Robotic manipulation in restricted spaces is still a great challenge as the conventional robotic hands have limited accessibility, dexterity and mobility in restricted spaces. To achieve better manipulation in confined spaces, this research presents the design and testing of a dexterous robotic finger. The proposed finger is designed based on the chopsticks used by Vietnamese people and has 3-DOFs for adaptive grasping without complex actuators. A mathematical model of contact forces between the components of fingers and grasped objects is created and investigated. A series of simulations are carried out on MATLAB to evaluate the motion behaviour, force transmission, workspace accessibility and grasp stability. Performance data are from simulation, prototype testing. An Adaptive Red Panda Optimized Dynamic Artificial Neural Network (ARPO-ANN) model is used to evaluate the quality of grasping, force distribution, and its effectiveness in manipulation. The proposed system successfully demonstrates high-level accuracy (98.2%); high success rate (98.4%); good grasp force (18.9 N); high accessibility; stability, and operational efficiency in the restricted space.
Keywords: accessibility enhancement, dexterous robotic finger, narrow operational spaces, robotic manipulation, vietnamese chopstick-inspired design
