Muscle synergies for reliable NAO arm motion control: An online simulation with real-time constraints

摘要

The design and implementation of a complete virtual model of a robotic system, by simulating components and control programs, can significantly impact the general efficiency of a project. Depending on the level of detail and accuracy of the simulation, there are various areas which can be investigated, all of which affect the development life cycle to a certain extent. This study describes a neuro-driven Human-Machine Interface based on the use of muscle synergies. The proposed strategy was evaluated on a NAO robot arm, by performing an online simulation with real-time constraints, within the Gazebo simulation environment. The obtained results show that it is possible to actively control an external device at all times, by using muscle synergies, without any subject-specific musculoskeletal model. Such a tecnology aims to effectively contribute on designing and developing new generation human-robot interfaces, and motion control algorithms for intelligent robotic devices.