Current Search:   Rahmatizadeh, Rouhollah (x)

View All Items

  • CSV Spreadsheet
(1 - 2 of 2)
Learning robotic manipulation from user demonstrations
Title
Learning robotic manipulation from user demonstrations.
Creator
Rahmatizadeh, Rouhollah, Boloni, Ladislau, Turgut, Damla, Jha, Sumit Kumar, University of Central Florida
Abstract / Description
Personal robots that help disabled or elderly people in their activities of daily living need to be able to autonomously perform complex manipulation tasks. Traditional approaches to this problem employ task-specific controllers. However, these must to be designed by expert programmers, are focused on a single task, and will perform the task as programmed, not according to the preferences of the user. In this dissertation, we investigate methods that enable an assistive robot to learn to...
Show morePersonal robots that help disabled or elderly people in their activities of daily living need to be able to autonomously perform complex manipulation tasks. Traditional approaches to this problem employ task-specific controllers. However, these must to be designed by expert programmers, are focused on a single task, and will perform the task as programmed, not according to the preferences of the user. In this dissertation, we investigate methods that enable an assistive robot to learn to execute tasks as demonstrated by the user. First, we describe a learning from demonstration (LfD) method that learns assistive tasks that need to be adapted to the position and orientation of the user's head. Then we discuss a recurrent neural network controller that learns to generate movement trajectories for the end-effector of the robot arm to accomplish a task. The input to this controller is the pose of related objects and the current pose of the end-effector itself. Next, we discuss how to extract user preferences from the demonstration using reinforcement learning. Finally, we extend this controller to one that learns to observe images of the environment and generate joint movements for the robot to accomplish a desired task. We discuss several techniques that improve the performance of the controller and reduce the number of required demonstrations. One of this is multi-task learning: learning multiple tasks simultaneously with the same neural network. Another technique is to make the controller output one joint at a time-step, therefore to condition the prediction of each joint on the previous joints. We evaluate these controllers on a set of manipulation tasks and show that they can learn complex tasks, overcome failure, and attempt a task several times until they succeed.
Show less
Date Issued
2017
Identifier
CFE0006908, ucf:51686
Format
Document (PDF)
PURL
http://purl.flvc.org/ucf/fd/CFE0006908
Energy efficient routing towards a mobile sink using virtual coordinates in a wireless sensor network
Title
Energy efficient routing towards a mobile sink using virtual coordinates in a wireless sensor network.
Creator
Rahmatizadeh, Rouhollah, Boloni, Ladislau, Turgut, Damla, Jha, Sumit, University of Central Florida
Abstract / Description
The existence of a coordinate system can often improve the routing in a wireless sensor network. While most coordinate systems correspond to the geometrical or geographical coordinates, in recent years researchers had proposed the use of virtual coordinates. Virtual coordinates depend only on the topology of the network as defined by the connectivity of the nodes, without requiring geographical information. The work in this thesis extends the use of virtual coordinates to scenarios where the...
Show moreThe existence of a coordinate system can often improve the routing in a wireless sensor network. While most coordinate systems correspond to the geometrical or geographical coordinates, in recent years researchers had proposed the use of virtual coordinates. Virtual coordinates depend only on the topology of the network as defined by the connectivity of the nodes, without requiring geographical information. The work in this thesis extends the use of virtual coordinates to scenarios where the wireless sensor network has a mobile sink. One reason to use a mobile sink is to distribute the energy consumption more evenly among the sensor nodes and thus extend the life-time of the network. We developed two algorithms, MS-DVCR and CU-DVCR which perform routing towards a mobile sink using virtual coordinates. In contrast to the baseline virtual coordinate routing MS-DVCR limits routing updates triggered by the sink movement to a local area around the sink. In contrast, CU-DVCR limits the route updates to a circular area on the boundary of the local area. We describe the design justification and the implementation of these algorithms. Using a set of experimental studies, we show that MS-DVCR and CU-DVCR achieve a lower energy consumption compared to the baseline virtual coordinate routing without any noticeable impact on routing performance. In addition, CU-DVCR provides a lower energy consumption than MS-DVCR for the case of a fast moving sink.
Show less
Date Issued
2014
Identifier
CFE0005402, ucf:50422
Format
Document (PDF)
PURL
http://purl.flvc.org/ucf/fd/CFE0005402