Other
Publications (11) View all
-
Article: A Learning Scheme for Reach to Grasp Movements: On EMG-Based Interfaces Using Task Specific Motion Decoding Models
[show abstract] [hide abstract]
ABSTRACT: A learning scheme based on Random Forests is used to discriminate between different reach to grasp movements in 3D space based on the myoelectric activity of human muscles of the upper arm and the forearm. Task specificity for motion decoding is introduced in two different levels: subspace to move towards and object to be grasped. The discrimination between the different reach to grasp strategies is accomplished with machine learning techniques for classification. The classification decision is then used in order to trigger an EMG-based task-specific motion decoding model. Task specific models manage to outperform "general" models providing better estimation accuracy. Thus the proposed scheme takes advantage of a framework incorporating both a classifier and a regressor that cooperate advantageously in order to split the task space. The proposed learning scheme can be easily used to a series of EMG-based interfaces that must operate in real time, providing data driven capabilities for multi-class problems, that occur in everyday life complex environments.IEEE Transactions on Information Technology in Biomedicine 12/2013; · 1.68 Impact Factor -
Conference Proceeding: Task Discrimination from Myoelectric Activity: A Learning Scheme for EMG-Based Interfaces
[show abstract] [hide abstract]
ABSTRACT: A learning scheme based on Random Forests is used to discriminate the task to be executed using only myoelectric activity from the upper limb. Three different task features can be discriminated: subspace to move towards, object to be grasped and task to be executed (with the object). The discrimination between the different reach to grasp movements is accomplished with a random forests classifier, which is able to perform efficient features selection, helping us to reduce the num-ber of EMG channels required for task discrimination. The proposed scheme can take advantage of both a clas-sifier and a regressor that cooperate advantageously to split the task space, providing better estimation accuracy with task-specific EMG-based motion decoding models, as reported in [1] and [2]. The whole learning scheme can be used by a series of EMG-based interfaces, that can be found in rehabilitation cases and neural prostheses.IEEE 13th International Conference on Rehabilitation Robotics (ICORR); 10/2013 -
Conference Proceeding: Telemanipulation with the DLR/HIT II Robot Hand Using a Dataglove and a Low Cost Force Feedback Device
[show abstract] [hide abstract]
ABSTRACT: In this paper a series of teleoperation and manipulation tasks are performed with the five fingered robot hand DLR/HIT II. Two different everyday life objects are used for the manipulation tasks; a small ball and a rectangular object. The joint-to-joint mapping methodology is used to map human to robot hand mo-tion, taking into account existing kinematic constraints such as synergistic characteristics and joint couplings. The Cyberglove II motion capture dataglove is used to measure human hand kinematics. A robot hand specific fast calibration procedure is used to map raw dataglove sensor values to human joint angles and subsequently through the mapping procedure, to DLR/HIT II joint angles. A novel low cost force feedback device is de-veloped, in order for the user to be able to detect contact and perceive the forces exerted by the robot fingertips, during manipulation tasks. The design of the force feedback device is based on RGB LEDs that provide visual feedback and vibration motors that provide vibro-tactile feedback.IEEE 21st Mediterranean Conference on Control and Automation (MED); 10/2013 -
Conference Proceeding: HandCorpus, a New Open-Access Repository for Sharing Experimental Data and Results on Human and Artificial Hands
[show abstract] [hide abstract]
ABSTRACT: We present the HandCorpus, a new repository where everyone can freely share and search for different kinds of experimental data about human and robotic hands. The goal is not only to allow researchers to replicate results for benchmarking, but also to reuse data from previous experiments. The HandCorpus provides an accurate and coherent record for citing data sets, giving due credit to authors. Data sets are hierarchically indexed and can be easily retrieved using keywords and advanced search operations. A blog, a newsletter, a publication repository and applications for mobile platforms and social networks are also provided.IEEE World Haptics Conference 2013; 04/2013 -
Conference Proceeding: Functional Anthropomorphism for Human to Robot Motion Mapping
[show abstract] [hide abstract]
ABSTRACT: In this paper we propose a generic methodology for human to robot motion mapping for the case of a robotic arm hand system, allowing anthropomorphism. For doing so we discriminate between Functional Anthropomorphism and Perceptional Anthropomorphism, focusing on the first to achieve anthropomorphic solutions of the inverse kinematics for a redundant robot arm. Regarding hand motion mapping, a "wrist" (end-effector) offset to compensate for differences between human and robot hand dimensions is applied and the fingertips mapping methodology is used. Two different mapping scenarios are also examined: mapping for teleoperation and mapping for autonomous operation. The proposed methodology can be applied to a variety of human robot interaction applications, that require a special focus on anthropomorphism.IEEE International Symposium on Robot and Human Interactive Communication (RoMan); 09/2012
