EMG oscillator model-based energy kernel method for characterizing muscle intrinsic property under isometric contraction

Improving the Transparency of an Exoskeleton Knee Joint Based on the Understanding of Motor Intent Using Energy Kernel Method of EMG

“…Considering the effect of noise, the energy kernel method treats the EMG signal in a short time interval as a harmonic oscillator. As presented in the earlier work [19], the harmonic oscillator hypothesis also leads to another characteristic quantity of "natural frequency" of EMG signal, the physical meaning of which is still under investigation, but initial analysis showed that this natural frequency is independent of MU firing rate, further supporting the oscillator hypothesis of EMG. The semiphenomenological model can be used as a cybernetic model because the inputs and outputs of the model are all external measureable quantities, and linear frames are extracted from basic in vitro experimental data, ensuring the high computational efficiency for engineering applications [13].…”

“…In our previous paper [19], this is conducted via a "linear fencing" method, which, however, is not fast enough to meet real-time demands. Here, a "discrete box counting" method is proposed.…”

“…In this work, we use a window width of N = 600 (with 1/30 increment, sampling rate, 2000 Hz), a thr of 3, and n1 = n2 =10. The detailed validation of the performance of energy kernel method has been presented in [19]. This method holds a high signal-to-noise (SN) ratio, and is very robust because it integrates the advantages of both RMS and MPF methods by simultaneously reflecting the recruitment level and firing rate (also indicated in (7)), especially for large span voluntary contraction levels.…”

“…To extract reliably the activation level from EMG, a promising way is to find the physical correspondence between the energy of the signal and the power of muscle based on the biomechanical relationship between the firing behavior of motoneuron and muscular contraction [18]. This paper adopts and improves the highly robust energy kernel method based on the oscillator model of EMG that we proposed earlier [19].…”

A Highly Efficient Semiphenomenological Model of a Half-Sarcomere for Real-Time Prediction of Mechanical Behavior

“…Although under artificial stimulating pulse trains the force transi ents of a bundle of muscle fibers are sawtooth-shaped (especially at low frequencies), here we only focus on the smoothed trends (dashed curves) because they bear practical engineering signifi cance when large numbers of fibers are voluntarily excited within a muscle [30]. Another important feature is that the relationship between the plateau force, or saturated [i (//,), and the firing rate is not linear, but can be fitted as an exponential as shown in Fig.…”