Source author record

Syn Schmitt

Syn Schmitt appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Biological Physics Robotics Artificial Intelligence cond-mat.soft Information Theory Machine Learning math.DS math.IT Neurons and Cognition

Catalog footprint

What is connected

5works

9topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2023arXiv

Learning with Muscles: Benefits for Data-Efficiency and Robustness in Anthropomorphic Tasks

Humans are able to outperform robots in terms of robustness, versatility, and learning of new tasks in a wide variety of movements. We hypothesize that highly nonlinear muscle dynamics play a large role in providing inherent stability, which is favorable to learning. While recent advances have been made in applying modern learning techniques to muscle-actuated systems both in simulation as well as in robotics, so far, no detailed analysis has been performed to show the benefits of muscles when learning from scratch. Our study closes this gap and showcases the potential of muscle actuators for core robotics challenges in terms of data-efficiency, hyperparameter sensitivity, and robustness.

preprint2022arXiv

Muscle pre-stimulation tunes viscous-like perturbation rejection in legged hopping

Muscle fibres possess unique visco-elastic properties, capable of generating a stabilising zero-delay response to unexpected perturbations. This instantaneous response -- termed "preflex" -- is crucial in the presence of neuro-transmission delays, which are particularly hazardous during fast locomotion due to the short stance duration. While the elastic contribution to preflexes has been studied extensively, research on the role of fibre viscosity due to the force-velocity relation remains unexplored. Moreover, muscle models predict conditions with saturated force-velocity relations resulting in reduced viscous-like fibre engagement. The goal of our study is to isolate and quantify the preflex force produced by the force-velocity relation. We use our approach to analyse two perturbed vertical hopping conditions, differing in their viscosity engagement at touch-down. Both cases showed stable hopping patterns despite the reduced viscous-like response to ground perturbations in the saturated case. Moreover, the force-velocity relation was not the predominant factor driving energy adjustment to disturbance intensity. From a robotics perspective, our results suggest that already a simple, constant damper mounted in parallel to an actuator could provide stabilising preflexes at and shortly after impact.

preprint2016arXiv

External control strategies for self-propelled particles: optimizing navigational efficiency in the presence of limited resources

We experimentally and numerically study the dependence of different navigation strategies regarding the effectivity of an active particle to reach a predefined target area. As the only control parameter, we vary the particle's propulsion velocity depending on its position and orientation relative to the target site. By introducing different figures of merit, e.g. the time to target or the total consumed propulsion energy, we are able to quantify and compare the efficiency of different strategies. Our results suggest, that each strategy to navigate towards a target, has its strengths and weaknesses and none of them outperforms the other in all regards. Accordingly, the choice of an ideal navigation strategy will strongly depend on the specific conditions and the figure of merit which should be optimized.

preprint2015arXiv

Evaluating Morphological Computation in Muscle and DC-motor Driven Models of Human Hopping

In the context of embodied artificial intelligence, morphological computation refers to processes which are conducted by the body (and environment) that otherwise would have to be performed by the brain. Exploiting environmental and morphological properties is an important feature of embodied systems. The main reason is that it allows to significantly reduce the controller complexity. An important aspect of morphological computation is that it cannot be assigned to an embodied system per se, but that it is, as we show, behavior- and state-dependent. In this work, we evaluate two different measures of morphological computation that can be applied in robotic systems and in computer simulations of biological movement. As an example, these measures were evaluated on muscle and DC-motor driven hopping models. We show that a state-dependent analysis of the hopping behaviors provides additional insights that cannot be gained from the averaged measures alone. This work includes algorithms and computer code for the measures.

preprint2014arXiv

Comparing different muscle activation dynamics using sensitivity analysis

In this paper, we mathematically compared two models of mammalian striated muscle activation dynamics proposed by Hatze and Zajac. Both models are representative of a broad variety of biomechanical models formulated as ordinary differential equations (ODEs). The models incorporate some parameters that directly represent known physiological properties. Other parameters have been introduced to reproduce empirical observations. We used sensitivity analysis as a mathematical tool for investigating the influence of model parameters on the solution of the ODEs. That is, we adopted a former approach for calculating such (first order) sensitivities. Additionally, we expanded it to treating initial conditions as parameters and to calculating second order sensitivities. The latter quantify the non-linearly coupled effect of any combination of two parameters. As a completion we used a global sensitivity analysis approach to take the variability of parameters into account. The method we suggest has numerous uses. A theoretician striving for model reduction may use it for identifying particularly low sensitivities to detect superfluous parameters. An experimenter may use it for identifying particularly high sensitivities to facilitate determining the parameter value with maximised precision.