TU Darmstadt / ULB / TUbiblio

An Approach for Runtime-Modifiable Behavior Control of Humanoid Rescue Robots

Schillinger, Philipp (2015):
An Approach for Runtime-Modifiable Behavior Control of Humanoid Rescue Robots.
Technische Universitaet Darmstadt, Department of Computer Science (SIM), [Master Thesis]

Abstract

This thesis describes a novel approach for modification of robot behaviors during runtime. Existing high-level robot control has very limited adaptability regarding unexpected disturbances. Possible uncertainties have to be known and explicitly considered in advance by defining strategies of how to react to these. This requirement confines the development of robust robots as for example required in complex and unstructured disaster mitigation scenarios. In order to overcome this limitation and to facilitate the development of more flexible high-level robot behavior control, the approach developed in this thesis enables to change the whole structure of behaviors even while they are executed. Therefore, the operator is able to incorporate situational knowledge gained during execution, and thus to compensate insufficient a priori knowledge about the status of the environment. The approach is proposed based on modeling behaviors as hierarchical state machines, allowing for modular composition and intuitive specification in different levels of abstraction. Detailed monitoring of the state of execution and occurred errors assists the operator when giving commands and adjusting the level of autonomy, utilizing the individual capabilities of both robot and operator in a cooperative manner. Nevertheless, the developed framework is able to cope with severe restrictions on the communication channel to the robot and is robust regarding runtime failure. In addition, verification of specified behaviors greatly reduces the risk of failure. As part of this thesis, the behavior engine FlexBE has been developed in order to implement and refine the promoted concepts. It comes with a comprehensive user interface and behavior editor and is practically applied in the upcoming DARPA Robotics Challenge Finals.

Item Type: Master Thesis
Erschienen: 2015
Creators: Schillinger, Philipp
Title: An Approach for Runtime-Modifiable Behavior Control of Humanoid Rescue Robots
Language: English
Abstract:

This thesis describes a novel approach for modification of robot behaviors during runtime. Existing high-level robot control has very limited adaptability regarding unexpected disturbances. Possible uncertainties have to be known and explicitly considered in advance by defining strategies of how to react to these. This requirement confines the development of robust robots as for example required in complex and unstructured disaster mitigation scenarios. In order to overcome this limitation and to facilitate the development of more flexible high-level robot behavior control, the approach developed in this thesis enables to change the whole structure of behaviors even while they are executed. Therefore, the operator is able to incorporate situational knowledge gained during execution, and thus to compensate insufficient a priori knowledge about the status of the environment. The approach is proposed based on modeling behaviors as hierarchical state machines, allowing for modular composition and intuitive specification in different levels of abstraction. Detailed monitoring of the state of execution and occurred errors assists the operator when giving commands and adjusting the level of autonomy, utilizing the individual capabilities of both robot and operator in a cooperative manner. Nevertheless, the developed framework is able to cope with severe restrictions on the communication channel to the robot and is robust regarding runtime failure. In addition, verification of specified behaviors greatly reduces the risk of failure. As part of this thesis, the behavior engine FlexBE has been developed in order to implement and refine the promoted concepts. It comes with a comprehensive user interface and behavior editor and is practically applied in the upcoming DARPA Robotics Challenge Finals.

Place of Publication: Technische Universitaet Darmstadt
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Simulation, Systems Optimization and Robotics Group
Date Deposited: 26 Jun 2019 07:49
Related URLs:
Export:
Suche nach Titel in: TUfind oder in Google

Optionen (nur für Redakteure)

View Item View Item