Proprioceptive Encoding in Spiking Neural Networks for Body Pitch Estimation

Veen, Thomas van der (2024) Proprioceptive Encoding in Spiking Neural Networks for Body Pitch Estimation. Master's Thesis / Essay, Applied Physics.

Preview	Text mAppP_2024_vanderVeenT.pdf Download (12MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (152kB)

Abstract

Robust insect climbing and locomotion requires a sense of body inclination relative to the substrate. Most insects lack dedicated posture sensing organs. Therefore, it is hypothesized that insects integrate high-level parameters, such as body pitch, from proprioceptive signals. However, specific details for the representation of body posture in the central nervous system of insects remain unknown. The objective of this thesis is to address this research gap by mimicking a portion of the stick insect Carausius morosus’ central nervous and proprioceptive system to estimate body pitch using unrestricted locomotion and climbing data. An existing spiking neural network was modified and extended to simulate tactile hair proprioceptors, descending interneurons , movement primitive interneurons, and posture neurons.The posture neuron estimated body pitch with an average error of approximately 28% Gaussian noise and the climbing classifier achieved a Matthew’s correlation coefficient (MCC) of 0.59. This suggests that proprioceptive information can be effectively processed using a spiking neural network (SNN) to simulate various stages of the proprioceptive system, ultimately estimating whole-body inclination relative to the substrate. Therefore, these results are further evidence that stick insects use proprioceptive feedback to create an internal representation of higher-order parameters, such as body posture.

Item Type:	Thesis (Master's Thesis / Essay)
Supervisor name:	Chicca, E.
Degree programme:	Applied Physics
Thesis type:	Master's Thesis / Essay
Language:	English
Date Deposited:	22 Apr 2024 07:19
Last Modified:	22 Apr 2024 07:19
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/32314

Actions (login required)

View Item