Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

HL: Fachverband Halbleiterphysik

HL 13: Poster Session III

HL 13.24: Poster

Dienstag, 28. September 2021, 13:30–16:30, P

Mimicking Delay-Based Self-Sustaining Gait Pattern Generators — Dennis Michaelis, •Sebastian Jenderny, and Karlheinz Ochs — Ruhr University Bochum, Chair of Digital Communication Systems, Bochum, Germany

Hardware implementations of gait pattern generators are an active field of research especially in robotics, where recent approaches are based on neural networks. Most of the latter implementations, however, only consider synaptic weight changes. In contrast to this, we design a gait pattern generator being able to learn and generate self-sustaining gait patterns based on a neural network adjusting its axonal delays. For this purpose, we synthesize a memristive circuit of a dynamic axon model serving as the basis for the neural network. Here, the circuit realization of the axon is based on Jaumann structures with memristors. A wave digital emulation of the resulting complete circuit verifies our approach by showing the successful learning and generation of self-sustained gait patterns of a dog.