Machine Learning Driven Prediction of the Behavior of Biohybrid Actuators

TL;DR

Biohybrid actuators built from skeletal muscle face high biological variability and nonlinear dynamics that hinder predictability. The paper demonstrates two ML-based modeling approaches on a two-post muscle-ring platform: static predictors (random forest and feedforward NN) to estimate peak force, and a dynamic LSTM-based digital twin to reproduce full force time series under electrical stimulation, achieving $R^2$ values up to $0.9425$ for static predictions and $R^2 = 0.9956$ for dynamics. A substantial dataset of 161 experiments with $123{,}786$ time steps (0.04 s sampling) under programmable stimulation underpins the models, with baseline pre-stimulation force significantly improving NN accuracy. The results establish ML-driven predictive foundations and a digital twin for biohybrid systems, enabling future adaptive control and robust deployment, while acknowledging limitations such as dataset size and open-loop dynamics that motivate closed-loop, multi-input extensions.

Abstract

Skeletal muscle-based biohybrid actuators have proved to be a promising component in soft robotics, offering efficient movement. However, their intrinsic biological variability and nonlinearity pose significant challenges for controllability and predictability. To address these issues, this study investigates the application of supervised learning, a form of machine learning, to model and predict the behavior of biohybrid machines (BHMs), focusing on a muscle ring anchored on flexible polymer pillars. First, static prediction models (i.e., random forest and neural network regressors) are trained to estimate the maximum exerted force achieved from input variables such as muscle sample, electrical stimulation parameters, and baseline exerted force. Second, a dynamic modeling framework, based on Long Short-Term Memory networks, is developed to serve as a digital twin, replicating the time series of exerted forces observed in response to electrical stimulation. Both modeling approaches demonstrate high predictive accuracy. The best performance of the static models is characterized by R2 of 0.9425, whereas the dynamic model achieves R2 of 0.9956. The static models can enable optimization of muscle actuator performance for targeted applications and required force outcomes, while the dynamic model provides a foundation for developing robustly adaptive control strategies in future biohybrid robotic systems.

Figures (8)

• Figure 1: RFR model accuracy and residual analysis for predicted force, (A) scatter plot of actual versus predicted values and (B) residual distribution.
• Figure 2: NN model accuracy and residual analysis for predicted force, (A) scatter plot of actual versus predicted values and (B) residual distribution.
• Figure 3: Updated NN model accuracy and residual analysis for predicted force, (A) scatter plot of actual versus predicted values and (B) residual distribution.
• Figure 4: Four examples of input windows of 10 data points, the true next value of force and the predicted one that LSTM provides after training.
• Figure 5: The actual time series for biphasic asymmetric charge‐balanced experiment with pulse modulation between $[2ms,10ms]$ at $30Hz$ and the forecast time series provided by LSTM.