Gastone Pietro Rosati Papini
|
Reinforcement Learning VS Episodic simulation with Lego Mind Storm |
|
()
martedì 13 febbraio 2024
The thesis aims to compare the use of RL with the use of episodic simulation in specific cases and verify its operation on a real device. For the thesis, it will be possible to use Lego Mind Storm for experimental campaigns. See the google drive folder for more detail: |
|
|
|
A neural network framework for modeling and control mechanic system |
|
()
lunedì 12 febbraio 2024
The idea of this thesis is to continue the development of an optimized framework for modeling mechanical systems using neural networks. The framework should support the design of structured neural networks from their conception to the final deployment on a microcontroller. The thesis may focus on only some aspects related to the framework. Please refer to the Neu4mes framework's GitHub for more details. See the google drive folder for more detail: |
|
|
|
Neural network optimization |
|
()
lunedì 12 febbraio 2024
Determination of a procedure for the definition of structure and optimization hyperparameter of a network when the dimension of the dataset is given. In particular, tuning of the hyperparameters of the training, the structure of the network and the characteristics of the dataset. More information can be found here: |
|
|
|
Low level vehicle modeling and control using neural networks |
|
Thesis for master students. ()
lunedì 12 febbraio 2024
The thesis aims to investigate in detail the use of structured neural networks for defining both models and control systems for small-scale vehicles that, however, represent dynamics comparable to road vehicles. See the google drive folder for more detail: |