Using Animatlab for Neuromechanical Analysis:
Linear Hill Parameter Calculation
Fletcher Young 1(B) , Alexander J. Hunt 2 , Hillel J. Chiel 1 , and Roger D. Quinn 1
1 Case Western Reserve University, Cleveland, OH 44106, USA
fletcher.young@case.edu
2 Portland State University, Portland, OR 97201, USA
Abstract. Modeling muscle-based locomotion in simulation requires reliable
interactions between kinematic motion and kinetic forces. Simplifications made
in simulation can sometimes obfuscate these interconnections and cause a system
to operate in unexpected ways. We present a methodology for calculating muscle
parameters that comprise a linear Hill muscle model in the neuromechanical software Animatlab. We examine two numerical issues in Animatlab that cause the
muscle tension to become discontinuous, show how violating certain constraints
can make the model inaccurate, and show how these issues can be avoided.
Keywords: Neuromechanics · Muscle model · Locomotion
1 Introduction
Animals are capable of adapting to their environment by adjusting their locomotion, and
maintaining equilibrium in the face of uncertain surroundings. For example, mountain
goats are able to scale sheer cliff faces in order to lick salt from a rock surface despite
having hooves as end-effectors [1]. A better understanding of interconnected biological systems that allow animals to adapt to their environment would benefit numerous
scientific disciplines including robotics, medicine, and neuroscience. Computer simulations of animal locomotion allow researchers to test biological mechanisms before
undertaking time-consuming and expensive animal experiments and can help guide
those experiments. New experiments then inform the simulation and the scientific cycle
continues.
Simulations that include both the body and nervous system of an organism, known
as neuromechanical models, provide insight into how biological systems coordinate
locomotion [2]. Researchers have developed neuromechanical models of many animals,
including cockroaches [3], rats [4, 5], and cats [6]. A predictable relationship between
neural stimulation and kinetic output is critical for developing neuromechanical models
that make useful predictions for experimental data. The utility of a computer simulation
is largely dependent on the degree to which it accurately models neural, physical, and
environmental systems, whose highly coupled nature determine the capabilities of the
organism [7, 8].
© Springer Nature Switzerland AG 2020
V. Vouloutsi et al. (Eds.): Living Machines 2020, LNAI 12413, pp. 402–414, 2020.
https://doi.org/10.1007/978-3-030-64313-3_38
Linear Hill Parameter Calculation
Fletcher Young 1(B) , Alexander J. Hunt 2 , Hillel J. Chiel 1 , and Roger D. Quinn 1
1 Case Western Reserve University, Cleveland, OH 44106, USA
fletcher.young@case.edu
2 Portland State University, Portland, OR 97201, USA
Abstract. Modeling muscle-based locomotion in simulation requires reliable
interactions between kinematic motion and kinetic forces. Simplifications made
in simulation can sometimes obfuscate these interconnections and cause a system
to operate in unexpected ways. We present a methodology for calculating muscle
parameters that comprise a linear Hill muscle model in the neuromechanical software Animatlab. We examine two numerical issues in Animatlab that cause the
muscle tension to become discontinuous, show how violating certain constraints
can make the model inaccurate, and show how these issues can be avoided.
Keywords: Neuromechanics · Muscle model · Locomotion
1 Introduction
Animals are capable of adapting to their environment by adjusting their locomotion, and
maintaining equilibrium in the face of uncertain surroundings. For example, mountain
goats are able to scale sheer cliff faces in order to lick salt from a rock surface despite
having hooves as end-effectors [1]. A better understanding of interconnected biological systems that allow animals to adapt to their environment would benefit numerous
scientific disciplines including robotics, medicine, and neuroscience. Computer simulations of animal locomotion allow researchers to test biological mechanisms before
undertaking time-consuming and expensive animal experiments and can help guide
those experiments. New experiments then inform the simulation and the scientific cycle
continues.
Simulations that include both the body and nervous system of an organism, known
as neuromechanical models, provide insight into how biological systems coordinate
locomotion [2]. Researchers have developed neuromechanical models of many animals,
including cockroaches [3], rats [4, 5], and cats [6]. A predictable relationship between
neural stimulation and kinetic output is critical for developing neuromechanical models
that make useful predictions for experimental data. The utility of a computer simulation
is largely dependent on the degree to which it accurately models neural, physical, and
environmental systems, whose highly coupled nature determine the capabilities of the
organism [7, 8].
© Springer Nature Switzerland AG 2020
V. Vouloutsi et al. (Eds.): Living Machines 2020, LNAI 12413, pp. 402–414, 2020.
https://doi.org/10.1007/978-3-030-64313-3_38