No abstract

Motor control is a set of time-varying muscle excitations which generate desired motions for a biomechanical system. Muscle excitations cannot be directly measured from live subjects. An alternative approach is to estimate muscle activations using inverse motion-driven simulation. In this article, we propose a deep reinforcement learning method to estimate the muscle excitations in simulated biomechanical systems. Here, we introduce a custom-made reward function which incentivizes faster point-to-point tracking of target motion. Moreover, we deploy two new techniques, namely, episode-based hard update and dual buffer experience replay, to avoid feedback training loops. The proposed method is tested in four simulated 2D and 3D environments with 6 to 24 axial muscles. The results show that the models were able to learn muscle excitations for given motions after nearly 100,000 simulated steps. Moreover, the root mean square error in point-to-point reaching of the target across experiments was less than 1% of the length of the domain of motion. Our reinforcement learning method is far from the conventional dynamic approaches as the muscle control is derived functionally by a set of distributed neurons. This can open paths for neural activity interpretation of this phenomenon.

This paper introduces Sound stream: a low-cost, tangible and ambidextrous controller which drives a dynamic muscle-based model of the human vocal tract for articulatory speech synthesis. The controller facilitates the multidimensional inputs which are mapped to the tongue muscles in a biomechanical modeling toolkit Artisynth using a microcontroller. As the vocal tract is a complex biological structure containing many muscles, it is a challenging and computationally expensive task to accommodate control for every muscle in the proposed scheme. So, we have followed a simplified approach by controlling the selective muscles for the efficient articulatory speech synthesis. The goal for designing an ambidextrous controller is to create new possibilities of controlling multiple parameters to vary the tongue position and shape simultaneously for generating various expressive vocal sounds. As a demonstration, the user learns to interact and control a mid-sagittal view of the tongue structure in Artisynth through a set of sensors using both hands. The Sound-Stream explores and evaluates the appropriate input and mapping methods to design a controllable speech synthesis engine. 1. Wang, J. et al. (2011) “Squeezy: Extending a multi-touch screen with force sensing objects for controlling articulatory synthesis,” in Proceedings on New Interfaces for Musical Expression, Oslo, Norway, pp. 531–532.

No abstract