Abstract
This paper presents a method for computing sensorimotor maps of braided continuum robots driven by pneumatic actuators. The method automatically creates a lattice-like representation of the sensorimotor map that preserves the topology of the input space by arranging its nodes into clusters of related data. Deformation trajectories can be simply represented with adjacent nodes whose values smoothly change along the lattice curve; this facilitates the computation of controls and the prediction of deformations in systems with unknown mechanical properties. The proposed model has an adaptive structure that can recalibrate to cope with changes in the mechanism or actuators. An experimental study with a robotic prototype is conducted to validate the proposed method.
Original language | English |
---|---|
Article number | 4 |
Journal | Frontiers Robotics AI |
Volume | 6 |
Issue number | FEB |
DOIs | |
Publication status | Published - 1 Jan 2019 |
Keywords
- Adaptive systems
- Continuum robots
- Neural networks
- Self-organizing maps
- Sensorimotor models
ASJC Scopus subject areas
- Computer Science Applications
- Artificial Intelligence