Virtual Stiffness: A Novel Biomechanical Approach to Estimate Limb Stiffness of a Multi-Muscle and Multi-Joint System

Abstract

In recent years, different groups have developed algorithms to control the stiffness of a robotic device through the electromyographic activity collected from a human operator. However, the approaches proposed so far require an initial calibration, have a complex subject-specific muscle model, or consider the activity of only a few pairs of antagonist muscles. This study described and tested an approach based on a biomechanical model to estimate the limb stiffness of a multi-joint, multi-muscle system from muscle activations. The “virtual stiffness” method approximates the generated stiffness as the stiffness due to the component of the muscle-activation vector that does not generate any endpoint force. Such a component is calculated by projecting the vector of muscle activations, estimated from the electromyographic signals, onto the null space of the linear mapping of muscle activations onto the endpoint force. The proposed method was tested by using an upper-limb model made of two joints and six Hill-type muscles and data collected during an isometric force-generation task performed with the upper limb. The null-space projection of the muscle-activation vector approximated the major axis of the stiffness ellipse or ellipsoid. The model provides a good approximation of the voluntary stiffening performed by participants that could be directly implemented in wearable myoelectric controlled devices that estimate, in real-time, the endpoint forces, or endpoint movement, from the mapping between muscle activation and force, without any additional calibrations.

Elbow Joint Stiffness Functional Scales Based on Hill’s Muscle Model and Genetic Optimization

“…Pathological conditions, such as post-stroke, lead to a reduction in joint stiffness and force due to the improper activation of antagonist muscle pairs. While several studies have investigated how to detect the rotational stiffness of lower-limb joints during locomotion, exploiting the cyclicity of gait or implementing musculoskeletal models driven by muscle activity, less attention has been given to the real-time estimation of upper limb stiffness modulation during non-cyclic movements [ 6 ].…”

Feasibility of Human Wrist-joint Neuromuscular System Identification Method Using Functional Electrical Stimulation in Clinical Examinations

“…We have proposed a method to evaluate the dynamic characteristics using a large robotic arm that consider muscle coordination [11], but this test is unsuitable for screening tests. Furthermore, several attempts have been made to obtain neuromuscular characteristics of joint movement using electromyography [12][13][14]; however, these require large equipment for stabilizing the neuromuscular state by deactivation or through other means.…”