AI3 Final Project

This project belongs to the course of Adaptive Embodied Locomotion Control Systems, which is part of the 3rd semester of MSc in Robot Systems at University of Southern Denmark.

For a more detailed report on the project, go to the repo's Wiki.

1. Project Description.

1.1. Contents.

The repo contains several directories and files:

• controller: Implementation of the controller for the real LocoKit robot in C++. Uses the LocoKitInterface API for C++.
• Neuron: C++ implementation of a class to define neurons as objects.
• PSN: C++ implementation of a class to define Phase Switching Netwoks (L9 - Modular Neural Control).
• RK4: C++ implementation of a class with mathematical methods to solve ODEs numerically with the Runge-Kutta approximation (L8 - Dynamic Computation and Muscle Model).
• simulation: Interface for simulation of the LocoKit robot controller using LPZROBOTS C++ framework.
• SO2: C++ implementation of a class that defines an SO2 Central Pattern Generator network (L8 - Dynamic Computation and Muscle Model).
• TransferFuntion: C++ implementation of a class with mathematical transfer functions commonly used in artificial neurons.
• VRN: C++ implementation of a class to decine Velocity Reduction Networks (L9 - Modular Neural Control).
• tests: Contains tests for the above projects.

Authors

Carlos Viescas Huerta.

cavie17@student.sdu.dk

References

• LocoKit robot.
• LocoKit C API.
• LocoKit Python API.
• LPZROBOTS C++ framework.
• LPZROBOTS git repository.
• Gorobots-edu git repository.