Full Record

New Search | Similar Records

Title Action learning from human demonstrations for personal robots:
Publication Date
Degree Level masters
University/Publisher Delft University of Technology
Abstract Household robots need to perform tasks specific for the owner. With Learning from demonstration (LfD) a robot can learn new tasks from human demonstrations, without requiring programming skills. This thesis investigates a novel representation of actions that can be learned by using only a 3d camera and an object tracker. The action representation is objectbased so it is independent of the morphology of the robot. The actions are represented using the average and standard deviation of multiple demonstrated trajectories with six degrees of freedom. The standard deviation serves as a weight factor for the required accuracy of the recognized or synthesized trajectory. Three novel methods proposed in this thesis aim to reduce variances in the demonstration that are not specific to the action. First the demonstrations are aligned in time using a novel action signature and a novel time warp algorithm. The time warp algorithm can approximate the alignment of multiple multidimensional signals in quadratic computing time. The third novel technique is a dynamically optimized choice of reference frame so variations in start and end position have little influence on the variance in trajectory. This method has been tested on a database of five actions repeatedly demonstrated by six subjects. The results show that it is possible to have a 90 percent action recognition rate with only three demonstrations in the database. It is also shown that a robot can use this action representation to synthesize four out of five actions with varying object positions.
Subjects/Keywords learning from demonstrations
Contributors Jonker, P.P.; Rudinac, M.
Language en
Rights (c) 2013 Rozemuller, C.G.
Country of Publication nl
Record ID oai:tudelft.nl:uuid:4dbcecf9-e143-4f16-b315-f4a509140b7d
Other Identifiers uuid:4dbcecf9-e143-4f16-b315-f4a509140b7d
Repository delft
Date Indexed 2017-06-19

Sample Search Hits | Sample Images

…referred to as: Learning by Demonstration (LbD), Programming by Demonstration (PbD), Learning by Experienced Demonstrations, Assembly Plan from Observation, Learning by Showing, Learning by Watching, Learning from Observation, behavioral…

…learned these techniques after analyzing the human demonstrations. • Recently researchers made a robot that could help assemble furniture after learning from observation [11] The only thing that all these examples have in common is that they are…

learning. With techniques such as imitation, simulation and sensory exploration the system should learn new knowledge from several demonstrations Tuning all the subsystems in an on-line manner. Its algorithms can be based on the ideas of statistical…

…C. G. Rozemuller 14 Related work on learning from demonstration Unknown Demonstration Multiple Demonstrations Recognition Start & End Position Learning Object Tracker Action Signature Object Tracker Align Trajectory average signature…

…51 51 52 A Learning from demonstration in literature 55 A-1 Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 A-2 Research groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56…

…this team. The LfD framework developed in this research should be applicable to this robot. 1-2 Learning from demonstration The ideal household robot should be able to take over any task that the user requires. Therefore everyone should be able to…

…demonstrating. Making a robot learn from human demonstrations will be the topic of this thesis. Master of Science Thesis C. G. Rozemuller 2 Introduction Figure 1-1: The personal robots developed by the Delft Robotics team. On the left is Robby developed in…

…2012 and on the right Lea developed in 2013 C. G. Rozemuller Master of Science Thesis 1-3 Examples 3 The field of Learning from Demonstration (LfD) 1 deals with the problem of transferring a policy (task) from a human teacher…