SMA actuators with telerehabilitation addresses the need for compact and efficient wearable devices, potentially improving patient outcomes through remote therapy.

This paper presents a new bioinspired whisker sensor system tailored for surface defect detection. A rotational whisker sensor is designed and installed on a homemade wheeled robot, enabling active detection across a wide area. The sensor array is made of carbon fibers reinforced with polyepoxides and controlled by an ESP8266 microcontroller via Wi-Fi. The simulation investigation using Sim-Scape software indicates that adjusting dynamic parameters like frequency and response speed improves sensor performance. The study proposes that this approach proves effective in identifying surface anomaly, especially in noisy conditions.