Oliver Rogalla

Learning Robot Behaviour and Skills Based on Human Demonstration and Advice: The Machine Learning Paradigm

Service robots require easy programming methods allowing the unexperienced human user to easily integrate motion and perception skills or complex problem solving strategies. To achieve this goal, robots should learn from operators how and what to do considering hard- and software constraints. Various approaches modelling the man-machine skill transfer have been proposed. Systems following the Programming by Demonstration (PbD) paradigm that were developed within the last decade are getting closer to this goal. However, most of these systems lack the possibility for the user to supervise and influence the process of program generation after the initial demonstration was performed. In this paper a principle learning methodology is discussed, which allows to transfer human skills and to supervise the learning process including subsymbolic and symbolic task knowledge. Here, several existing approaches will be discussed and compared to each other. Moreover, a system approach is presented, integrating the overall process of skill transfer from a human to a robotic manipulation system. One major goal is to modify information gained by the demonstration in that way that different target systems are supported. The resulting PbD-system yields towards a hybrid learning approach in robotics to support natural programming based on human demonstrations and user advice.

Using gesture and speech control for commanding a robot assistant

Giving advice to a mobile robot assistant still requires classical user interfaces. A more intuitive way of commanding is achieved by verbal or gesture commands. In this article, we present new approaches and enhancements for established methods that are in use in our laboratory. Our aim is to interact with a robot using natural and direct communication techniques to facilitate robust performance of simple tasks. Within this paper, we describe the robots vision and speech recognition system. Then, we display robot control for selecting the appropriate robot reaction for solving basic manipulation tasks.

Understanding users intention: programming fine manipulation tasks by demonstration

The Programming by Demonstration (PbD) paradigm enable programming of service robots by inexperienced human users. The main goal of these systems is to allow the inexperienced human user to easily integrate motion and perception skills or complex problem solving strategies. Unfortunately, actual PbD systems deal only with manipulation based on Pick & Place operations. For complex service tasks these are insufficient. Therefore, this paper describes how fine manipulations like detecting screw movements can be recognized by a PbD system. In order to do this, finger movements and forces on the fingertips are gathered and analyzed while an object is grasped. This assumes sensory employment like a data glove and integrated tactile sensors. An overview of the used tactile sensors and the gathered signals is given. Furthermore the segmentation of users demonstration and the classification of the recognized dynamic grasp is pointed out. For classifying dynamic grasps a time delay method based on a Support Vector Machine (SVM) is used. Finally the symbolic representation of service tasks is briefly illustrated.

Programming service tasks in household environments by human demonstration

Robot assistants will only reach a mass consumer market when they are easy to use. This applies especially to the way a user programs his robot system. The only approach that enables a non-expert robot user to teach a system complex tasks is programming by demonstration. This paper explains the basic concepts for mapping typical human actions performed in a household to a robot system: the recognition of the particular user actions, the task representation and the mapping strategy itself. The execution of a mapped program can then be performed on a real robot. An experiment is presented that was carried out concerning a table laying task and proving the feasibility of this approach.

Interactive Robot Programming Based on Human Demonstration and Advice

Service robots require interactive programming interfaces that allow users without programming experience to easily instruct the robots. Systems following the Programming by Demonstration (PbD) paradigm that were developed within the last years are getting closer to this goal. However, most of these systems lack the possibility for the user to supervise and influence the process of program generation after the initial demonstration was performed. In this paper an approach is presented, that enables the user to supervise the entire program generation process and to annotate, and edit system hypotheses. Moreover, the knowledge representation and algorithms presented enable the user to generalise the generated program by annotating conditions and object selection criteria via a 3D simulation and graphical user interface. The resulting PbD-system widens the PbD approach in robotics to programming based on human demonstrations and user annotations.

Learning and Communication in Multi-Agent Systems

This paper discusses the significance of communication between individual agents that are embedded into learning Multi-Agent Systems. For several learning tasks occurring within a Multi-Agent System, communication activities are investigated and the need for a mutual understanding of agents participating in the learning process is made explicit. Thus, the need for a common ontology to exchange learning-related information is shown. Building this ontology is an additional learning task that is not only extremely important, but also extremely difficult. We propose a. solution that is motivated by the human ability to understand each other even in the absence of a common language by using alternative communication channels, such as gestures. We show some results for the task of cooperative material handling by several manipulators.

A sensor fusion approach for PbD

Since programming by demonstration (PbD) approaches have reached prime importance in interactive robot programming, sensor technology for tracking user actions and user behavior have become more and more important. However, traditional methods based on single sensor system input are already at their limits, since PbD is an application area where real-time requirement do play an important role. Thus, a sensor fusion approach is proposed which serves input data for a finite state automaton. The input sources are on the one hand a data-glove which classifies different grips and on the other hand a movable camera head which tracks the movements of the data glove as well as estimates object positions. Both sensor sources use time efficient algorithms, since sensory data must be processed in real-time. The efficiency of this approach is proven in a PbD environment were flexible infusion bags are handled and respective actions and object positions are recognized.

Integrating skills into multi-agent systems

Currently, an important topic of robotic research is the design and development of multi-agent robot systems (MASs). In these a number of autonomous robots cooperate and coordinate themselves in order to pursue given goals. The agents of an MAS not only have to work autonomously or in cooperation with other agents, but in dynamic, relatively unstructured environments. Therefore, the agents require agent-specific but flexible skills to cope with their tasks and the environments variability. On the other hand, the actions to be performed by agents in an MAS have to meet certain requirements imposed by the MASs structure. The representation of actions has to support planning, inter-agent communication, task negotiation etc. In this paper, we describe a method of combining the agent-specific nature of skills with the requirements for a general action knowledge representation inherent to MASs, by presenting elementary operations (EOs) that provide an appropriate interface.

Communication and propagation of action knowledge in multi-agent systems

Abstract Multi-agent systems (MASs) do play an important role in the construction of fault tolerant and robust robot systems. One major advantage of MAS is the fact that multiple agents work towards a common goal, having different skills for specific subtasks. Usually, agents have to use a common description of the actions to be carried out. Since agents can join and leave the MAS at any time, it is important that knowledge acquired by single agents can be transferred or propagated between agents, to ensure that knowledge is not lost, in the case that an agent leaves the system. In this paper, techniques will be presented that enable representation of common extendable action knowledge for task solutions in an agent’s knowledge base and additionally, algorithms for propagating this knowledge between agents efficiently and with minimum required communication effort.

Distributing Programs in Multi-Agent Systems

Multi-Agent Systems (MAS) do play an important role in the construction of fault tolerant and robust robot systems. One major advantage of MAS is the fact that multiple agents work towards a common goal, having different skills for specific subtasks. Usually, agents have to use a common description of the actions to be carried out. Since agents can join and leave the MAS at any time, it is important that knowledge acquired by single agents can be transferred or propagated between agents, to ensure that knowledge is not lost, in the case that an agent leaves the system. In this paper, techniques will be presented that enable representation of common extendable action knowledge for task solutions in an agent’s knowledge base and additionally, algorithms for propagating this knowledge between agents efficiently and with minimum required communication effort.