Gassilloud, Andreas (2025) A cerebellar-inspired architecture to control a soft robot with redundant Degrees of Freedom. Master's Thesis / Essay, Artificial Intelligence.
|
Text
MasterThesisGassilloudAndreasS4201523.pdf Download (10MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
Toestemming.pdf Restricted to Registered users only Download (180kB) |
Abstract
This thesis expands on previous work using a cerebellar-inspired control architecture for robots operating under kinematic parameter uncertainty. The system combines an inverse kinematics approximator with a correction component based on Echo State Networks (ESNs). This removes the need for precise kinematic parameters required by traditional approaches based on mathematical models, making the architecture well suited for robots where parameter estimation is impractical, such as soft robots with non-rigid actuators or those with redundant degrees of freedom (DoF). The cerebellar-inspired control architecture is validated on the soft robot "Affetto," focusing on its redundant DoF arm. Three correction strategies are tested: (1) adjusting the desired end-effector input to the approximator, (2) correcting its joint angle outputs, and (3) applying both corrections simultaneously. Two ESN based correction components are compared: Ensembles of standard leaky integrator ESNs and reBASICS-type ESN ensembles. Results show that correction pathways can either interfere or combine constructively, depending on the strength of the ESN base correction component. The reBASICS-type ensembles significantly outperform the standard-type ESN ensembles, attributed to lower degree of correlation between the modules of the ensemble.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Jaeger, H. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 18 Nov 2025 13:12 |
| Last Modified: | 18 Nov 2025 13:12 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/37134 |
Actions (login required)
 |
View Item |