Schumacher, Christian (2020)
Motor and Sensor Network Topologies as Translators between Motor Control and Human Locomotion.
Technische Universität Darmstadt
doi: 10.25534/tuprints-00011846
Ph.D. Thesis, Primary publication
|
Text
PhD_Schumacher_Final.pdf Copyright Information: CC BY-SA 4.0 International - Creative Commons, Attribution ShareAlike. Download (32MB) | Preview |
Item Type: | Ph.D. Thesis | ||||
---|---|---|---|---|---|
Type of entry: | Primary publication | ||||
Title: | Motor and Sensor Network Topologies as Translators between Motor Control and Human Locomotion | ||||
Language: | English | ||||
Referees: | Seyfarth, Prof. Dr. Andre ; Ijspeert, Prof. Dr. Auke Jan | ||||
Date: | June 2020 | ||||
Place of Publication: | Darmstadt | ||||
DOI: | 10.25534/tuprints-00011846 | ||||
Abstract: | Human locomotion requires a complex interplay of the mechanical, sensor, neural and motor systems of motor control. Still, cyclic locomotion tasks such as walking and running can be described by simple mechanical concepts, as identified by biomechanical Template models. This dissertation asks how the complex motor control system produces simple patterns of human locomotion and aims to identify mechanisms of motor control to bridge the gap between the sophisticated body morphology and the low-dimensional structure of locomotion. Inspired by previous research, this work hypothesises that networks of the motor and sensor systems can provide structural solutions to simplify motor control. Three studies investigate the role of these networks for generating elementary behaviours of locomotion. Two works evaluate the motor network during steady-state locomotion and the response to unforeseen balance perturbations, while one study addresses sensor networks in dynamic hopping motions. The first study, a meta-analysis, reviews the specific role of biarticular muscles for multi-joint coordination. The framework of locomotor subfunctions helps to categorise the diverse literature from biomechanics, biology and robotics. Conceptual models indicate that biarticular muscles can sense and act in global leg coordinates (leg length and orientation) instead of joint coordinates if the leg design approximates the human anatomy. Evidence from human experiments shows benefits for coordinating the segmented leg, improving motion economy and controlling angular momentum. While many of these principles provide the potential to enhance robotic system designs, these concepts await further exploitation in robotic hardware. While the main body of experimental studies focused on steady-state motions, only a limited amount of research comprised unforeseen perturbation scenarios to study biarticular muscles. To fill this gap of knowledge, the second work investigates the role of biarticular muscles to realign the upper-body after postural balance perturbations. A new device is used to produce specific, impulse-like pitch perturbations with only minimal effects on the centre of mass dynamics. Subjects respond with intense reflex activity in biarticular thigh muscles, while monoarticular muscles respond only moderately. This study provides strong evidence that biarticular muscles are preferably used to realign the upper-body. The final study investigates the function of blended reflex pathways to generate steady-state hopping patterns. A neuromechanical simulation model predicts pathway-specific activation patterns resulting in different motion characteristics: performance, efficiency and safety. These results indicate that elementary compositions in the sensor network can produce task-relevant behaviours. Novel sensor-motor maps visualise the solution space of the predicted hopping patterns to evaluate the specific feedback contributions for generating a repulsive leg function. These maps are compact, united and consistent over changes in body morphology and environment, which suggests a simple learning problem. In summary, this thesis investigates peripheral networks of motor control, namely networks of muscle and sensory mechanoreceptors, and their function for simplifying the control task of the neural system. The main contribution of this work is the identification of the supporting role of these networks that allows the neural system to exploit the fundamental dynamics of locomotion. These results stimulate future research on human experiments and robotic demonstrators to validate the conceptual theory presented here. |
||||
Alternative Abstract: |
|
||||
URN: | urn:nbn:de:tuda-tuprints-118463 | ||||
Classification DDC: | 500 Science and mathematics > 500 Science 500 Science and mathematics > 570 Life sciences, biology 600 Technology, medicine, applied sciences > 610 Medicine and health 600 Technology, medicine, applied sciences > 620 Engineering and machine engineering 700 Arts and recreation > 796 Sports |
||||
Divisions: | 03 Department of Human Sciences > Institut für Sportwissenschaft > Sportbiomechanik | ||||
Date Deposited: | 06 Aug 2020 12:04 | ||||
Last Modified: | 06 Aug 2020 12:04 | ||||
URI: | https://tuprints.ulb.tu-darmstadt.de/id/eprint/11846 | ||||
PPN: | |||||
Export: |
View Item |