TU Darmstadt / ULB / TUbiblio

A Framework for Adaptive Feedforward Motor-Control for Unmanned Ground Vehicles

Ommer, Nicolai (2016):
A Framework for Adaptive Feedforward Motor-Control for Unmanned Ground Vehicles.
Department of Computer Science (SIM, Technische Universitaet Darmstadt, Department of Computer Science (SIM), [Master Thesis]

Abstract

Autonomous robots require precise motion models to interact in their environment. Deviations from their planned path may result in collisions or inefficient motion trajectories. Motion control underlies uncertainties such as unknown ground, contact forces, hardware inaccuracies and hardware wear that can be addressed by a learned compensation model to improve accuracy. While there are many proposals for learning a motion model offline, in recent years online learning methods became more widespread. These methods enable adaptation during runtime to compensate hardware failures or to adapt to new terrain. In this thesis different online learning methods were integrated into a new framework based on ROS (Robot Operating System) for an online adaptive feedforward controller which is based on an adaptive compensation model. The framework was applied to an omnidirectional soccer robot and a tracked rescue robot, but is designed to be applicable to other systems as well.

Item Type: Master Thesis
Erschienen: 2016
Creators: Ommer, Nicolai
Title: A Framework for Adaptive Feedforward Motor-Control for Unmanned Ground Vehicles
Language: English
Abstract:

Autonomous robots require precise motion models to interact in their environment. Deviations from their planned path may result in collisions or inefficient motion trajectories. Motion control underlies uncertainties such as unknown ground, contact forces, hardware inaccuracies and hardware wear that can be addressed by a learned compensation model to improve accuracy. While there are many proposals for learning a motion model offline, in recent years online learning methods became more widespread. These methods enable adaptation during runtime to compensate hardware failures or to adapt to new terrain. In this thesis different online learning methods were integrated into a new framework based on ROS (Robot Operating System) for an online adaptive feedforward controller which is based on an adaptive compensation model. The framework was applied to an omnidirectional soccer robot and a tracked rescue robot, but is designed to be applicable to other systems as well.

Place of Publication: Department of Computer Science (SIM
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Simulation, Systems Optimization and Robotics Group
Date Deposited: 26 Jun 2019 07:47
Related URLs:
Export:
Suche nach Titel in: TUfind oder in Google

Optionen (nur für Redakteure)

View Item View Item