Human Skill Transfer System via Novint Falcon

Size: px
Start display at page:

Download "Human Skill Transfer System via Novint Falcon"

Transcription

1 Human Skill Transfer System via Novint Falcon Tarinee Tonggoed and Siam Charoenseang Abstract This paper presents a skill transfer system of hand movement via Novint Falcon. In the research, expert can demonstrate his or her hand movement skill through the 3 DOF device while the novice is able to practice how to perform that trajectory on the same device by himself or herself. The proposed skill transfer system composes of skill modeling, skill reproduction, skill playback, and skill evaluation. During skill modeling, the mixture model algorithm is selected to model the hand s trajectory. The trajectory data can be reproduced by applying a mixture regression algorithm in the reproduction process. Skill playback then receives reproduced data to present the modeled trajectory of the device to the novice. After training, the novice will perform the trajectory that he or she already practiced through the device. Then, the skill evaluation process will determine the similarities of spatial and temporal values between the modeled trajectory and the captured novice s trajectory. The experimental results show that this proposed system is efficient enough for modeling and reproducing the captured trajectory in the human skill transfer from the expert to the novice. I. INTRODUCTION During the past decades, many research works have contributed on human s skill modeling and they also transferred some skills related to manipulation to the robots. This field of research also covers the robot programming by demonstration (PbD) or learning from demonstration (LfD) and there are several approaches to apply the learning methods to the robots. Yang[1] and Calinon[2] have investigated the effectiveness of using Hidden Markov Model in skill modeling. Another interesting method for modeling skill is a mixture model (GMM) which is used in[3] and [4]. Besides robot programming by demonstration, the skill transferring to human operator is also an attractive research field. Nechyba and Yangsheng studied about skill transferring from an expert to a less-skilled operator[5]. By using Neural Network in the learning process, they focused on transferring human control strategy in a simulated inverted pendulum system from an expert user to a less-experienced. Kazuyuki HENMI presented a calligraphy transfer skill system using haptic virtual reality technology [6]. Jorge Solis proposed the idea of using the haptic system to exert the proper force to help user to perform tasks appropriately [7]. There are not many research works which cover both expert s skill modeling and transferring to novice with multiple DOF devices. system. Usually, they cover only 1 or 2 DOF skill transfer In this paper, a development of a skill transfer system that can teach the non-expert with a skill related to the expert s hand trajectory using a 3-DOF Novint Falcon touch device is proposed. mixture model is chosen for modeling the expert s manipulation skill. The component number of mixture model is an important parameter for obtaining an efficient model. In this research, an estimation of the component number based on the changes of trajectory direction will be discussed. Furthermore, the regenerated data sent to the device is also reproduced by using mixture regression. Then, the system will evaluate the temporal and spatial similarities between the novice s hand trajectory and the modeled expert s trajectory. The preliminary experimental results are collected and presented to show the performance of the proposed system. This paper is organized as follows. System Overview section covers details of the whole system. Experimental setup and results section show the experimental results of the proposed system. Finally, conclusions and further works are presented in the last section. A. System Overview II. SYSTEM STRUCTURE The proposed system consists of an expert, a novice, a 3D Novint Falcon touch device [8], and a computer as shown in Fig.1. In this system, the expert plays the role of a teacher to transfer motion skill by holding a gripper of the Novint Falcon to perform the hand movement via that device. The Novint Falcon acts as both a sensor and an actuator. As a role of the sensor, the device will capture and send the hand trajectory of expert to the computer. In the other hand, the Novint Falcon will present touch feedback accordingly to the modeled movement to the novice during the training. The computer will record both spatial and temporal data during system training for skill capture and novice training for skill transfer. The computer will build the hand s trajectory model from the recorded data and evaluate the trajectory similarity of novice s movement and the captured path. In addition, the novice can use the Novint Falcon to practice the captured skill which is reproduced by the computer. Tarinee Tonggoed is with the Institute of Field Bobotics, King Mongkut s University of Technology, Thonburi, Bangkok, Thailand, (corresponding author to provide phone: +(66) , ; fax: +(66) ; @st.kmutt.ac.th). Siam Chareonseang is with the Institute of Field Bobotics, King Mongkut s University of Technology, Thonburi, Bangkok, Thailand, (corresponding author to provide phone: +(66) , ; fax: +(66) ; siam@fibo.kmutt.ac.th).

2 A. Mixture Model (GMM) Data modeling from a mixture of distribution K is the main idea of mixture model. For example, if there are n data sequences and each data sequence has a length of T, data set contains N = nt data points. Let data N set be α = {α j } j=1 which α j = {α t, α s } and p(α j ) is defined as the posterior probability. By using Bayes theorem, p(α j ) is computed as in B. System Data Flow Figue 1 System Overview In system training, the expert will demonstrate the hand movement trajectory through the Novint Falcon. The data gathering module then records that trajectory in the form of position (x, y, and z) and time (t). The recorded data will be sent to the trajectory modeling module which builds the captured trajectory s model by using the mixture model algorithm. Next, the system will reproduce the captured trajectory from the model by the reproduction module. steps and Cartesian positions are the outputs of this module. After the reproduction process, the playbacked data will be sent to the robot controller module which is responsible to control the actuator s movement using PID control algorithm. Forces in x, y, and z axes are generated from that module and sent to the Novint Falcon module to perform its movements. This will guide the novice to learn and practice the modeled trajectory obtained from the expert. During the evaluation phase, the data gathering module will capture and send the obtained data which are time steps (t) and position (x, y, and z) of the novice s hand trajectory to the evaluation module. The module will compare the likeness between the expert s trajectory model and the novice s trajectory of using the Novint Falcon. The system data flow in Fig.2. p(α j ) = K k=1 p(k)p(α j k) (1)[9] p(k) are priors of each K and p(α j k) is a conditional probability density function as in p(α j k) = N(α j ; μ k, k ) (2) [9] From above equations, GMM parameters are K {π k, μ k, k } k=1. Where π k, μ k and k are prior, mean, and covariance matrix of the kth. For learning of GMM parameters, k-mean clustering technique is selected to initiate the GMM s parameters and Expectation- Maximization algorithm is used to train those parameters. In this paper α j is a D-dimensional data set with D = 4. The data sequences have n = 5 and each sequence s length is T = 40. B. Mixture Regression (GMR) Mixture Regression is an algorithm for data recovering from the trained GMM. From the previous section, GMM is defined by prior(π k ), mean(μ k ), and covariance ( k ) matrices. The condition expectation of α s given α t can be described as in Where, K p(α s α t ) ~ k=1 β k N(α k, k ) (3) [9] β k = p(k α t ) (4) [9] In this research, α t is the time step with length of 40 and step size is 1. Finally, the position in x, y, and z at each given time step can be recovered. Figure 2 System Data Flow III. LEARNING ALGORITHM AND ROBOT CONTROLLING The learning algorithms applied in this research consist of mixture model and mixture regression. The first one is applied in skill modeling process. The estimation of component number based on the direction changing will be discussed. In the reproduction process, the mixture regression is the method for data recovering. In addition, a PID controller is implemented to control the Novint Falcon s movement. C. Estimation of Component Number The number of component is the important initial parameter for updating in the expectation maximization (EM) algorithm. The idea of estimating that number in this proposed paper is based on the changes of trajectory s direction. In the research, spatial and temporal data in 4 dimensions are collected. Before the data modeling process, the other important process is the data preprocessing for obtaining more clean and meaningful data. In this research, the changing of direction is contributed for good feature extraction. The idea of this algorithm is to calculate the angle between the spatial data which are the positions in x, y, and z axes at time step t and t+1 as shown in Fig.3. After the calculation of angle, the result of that process can be classified into 8 different groups.

3 Figure 3 Direction Segmentation Start Read data 1 set Calculate Slope between data(t) and data(t+1) No Classify Slope into Group (8 groups) Figure 5 Flow Chart of Data Clustering based on Direction Changing E. PID controller Is data empty? End Yes Figure 4 Flow Chart of Data Preprocessing At the beginning, the one set of recorded data is read and the moving average with window size of 4 is applied for smoothing the raw data. After that, angle between time step or slope will be calculated and classified into each group. This process will be repeated until there is no more recorded data. The output of this process is the set of labeled data which is a feature vector. The flow chart of this data preprocessing can be shown in Fig.4. From the data preprocessing, the labeled data sets are obtained. The next process is the process for grouping data of the same label into the same cluster. The algorithm of grouping is shown in Fig.5. The last process is to find the maximum value of cluster group for each data set and average those values. In this process, the averaged component number can be obtained. A PID controller is applied in the robot controller for this system. Since the robot s workspace is not large, the integral term of the controller does not effect much on this system. The control equation of PID controller can be demonstrated as in Equation 5. u(t) = K P e(t) + K I e(t) dt + K D d Where, dt u(t) is Output signal to process e(t) is Error e(t) = r(t) y(t) r(t) is Setpoint y(t) is Actual Output e(t) (5)[10] For this system, r (t) is the target position which is obtained from the reproduction process and y(t) is the end effector current position. Ziegler-Nichols method is applied for the PID parameter tuning. The final gains are Kp = 10, Ki = 0.001, and Kd = IV. EXPERIMENTAL SETUP AND RESULTS The experimental setup is shown in Fig.6 which illustrates the expert or novice, a computer, and a Novint Falcon touch device shown in Fig. 7 which is a 3-DOF parallel delta robot with position resolution of 400 dpi in 4 x 4 x 4 workspace[8]. One expert and seven novices were asked to perform two main sets of experiments.

4 Figure 6 Experimental Setup implementation of skill transfer from the expert to the robot and the robot to the novice. A. The Experimental Results of Estimation of Component Number To evaluate this proposed algorithm, the incremental k-mean is selected to be compared with the estimation of component number. Bayesian information criterion () is selected as the criterion of stopping condition. The two 40-point trajectories are used in this research to evaluate the performance of this proposed algorithm. The experimental results are shown in Table 1 and 2. TABLE I. RESULTS OF THE PROPOSED ALGORITHM AND INCREMENTAL K-MEAN ALGORITHM FOR GESTURE 1 Figue 7 Novint Falcon [8] In the research, the Novint Falcon is used as both an input and an output device. The haptics device abstraction layer (HDAL) SDK [11] is implemented to manipulate the Novint Falcon s movement. There are two trajectories that expert will have to demonstrate to the robot as shown in Fig.8 and Fig.9 No Proposed Algorithm Incremental K-Mean Avg a ) Pull Forward b) Pull Up c) Push Down Figure 8 The First Demonstrated Trajectory of Hand Movement a ) Pull Left b) Pull Right c) Pull Forward Figure 9 The Second Demonstrated Trajectory of Hand Movement Three experimental sets were conducted to evaluate the effectiveness of this system. The first set is to test the efficiency of the proposed estimation method for component number The second set is implemented to test the system performance of trajectory transferring from the expert to the robot. The performance in skill transferring from the robot to the novice is also investigated in the third experimental set. The system performance of trajectory transferring involves with recording time, training time, and the reproducing time which will be also measured. For the data set with a size of 3 x 40, the recording process takes about 4.78 seconds, the reproducing process using mixture regression consumes about 3.6 milliseconds. Averaged training time will be also demonstrated in the experimental results of estimation component number. For this research, the consumed times in recording, modeling, and reproduction processes are fast enough for the No TABLE II. RESULTS OF THE PROPOSED ALGORITHM AND INCREMENTAL K-MEAN ALGORITHM FOR GESTURE 2 Proposed Algorithm Incremental K-Mean Avg The experimental results in Table 1 show that the component numbers after applying the proposed algorithm in gesture 1is equal to 7. The averaged modeling time is 0.44 seconds and the averaged of value is For applying the incremental k-mean algorithm, the averaged value of the proposed algorithm is used as the stopping criterion of this method. The averaged component number is 7.6 but 8 is used in actual implementation.the averaged time consuming in the modeling process is 0.50 seconds and the value is Table 2 presents that the proposed algorithm took 0.27 seconds in the modeling process. The averaged value is equal to and the component number is 6. For the incremental k-mean algorithm, time consuming in the

5 modeling process took 0.27 seconds, and value and component number are and 6, respectively. These results indicate that the proposed algorithm can obtain similar performance on these specific data as the one obtained from the incremental k-mean algorithm. B. The Experimental Results of System Performance of Trajectory Transferring from the Expert to the Robot In this research, log-likelihood is selected to evaluate the similarity between the trajectory of the Novint Falcon after being trained and the modeled trajectory obtained from the expert. The results of this evaluation show that an averaged log-likelihood of the Novint Falcon s trajectory after it performed first trajectory and second trajectory are and -5.6, respectively. Based on those values and expert s decisions, it can be concluded that the Novnint Falcon can perform the alike movement as the expected trajectory which are created from the trained model. Fig. 10 presents that performance of using GMM and GMR is effective enough for the uses in trajectory modeling and reproduction processes. C. The Experimental Results of System Performance of Trajectory Transferring from the Robot to the Novice The experiment s objective is to evaluate the effectiveness of the novice s trajectory training through the Novint Falcon. During the evaluation process, the novice is asked to practice to move his/her hand following the Novint Falcon s movement under the eye-closed and eye-opened conditions. After training under each condition, the novice performed the trajectory that he/she already has practiced on the Novint Falcon. The system will evaluate the resemblance of novice s trajectory with the model s one. The experimental results are shown in Fig.11 and Fig.12. Figure 11 Novice's Log-likelihood Plot for Gesture 1 a) Novint Falcon s First Trajectory Figure 12 Novice's Log-likelihood Plot for Gesture 2 b) Novint Falcon s Second Trajectory Figure 10 Trajectories Performed by the Novint Falcon vs. the Expected Trajectory Model Fig.11 demonstrates that the log-likelihood plots obtained from the training and playing back with the eye-opened condition in first trajectory are about 57% better than the ones obtained from the learning and playing back with the

6 eye-closed condition. For the plots of the second trajectory shown in Fig. 12, the training with the eye-opened condition gives a better performance than the one with eye-closed condition upto 85%. Also, the play back of movement with eye-opened condition provides 57% better performance than another. This means the novices can learn the manipulation skill better under the eye-opened condition. D. Conclusions and Further Works The hand movement skill transfer from the expert to the novice through the Novint Falcon was presented. The proposed system covered human s trajectory modeling and reproduction by machine learning algorithms and demonstrated the real implementations of manipulation skill transfer from the expert to the device and the device to the novice. In the process of trajectory modeling, mixture model is applied for building the captured trajectory s model. The estimation method of component number based on the changes of trajectories direction is proposed and compared with the incremental k- mean algorithm. The proposed method can provide similar performance as the well-known incremental k-mean algorithm. mixture regression is a selected method in the reproduction process to regenerate the data accordingly to the captured model. The Novint Falcon is a 3-DOF parallel robot which plays the main roles of sensor and actuator in this proposed system. The expert demonstrates the hand trajectories through Novint Falcon and the system captures and builds their path models. The regenerated path data are presented through the Novint Falcon movement in the playback and novice training processes. The experimental results indicate that the novices can learn and playback with the systems under eye-opened condition better than the ones under eye-closed condition. Several interesting points of work can be conducted to see whether the proposed system is robust enough to cope with some more sophisticated tasks which require more complex manipulation with multiple degrees of freedom. This leads in modification of robot to obtain more degrees of freedom and more accurate control. In addition, some augmented information could be implemented to provide graphical advised information in real time during the training and playing back processes. Also, this system can be enhanced for arm/hand rehabilitations along with some force/tactile feedbacks via the proposed skill transfer system. [4] Y. Lin, S. Ren, M. Clevenger, and Y. Sun, "Learning Grasping Force from Demonstration," in 2012 IEEE International Conference on Robotics and Automation USA, [5] M. C. Nechyba and X. Yangsheng, "Human skill transfer: neural networks as learners and teachers," in Intelligent Robots and Systems 95. 'Human Robot Interaction and Cooperative Robots', Proceedings IEEE/RSJ International Conference on, 1995, pp vol.3. [6] K. HENMI and T. YOSHIKAWA, "Virtual Lessoin and Its Application to Virtual Calligraphy System " in International Conference on Robotics & Automation Leuven, Belgium [7] J. Solis, C. A. Avizzano, and M. Bergamasco, "Teaching to Write Japanese Characters using a Haptic Interface " in HAPTIC' 02, [8] S. Martin and N. Hillier, "Characterisation of the Novint Falcon Haptic Device for Application as a Robot Manipulator," in Australasian Conference on Robotics and Automation (ACRA) Sydney, Australia, [9] S. Calinon, Robot Programming by Demonstration: A Probabilistic Approach: EPFL Press, ] "PID controller," [11] N. T. Incorporated, "Haptic Device Abstraction Layer (HDAL) " the USA, REFERENCES [1] J. Yang, Y. Xu, and C. S. ChenZ, "Hidden Markov model approach to skill learning and its application to telerobotics," Robotics & Control Systems vol. 10, [2] S. Calinon, F. D halluin, E. L. Sauser, D. G. Caldwell, and A. G. Billard, "Learning and Reproduction of Gestures by Imitation," IEEE Robotics & Automation Magazine, vol. 17, [3] S. Calinon and A. Billard, "Incremental learning of gestures by imitation in a humanoid robot," in 2007 ACM/IEEE International Conference on Human-Robot Interaction, 2007.

Kinesthetic Teaching via Fast Marching Square

Kinesthetic Teaching via Fast Marching Square Kinesthetic Teaching via Fast Marching Square Javier V. Gómez, David Álvarez, Santiago Garrido and Luis Moreno Abstract This paper presents a novel robotic learning technique based on Fast Marching Square

More information

A New Algorithm for Measuring and Optimizing the Manipulability Index

A New Algorithm for Measuring and Optimizing the Manipulability Index A New Algorithm for Measuring and Optimizing the Manipulability Index Mohammed Mohammed, Ayssam Elkady and Tarek Sobh School of Engineering, University of Bridgeport, USA. Mohammem@bridgeport.edu Abstract:

More information

A Calligraphy Robot - Callibot: Design, Analysis and Applications

A Calligraphy Robot - Callibot: Design, Analysis and Applications Proceeding of the IEEE International Conference on Robotics and Biomimetics (ROBIO) Shenzhen, China, December 13 A Calligraphy Robot - Callibot: Design, Analysis and Applications Yuandong Sun and Yangsheng

More information

Robot Learning by Demonstration

Robot Learning by Demonstration School of Computer Science Internship Project Report Robot Learning by Demonstration Author: Kaushik Subramanian Rutgers University Supervisor: Dr. Gerhard Lakemeyer Head of Knowledge Based Systems Group

More information

A New Algorithm for Measuring and Optimizing the Manipulability Index

A New Algorithm for Measuring and Optimizing the Manipulability Index DOI 10.1007/s10846-009-9388-9 A New Algorithm for Measuring and Optimizing the Manipulability Index Ayssam Yehia Elkady Mohammed Mohammed Tarek Sobh Received: 16 September 2009 / Accepted: 27 October 2009

More information

PSO based Adaptive Force Controller for 6 DOF Robot Manipulators

PSO based Adaptive Force Controller for 6 DOF Robot Manipulators , October 25-27, 2017, San Francisco, USA PSO based Adaptive Force Controller for 6 DOF Robot Manipulators Sutthipong Thunyajarern, Uma Seeboonruang and Somyot Kaitwanidvilai Abstract Force control in

More information

A Visualization Tool to Improve the Performance of a Classifier Based on Hidden Markov Models

A Visualization Tool to Improve the Performance of a Classifier Based on Hidden Markov Models A Visualization Tool to Improve the Performance of a Classifier Based on Hidden Markov Models Gleidson Pegoretti da Silva, Masaki Nakagawa Department of Computer and Information Sciences Tokyo University

More information

Task-level Imitation Learning using Variance-based Movement Optimization

Task-level Imitation Learning using Variance-based Movement Optimization 29 IEEE International Conference on Robotics and Automation Kobe International Conference Center Kobe, Japan, May 2-7, 29 Task-level Imitation Learning using Variance-based Movement Optimization Manuel

More information

Machine Learning A W 1sst KU. b) [1 P] Give an example for a probability distributions P (A, B, C) that disproves

Machine Learning A W 1sst KU. b) [1 P] Give an example for a probability distributions P (A, B, C) that disproves Machine Learning A 708.064 11W 1sst KU Exercises Problems marked with * are optional. 1 Conditional Independence I [2 P] a) [1 P] Give an example for a probability distribution P (A, B, C) that disproves

More information

Learning from Demonstrations with Partially Observable Task Parameters

Learning from Demonstrations with Partially Observable Task Parameters Learning from Demonstrations with Partially Observable Task Parameters Tohid Alizadeh 1, Sylvain Calinon 1, and Darwin G. Caldwell 1 Abstract Robot learning from demonstrations requires the robot to learn

More information

Incremental Learning of Skills in a Task-Parameterized Gaussian Mixture Model

Incremental Learning of Skills in a Task-Parameterized Gaussian Mixture Model Noname manuscript No. (will be inserted by the editor) Incremental Learning of Skills in a Task-Parameterized Gaussian Mixture Model Jose Hoyos Flavio Prieto Guillem Alenyà Carme Torras Received: date

More information

INSTITUTE OF AERONAUTICAL ENGINEERING

INSTITUTE OF AERONAUTICAL ENGINEERING Name Code Class Branch Page 1 INSTITUTE OF AERONAUTICAL ENGINEERING : ROBOTICS (Autonomous) Dundigal, Hyderabad - 500 0 MECHANICAL ENGINEERING TUTORIAL QUESTION BANK : A7055 : IV B. Tech I Semester : MECHANICAL

More information

Path Planning for a Robot Manipulator based on Probabilistic Roadmap and Reinforcement Learning

Path Planning for a Robot Manipulator based on Probabilistic Roadmap and Reinforcement Learning 674 International Journal Jung-Jun of Control, Park, Automation, Ji-Hun Kim, and and Systems, Jae-Bok vol. Song 5, no. 6, pp. 674-680, December 2007 Path Planning for a Robot Manipulator based on Probabilistic

More information

EEE 187: Robotics Summary 2

EEE 187: Robotics Summary 2 1 EEE 187: Robotics Summary 2 09/05/2017 Robotic system components A robotic system has three major components: Actuators: the muscles of the robot Sensors: provide information about the environment and

More information

Incremental Local Online Gaussian Mixture Regression for Imitation Learning of Multiple Tasks

Incremental Local Online Gaussian Mixture Regression for Imitation Learning of Multiple Tasks Incremental Local Online Gaussian Mixture Regression for Imitation Learning of Multiple Tasks Thomas Cederborg, Ming Li, Adrien Baranes and Pierre-Yves Oudeyer Abstract Gaussian Mixture Regression has

More information

Resolution of spherical parallel Manipulator (SPM) forward kinematic model (FKM) near the singularities

Resolution of spherical parallel Manipulator (SPM) forward kinematic model (FKM) near the singularities Resolution of spherical parallel Manipulator (SPM) forward kinematic model (FKM) near the singularities H. Saafi a, M. A. Laribi a, S. Zeghloul a a. Dept. GMSC, Pprime Institute, CNRS - University of Poitiers

More information

Artificial Neural Network-Based Prediction of Human Posture

Artificial Neural Network-Based Prediction of Human Posture Artificial Neural Network-Based Prediction of Human Posture Abstract The use of an artificial neural network (ANN) in many practical complicated problems encourages its implementation in the digital human

More information

An Efficient Model Selection for Gaussian Mixture Model in a Bayesian Framework

An Efficient Model Selection for Gaussian Mixture Model in a Bayesian Framework IEEE SIGNAL PROCESSING LETTERS, VOL. XX, NO. XX, XXX 23 An Efficient Model Selection for Gaussian Mixture Model in a Bayesian Framework Ji Won Yoon arxiv:37.99v [cs.lg] 3 Jul 23 Abstract In order to cluster

More information

Regression-Based Online Situation Recognition for Vehicular Traffic Scenarios

Regression-Based Online Situation Recognition for Vehicular Traffic Scenarios Regression-Based Online Situation Recognition for Vehicular Traffic Scenarios Daniel Meyer-Delius Jürgen Sturm Wolfram Burgard Abstract In this paper, we present an approach for learning generalized models

More information

Movement Learning and Control for Robots in Interaction

Movement Learning and Control for Robots in Interaction Movement Learning and Control for Robots in Interaction Dr.-Ing. Michael Gienger Honda Research Institute Europe Carl-Legien-Strasse 30 63073 Offenbach am Main Seminar Talk Machine Learning Lab, University

More information

Imitation Learning with Generalized Task Descriptions

Imitation Learning with Generalized Task Descriptions Imitation Learning with Generalized Task Descriptions Clemens Eppner Jürgen Sturm Maren Bennewitz Cyrill Stachniss Wolfram Burgard Abstract In this paper, we present an approach that allows a robot to

More information

10/25/2018. Robotics and automation. Dr. Ibrahim Al-Naimi. Chapter two. Introduction To Robot Manipulators

10/25/2018. Robotics and automation. Dr. Ibrahim Al-Naimi. Chapter two. Introduction To Robot Manipulators Robotics and automation Dr. Ibrahim Al-Naimi Chapter two Introduction To Robot Manipulators 1 Robotic Industrial Manipulators A robot manipulator is an electronically controlled mechanism, consisting of

More information

Optimization of Observation Membership Function By Particle Swarm Method for Enhancing Performances of Speaker Identification

Optimization of Observation Membership Function By Particle Swarm Method for Enhancing Performances of Speaker Identification Proceedings of the 6th WSEAS International Conference on SIGNAL PROCESSING, Dallas, Texas, USA, March 22-24, 2007 52 Optimization of Observation Membership Function By Particle Swarm Method for Enhancing

More information

Skill. Robot/ Controller

Skill. Robot/ Controller Skill Acquisition from Human Demonstration Using a Hidden Markov Model G. E. Hovland, P. Sikka and B. J. McCarragher Department of Engineering Faculty of Engineering and Information Technology The Australian

More information

Robotics. Chapter 25-b. Chapter 25-b 1

Robotics. Chapter 25-b. Chapter 25-b 1 Robotics Chapter 25-b Chapter 25-b 1 Particle Filtering Particle filtering uses a population of particles (each particle is a state estimate) to localize a robot s position. This is called Monte Carlo

More information

Robots are built to accomplish complex and difficult tasks that require highly non-linear motions.

Robots are built to accomplish complex and difficult tasks that require highly non-linear motions. Path and Trajectory specification Robots are built to accomplish complex and difficult tasks that require highly non-linear motions. Specifying the desired motion to achieve a specified goal is often a

More information

Three-Dimensional Sensors Lecture 6: Point-Cloud Registration

Three-Dimensional Sensors Lecture 6: Point-Cloud Registration Three-Dimensional Sensors Lecture 6: Point-Cloud Registration Radu Horaud INRIA Grenoble Rhone-Alpes, France Radu.Horaud@inria.fr http://perception.inrialpes.fr/ Point-Cloud Registration Methods Fuse data

More information

Extended Hopfield Network for Sequence Learning: Application to Gesture Recognition

Extended Hopfield Network for Sequence Learning: Application to Gesture Recognition Extended Hopfield Network for Sequence Learning: Application to Gesture Recognition André Maurer, Micha Hersch and Aude G. Billard Ecole Polytechnique Fédérale de Lausanne (EPFL) Swiss Federal Institute

More information

Learning to bounce a ball with a robotic arm

Learning to bounce a ball with a robotic arm Eric Wolter TU Darmstadt Thorsten Baark TU Darmstadt Abstract Bouncing a ball is a fun and challenging task for humans. It requires fine and complex motor controls and thus is an interesting problem for

More information

Intermediate Desired Value Approach for Continuous Transition among Multiple Tasks of Robots

Intermediate Desired Value Approach for Continuous Transition among Multiple Tasks of Robots 2 IEEE International Conference on Robotics and Automation Shanghai International Conference Center May 9-3, 2, Shanghai, China Intermediate Desired Value Approach for Continuous Transition among Multiple

More information

CALCULATING TRANSFORMATIONS OF KINEMATIC CHAINS USING HOMOGENEOUS COORDINATES

CALCULATING TRANSFORMATIONS OF KINEMATIC CHAINS USING HOMOGENEOUS COORDINATES CALCULATING TRANSFORMATIONS OF KINEMATIC CHAINS USING HOMOGENEOUS COORDINATES YINGYING REN Abstract. In this paper, the applications of homogeneous coordinates are discussed to obtain an efficient model

More information

Methodology for Translating Upper Extremity Motion to Haptic Interfaces

Methodology for Translating Upper Extremity Motion to Haptic Interfaces Methodology for Translating Upper Extremity Motion to Haptic Interfaces Nalini Vishnoi and Cody Narber and Zoran Duric Department of Computer Science George Mason University Fairfax, VA Email: {nvishnoi,

More information

CS839: Probabilistic Graphical Models. Lecture 10: Learning with Partially Observed Data. Theo Rekatsinas

CS839: Probabilistic Graphical Models. Lecture 10: Learning with Partially Observed Data. Theo Rekatsinas CS839: Probabilistic Graphical Models Lecture 10: Learning with Partially Observed Data Theo Rekatsinas 1 Partially Observed GMs Speech recognition 2 Partially Observed GMs Evolution 3 Partially Observed

More information

The University of Missouri - Columbia Electrical & Computer Engineering Department EE4330 Robotic Control and Intelligence

The University of Missouri - Columbia Electrical & Computer Engineering Department EE4330 Robotic Control and Intelligence The University of Missouri - Columbia Final Exam 1) Clear your desk top of all handwritten papers and personal notes. You may keep only your textbook, a cheat sheet, the test paper, a calculator and a

More information

Robot learning by observing humans activities and modeling failures.

Robot learning by observing humans activities and modeling failures. Robot learning by observing humans activities and modeling failures. Stefano Michieletto 1 and Alberto Rizzi 1 and Emanuele Menegatti 1 Abstract People needs are varied and different. Service robotics

More information

Learning Multiple Collaborative Tasks with a Mixture of Interaction Primitives

Learning Multiple Collaborative Tasks with a Mixture of Interaction Primitives Learning Multiple Collaborative Tasks with a Mixture of Interaction Primitives Marco Ewerton 1, Gerhard Neumann 1, Rudolf Lioutikov 1, Heni Ben Amor 2, Jan Peters 1,3 and Guilherme Maeda 1 Abstract Robots

More information

An Interactive Technique for Robot Control by Using Image Processing Method

An Interactive Technique for Robot Control by Using Image Processing Method An Interactive Technique for Robot Control by Using Image Processing Method Mr. Raskar D. S 1., Prof. Mrs. Belagali P. P 2 1, E&TC Dept. Dr. JJMCOE., Jaysingpur. Maharashtra., India. 2 Associate Prof.

More information

Reinforcement Learning of Clothing Assistance with a Dual-arm Robot

Reinforcement Learning of Clothing Assistance with a Dual-arm Robot Reinforcement Learning of Clothing Assistance with a Dual-arm Robot Tomoya Tamei, Takamitsu Matsubara, Akshara Rai, Tomohiro Shibata Graduate School of Information Science,Nara Institute of Science and

More information

Design & Kinematic Analysis of an Articulated Robotic Manipulator

Design & Kinematic Analysis of an Articulated Robotic Manipulator Design & Kinematic Analysis of an Articulated Robotic Manipulator Elias Eliot 1, B.B.V.L. Deepak 1*, D.R. Parhi 2, and J. Srinivas 2 1 Department of Industrial Design, National Institute of Technology-Rourkela

More information

Supplementary Figure 1. Decoding results broken down for different ROIs

Supplementary Figure 1. Decoding results broken down for different ROIs Supplementary Figure 1 Decoding results broken down for different ROIs Decoding results for areas V1, V2, V3, and V1 V3 combined. (a) Decoded and presented orientations are strongly correlated in areas

More information

ROBOT programming by demonstration (PbD) refers to

ROBOT programming by demonstration (PbD) refers to IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART B: CYBERNETICS, VOL. 42, NO. 4, AUGUST 2012 1039 Trajectory Learning for Robot Programming by Demonstration Using Hidden Markov Model and Dynamic

More information

TRAINING A ROBOTIC MANIPULATOR

TRAINING A ROBOTIC MANIPULATOR ME 4773/5493 Fundamental of Robotics Fall 2016 San Antonio, TX, USA TRAINING A ROBOTIC MANIPULATOR Jonathan Sackett Dept. of Mechanical Engineering San Antonio, TX, USA 78249 jonathan.sackett@utsa.edu

More information

Arm Movement Recorder

Arm Movement Recorder Energy Research Journal 1 (2): 126-130, 2010 ISSN 1949-0151 2010 Science Publications Arm Movement Recorder Jakkrapun Chuanasa and Szathys Songschon Department of Mechanical Engineering, King Mongkut s

More information

A Two-stage Scheme for Dynamic Hand Gesture Recognition

A Two-stage Scheme for Dynamic Hand Gesture Recognition A Two-stage Scheme for Dynamic Hand Gesture Recognition James P. Mammen, Subhasis Chaudhuri and Tushar Agrawal (james,sc,tush)@ee.iitb.ac.in Department of Electrical Engg. Indian Institute of Technology,

More information

Flexible Modeling and Simulation Architecture for Haptic Control of Maritime Cranes and Robotic Arms

Flexible Modeling and Simulation Architecture for Haptic Control of Maritime Cranes and Robotic Arms Flexible Modeling and Simulation Architecture for Haptic Control of Maritime Cranes and Robotic Arms F. Sanfilippo, H. P. Hildre, V. Æsøy and H.X. Zhang Department of Maritime Technology and Operation

More information

Handwriting Trajectory Movements Controlled by a Bêta-Elliptic Model

Handwriting Trajectory Movements Controlled by a Bêta-Elliptic Model Handwriting Trajectory Movements Controlled by a Bêta-Elliptic Model Hala BEZINE, Adel M. ALIMI, and Nabil DERBEL Research Group on Intelligent Machines (REGIM) Laboratory of Intelligent Control and Optimization

More information

Learning to Control Planar Hitting Motions in a Minigolf-like Task

Learning to Control Planar Hitting Motions in a Minigolf-like Task Learning to Control Planar Hitting Motions in a Minigolf-like Task Klas Kronander, Mohammad S.M. Khansari-Zadeh and Aude Billard Learning Algorithms and Systems Laboratory (LASA) Ecole Polytechnique Federale

More information

10-701/15-781, Fall 2006, Final

10-701/15-781, Fall 2006, Final -7/-78, Fall 6, Final Dec, :pm-8:pm There are 9 questions in this exam ( pages including this cover sheet). If you need more room to work out your answer to a question, use the back of the page and clearly

More information

Machine Learning A W 1sst KU. b) [1 P] For the probability distribution P (A, B, C, D) with the factorization

Machine Learning A W 1sst KU. b) [1 P] For the probability distribution P (A, B, C, D) with the factorization Machine Learning A 708.064 13W 1sst KU Exercises Problems marked with * are optional. 1 Conditional Independence a) [1 P] For the probability distribution P (A, B, C, D) with the factorization P (A, B,

More information

Robot learning for ball bouncing

Robot learning for ball bouncing Robot learning for ball bouncing Denny Dittmar Denny.Dittmar@stud.tu-darmstadt.de Bernhard Koch Bernhard.Koch@stud.tu-darmstadt.de Abstract For robots automatically learning to solve a given task is still

More information

SCREW-BASED RELATIVE JACOBIAN FOR MANIPULATORS COOPERATING IN A TASK

SCREW-BASED RELATIVE JACOBIAN FOR MANIPULATORS COOPERATING IN A TASK ABCM Symposium Series in Mechatronics - Vol. 3 - pp.276-285 Copyright c 2008 by ABCM SCREW-BASED RELATIVE JACOBIAN FOR MANIPULATORS COOPERATING IN A TASK Luiz Ribeiro, ribeiro@ime.eb.br Raul Guenther,

More information

Matlab Simulator of a 6 DOF Stanford Manipulator and its Validation Using Analytical Method and Roboanalyzer

Matlab Simulator of a 6 DOF Stanford Manipulator and its Validation Using Analytical Method and Roboanalyzer Matlab Simulator of a 6 DOF Stanford Manipulator and its Validation Using Analytical Method and Roboanalyzer Maitreyi More 1, Rahul Abande 2, Ankita Dadas 3, Santosh Joshi 4 1, 2, 3 Department of Mechanical

More information

An Interactive Software Environment for Gait Generation and Control Design of Sony Legged Robots

An Interactive Software Environment for Gait Generation and Control Design of Sony Legged Robots An Interactive Software Environment for Gait Generation and Control Design of Sony Legged Robots Dragos Golubovic and Huosheng Hu Department of Computer Science, University of Essex, Colchester CO4 3SQ,

More information

Force control of redundant industrial robots with an approach for singularity avoidance using extended task space formulation (ETSF)

Force control of redundant industrial robots with an approach for singularity avoidance using extended task space formulation (ETSF) Force control of redundant industrial robots with an approach for singularity avoidance using extended task space formulation (ETSF) MSc Audun Rønning Sanderud*, MSc Fredrik Reme**, Prof. Trygve Thomessen***

More information

FREE SINGULARITY PATH PLANNING OF HYBRID PARALLEL ROBOT

FREE SINGULARITY PATH PLANNING OF HYBRID PARALLEL ROBOT Proceedings of the 11 th International Conference on Manufacturing Research (ICMR2013), Cranfield University, UK, 19th 20th September 2013, pp 313-318 FREE SINGULARITY PATH PLANNING OF HYBRID PARALLEL

More information

Unsupervised Learning

Unsupervised Learning Unsupervised Learning Learning without Class Labels (or correct outputs) Density Estimation Learn P(X) given training data for X Clustering Partition data into clusters Dimensionality Reduction Discover

More information

Robotics. SAAST Robotics Robot Arms

Robotics. SAAST Robotics Robot Arms SAAST Robotics 008 Robot Arms Vijay Kumar Professor of Mechanical Engineering and Applied Mechanics and Professor of Computer and Information Science University of Pennsylvania Topics Types of robot arms

More information

Inference and Representation

Inference and Representation Inference and Representation Rachel Hodos New York University Lecture 5, October 6, 2015 Rachel Hodos Lecture 5: Inference and Representation Today: Learning with hidden variables Outline: Unsupervised

More information

Learning Reactive Robot Behavior for Autonomous Valve Turning

Learning Reactive Robot Behavior for Autonomous Valve Turning 24 4th IEEE-RAS International Conference on Humanoid Robots (Humanoids) November 8-2, 24. Madrid, Spain Learning Reactive Robot Behavior for Autonomous Valve Turning Seyed Reza Ahmadzadeh, Petar Kormushev,

More information

Learning Motion Skills from Expert Demonstrations and Own Experience using Gaussian Process Regression

Learning Motion Skills from Expert Demonstrations and Own Experience using Gaussian Process Regression ISR / ROBOTIK 21 Learning Motion Skills from Expert Demonstrations and Own Experience using Gaussian Process Regression Kathrin Gräve, Jörg Stückler, and Sven Behnke Autonomous Intelligent Systems Group,

More information

Trajectory Reconstruction with NURBS Curves for Robot Programming by Demonstration

Trajectory Reconstruction with NURBS Curves for Robot Programming by Demonstration Trajectory Reconstruction with NURBS Curves for Robot Programming by Demonstration Jacopo Aleotti, Stefano Caselli, Giuliano Maccherozzi RIMLab - Robotics and Intelligent Machines Laboratory Dipartimento

More information

arxiv: v2 [cs.ro] 17 Sep 2018

arxiv: v2 [cs.ro] 17 Sep 2018 Imitating Human Search Strategies for Assembly Dennis Ehlers, Markku Suomalainen, Jens Lundell and Ville Kyrki arxiv:189.486v2 [cs.ro] 17 Sep 218 Abstract We present a Learning from Demonstration method

More information

ECE276A: Sensing & Estimation in Robotics Lecture 11: Simultaneous Localization and Mapping using a Particle Filter

ECE276A: Sensing & Estimation in Robotics Lecture 11: Simultaneous Localization and Mapping using a Particle Filter ECE276A: Sensing & Estimation in Robotics Lecture 11: Simultaneous Localization and Mapping using a Particle Filter Lecturer: Nikolay Atanasov: natanasov@ucsd.edu Teaching Assistants: Siwei Guo: s9guo@eng.ucsd.edu

More information

Variable-resolution Velocity Roadmap Generation Considering Safety Constraints for Mobile Robots

Variable-resolution Velocity Roadmap Generation Considering Safety Constraints for Mobile Robots Variable-resolution Velocity Roadmap Generation Considering Safety Constraints for Mobile Robots Jingyu Xiang, Yuichi Tazaki, Tatsuya Suzuki and B. Levedahl Abstract This research develops a new roadmap

More information

Weakly supervised segmentation and

Weakly supervised segmentation and Weakly supervised segmentation and Firma convenzione Politecnico recognition di Milano of surgical e Veneranda gestures Fabbrica del Duomo di Milano Tutor: Elena De Momi, PhD Co-tutor: Danail Stoyanov,

More information

Monocular Human Motion Capture with a Mixture of Regressors. Ankur Agarwal and Bill Triggs GRAVIR-INRIA-CNRS, Grenoble, France

Monocular Human Motion Capture with a Mixture of Regressors. Ankur Agarwal and Bill Triggs GRAVIR-INRIA-CNRS, Grenoble, France Monocular Human Motion Capture with a Mixture of Regressors Ankur Agarwal and Bill Triggs GRAVIR-INRIA-CNRS, Grenoble, France IEEE Workshop on Vision for Human-Computer Interaction, 21 June 2005 Visual

More information

Development of a Master Slave System with Force Sensing Using Pneumatic Servo System for Laparoscopic Surgery

Development of a Master Slave System with Force Sensing Using Pneumatic Servo System for Laparoscopic Surgery 2007 IEEE International Conference on Robotics and Automation Roma, Italy, 10-14 April 2007 WeC8.3 Development of a Master Slave System with Force Sensing Using Pneumatic Servo System for Laparoscopic

More information

Force-Control for the Automated Footwear Testing System

Force-Control for the Automated Footwear Testing System Chapter 2 Force-Control for the Automated Footwear Testing System Problem presented by: Dr. Gerald K. Cole (Biomechanigg Research Inc.) Mentors: Greg Lewis (University of Ontario IT), Rex Westbrook (University

More information

FACIAL MOVEMENT BASED PERSON AUTHENTICATION

FACIAL MOVEMENT BASED PERSON AUTHENTICATION FACIAL MOVEMENT BASED PERSON AUTHENTICATION Pengqing Xie Yang Liu (Presenter) Yong Guan Iowa State University Department of Electrical and Computer Engineering OUTLINE Introduction Literature Review Methodology

More information

Singularity Handling on Puma in Operational Space Formulation

Singularity Handling on Puma in Operational Space Formulation Singularity Handling on Puma in Operational Space Formulation Denny Oetomo, Marcelo Ang Jr. National University of Singapore Singapore d oetomo@yahoo.com mpeangh@nus.edu.sg Ser Yong Lim Gintic Institute

More information

Dynamics Analysis for a 3-PRS Spatial Parallel Manipulator-Wearable Haptic Thimble

Dynamics Analysis for a 3-PRS Spatial Parallel Manipulator-Wearable Haptic Thimble Dynamics Analysis for a 3-PRS Spatial Parallel Manipulator-Wearable Haptic Thimble Masoud Moeini, University of Hamburg, Oct 216 [Wearable Haptic Thimble,A Developing Guide and Tutorial,Francesco Chinello]

More information

Machine Learning A WS15/16 1sst KU Version: January 11, b) [1 P] For the probability distribution P (A, B, C, D) with the factorization

Machine Learning A WS15/16 1sst KU Version: January 11, b) [1 P] For the probability distribution P (A, B, C, D) with the factorization Machine Learning A 708.064 WS15/16 1sst KU Version: January 11, 2016 Exercises Problems marked with * are optional. 1 Conditional Independence I [3 P] a) [1 P] For the probability distribution P (A, B,

More information

Simulation-Based Design of Robotic Systems

Simulation-Based Design of Robotic Systems Simulation-Based Design of Robotic Systems Shadi Mohammad Munshi* & Erik Van Voorthuysen School of Mechanical and Manufacturing Engineering, The University of New South Wales, Sydney, NSW 2052 shadimunshi@hotmail.com,

More information

A Non-Linear Manifold Alignment Approach to Robot Learning from Demonstrations

A Non-Linear Manifold Alignment Approach to Robot Learning from Demonstrations Paper: from Demonstrations Ndivhuwo Makondo 1, 2, Michihisa Hiratsuka 3, Benjamin Rosman 2, 4, and Osamu Hasegawa 5 1 Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology

More information

Fault Tolerant Supervisory Control of Human Interactive Robots

Fault Tolerant Supervisory Control of Human Interactive Robots Fault Tolerant Supervisory Control of Human Interactive Robots H.-B. Kuntze, C.-W. Frey, K. Giesen, G. Milighetti IITB für Karlsruhe, Germany Kuntze@iitb.fraunhofer.de page 1 Content 1 Introduction 2 Supervisory

More information

Master Thesis. Robotic Calligraphy: Learning From Character Images. Omair Ali. Dortmund, October 06, 2015

Master Thesis. Robotic Calligraphy: Learning From Character Images. Omair Ali. Dortmund, October 06, 2015 TU Dortmund University and TU München University Computational Intelligence Research Group Dynamic Human Robot Interaction Group Prof. Dr. Günter Rudolph Prof. Dongheui Lee Master Thesis Robotic Calligraphy:

More information

On Learning the Statistical Representation of a Task and Generalizing it to Various Contexts

On Learning the Statistical Representation of a Task and Generalizing it to Various Contexts On Learning the Statistical epresentation of a Task and Generalizing it to Various Contexts Sylvain Calinon, Florent Guenter and Aude Billard LASA Laboratory, School of Engineering (EPFL), C-5 Lausanne,

More information

Introduction to Trajectory Clustering. By YONGLI ZHANG

Introduction to Trajectory Clustering. By YONGLI ZHANG Introduction to Trajectory Clustering By YONGLI ZHANG Outline 1. Problem Definition 2. Clustering Methods for Trajectory data 3. Model-based Trajectory Clustering 4. Applications 5. Conclusions 1 Problem

More information

Inverse Kinematics Analysis for Manipulator Robot With Wrist Offset Based On the Closed-Form Algorithm

Inverse Kinematics Analysis for Manipulator Robot With Wrist Offset Based On the Closed-Form Algorithm Inverse Kinematics Analysis for Manipulator Robot With Wrist Offset Based On the Closed-Form Algorithm Mohammed Z. Al-Faiz,MIEEE Computer Engineering Dept. Nahrain University Baghdad, Iraq Mohammed S.Saleh

More information

Lighting- and Occlusion-robust View-based Teaching/Playback for Model-free Robot Programming

Lighting- and Occlusion-robust View-based Teaching/Playback for Model-free Robot Programming Lighting- and Occlusion-robust View-based Teaching/Playback for Model-free Robot Programming *Yusuke MAEDA (Yokohama National University) Yoshito SAITO (Ricoh Corp) Background Conventional Teaching/Playback

More information

Representability of Human Motions by Factorial Hidden Markov Models

Representability of Human Motions by Factorial Hidden Markov Models Proceedings of the 2007 IEEE/RSJ International Conference on Intelligent Robots and Systems San Diego, CA, USA, Oct 29 - Nov 2, 2007 WeC10.1 Representability of Human Motions by Factorial Hidden Markov

More information

University of Applied Sciences. Hochschule Ravensburg-Weingarten Fakultät Elektrotechnik und Informatik. Master Thesis

University of Applied Sciences. Hochschule Ravensburg-Weingarten Fakultät Elektrotechnik und Informatik. Master Thesis University of Applied Sciences Hochschule Ravensburg-Weingarten Fakultät Elektrotechnik und Informatik Master Thesis Adaptive Learning from Demonstration using Dynamic Movement Primitives August 2012 Author:

More information

9.913 Pattern Recognition for Vision. Class I - Overview. Instructors: B. Heisele, Y. Ivanov, T. Poggio

9.913 Pattern Recognition for Vision. Class I - Overview. Instructors: B. Heisele, Y. Ivanov, T. Poggio 9.913 Class I - Overview Instructors: B. Heisele, Y. Ivanov, T. Poggio TOC Administrivia Problems of Computer Vision and Pattern Recognition Overview of classes Quick review of Matlab Administrivia Instructors:

More information

Learning and Generalization of Motor Skills by Learning from Demonstration

Learning and Generalization of Motor Skills by Learning from Demonstration Learning and Generalization of Motor Skills by Learning from Demonstration Peter Pastor, Heiko Hoffmann, Tamim Asfour, and Stefan Schaal Abstract We provide a general approach for learning robotic motor

More information

DEVELOPMENT OF TELE-ROBOTIC INTERFACE SYSTEM FOR THE HOT-LINE MAINTENANCE. Chang-Hyun Kim, Min-Soeng Kim, Ju-Jang Lee,1

DEVELOPMENT OF TELE-ROBOTIC INTERFACE SYSTEM FOR THE HOT-LINE MAINTENANCE. Chang-Hyun Kim, Min-Soeng Kim, Ju-Jang Lee,1 DEVELOPMENT OF TELE-ROBOTIC INTERFACE SYSTEM FOR THE HOT-LINE MAINTENANCE Chang-Hyun Kim, Min-Soeng Kim, Ju-Jang Lee,1 Dept. of Electrical Engineering and Computer Science Korea Advanced Institute of Science

More information

A Portable 3-Dimensional Digitizer for Computer Graphics 3D Modeling

A Portable 3-Dimensional Digitizer for Computer Graphics 3D Modeling A Portable 3-Dimensional Digitizer for Computer Graphics 3D Modeling SURIYONG LERTKULVANICH 1 AND NITHI BURANAJANT 2 1 Computer and Information Technology Department 2 Production Engineering Department

More information

Short Survey on Static Hand Gesture Recognition

Short Survey on Static Hand Gesture Recognition Short Survey on Static Hand Gesture Recognition Huu-Hung Huynh University of Science and Technology The University of Danang, Vietnam Duc-Hoang Vo University of Science and Technology The University of

More information

Robotics. Lecture 5: Monte Carlo Localisation. See course website for up to date information.

Robotics. Lecture 5: Monte Carlo Localisation. See course website  for up to date information. Robotics Lecture 5: Monte Carlo Localisation See course website http://www.doc.ic.ac.uk/~ajd/robotics/ for up to date information. Andrew Davison Department of Computing Imperial College London Review:

More information

Robotics/Perception II

Robotics/Perception II Robotics/Perception II Artificial Intelligence and Integrated Computer Systems Division (AIICS) Outline Sensors - summary Computer systems Robotic architectures Mapping and Localization Motion planning

More information

Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information

Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information Ana González, Marcos Ortega Hortas, and Manuel G. Penedo University of A Coruña, VARPA group, A Coruña 15071,

More information

Human skill capturing and modelling using wearable devices

Human skill capturing and modelling using wearable devices Loughborough University Institutional Repository Human skill capturing and modelling using wearable devices This item was submitted to Loughborough University's Institutional Repository by the/an author.

More information

Design and building motion capture system using transducer Microsoft kinect to control robot humanoid

Design and building motion capture system using transducer Microsoft kinect to control robot humanoid Design and building motion capture system using transducer Microsoft kinect to control robot humanoid Gun Gun Maulana 1 *, Yuliadi Erdani 1, Aris Budiyarto 1, and Wahyudi Purnomo 1 1 Politeknik Manufaktur

More information

Construction Tele-robot System With Virtual Reality

Construction Tele-robot System With Virtual Reality Construction Tele-robot System With Virtual Reality Hironao Yamada Department of Mechanical & Systems Engineering Gifu University Gifu, Japan yamada@gifu-u.ac.jp Ni Tao Virtual System Laboratory Gifu University

More information

Normalized Texture Motifs and Their Application to Statistical Object Modeling

Normalized Texture Motifs and Their Application to Statistical Object Modeling Normalized Texture Motifs and Their Application to Statistical Obect Modeling S. D. Newsam B. S. Manunath Center for Applied Scientific Computing Electrical and Computer Engineering Lawrence Livermore

More information

Research Subject. Dynamics Computation and Behavior Capture of Human Figures (Nakamura Group)

Research Subject. Dynamics Computation and Behavior Capture of Human Figures (Nakamura Group) Research Subject Dynamics Computation and Behavior Capture of Human Figures (Nakamura Group) (1) Goal and summary Introduction Humanoid has less actuators than its movable degrees of freedom (DOF) which

More information

Note Set 4: Finite Mixture Models and the EM Algorithm

Note Set 4: Finite Mixture Models and the EM Algorithm Note Set 4: Finite Mixture Models and the EM Algorithm Padhraic Smyth, Department of Computer Science University of California, Irvine Finite Mixture Models A finite mixture model with K components, for

More information

Application of Principal Components Analysis and Gaussian Mixture Models to Printer Identification

Application of Principal Components Analysis and Gaussian Mixture Models to Printer Identification Application of Principal Components Analysis and Gaussian Mixture Models to Printer Identification Gazi. Ali, Pei-Ju Chiang Aravind K. Mikkilineni, George T. Chiu Edward J. Delp, and Jan P. Allebach School

More information

Inverse Kinematics of 6 DOF Serial Manipulator. Robotics. Inverse Kinematics of 6 DOF Serial Manipulator

Inverse Kinematics of 6 DOF Serial Manipulator. Robotics. Inverse Kinematics of 6 DOF Serial Manipulator Inverse Kinematics of 6 DOF Serial Manipulator Robotics Inverse Kinematics of 6 DOF Serial Manipulator Vladimír Smutný Center for Machine Perception Czech Institute for Informatics, Robotics, and Cybernetics

More information

Motion Interpretation and Synthesis by ICA

Motion Interpretation and Synthesis by ICA Motion Interpretation and Synthesis by ICA Renqiang Min Department of Computer Science, University of Toronto, 1 King s College Road, Toronto, ON M5S3G4, Canada Abstract. It is known that high-dimensional

More information

Inverse KKT Motion Optimization: A Newton Method to Efficiently Extract Task Spaces and Cost Parameters from Demonstrations

Inverse KKT Motion Optimization: A Newton Method to Efficiently Extract Task Spaces and Cost Parameters from Demonstrations Inverse KKT Motion Optimization: A Newton Method to Efficiently Extract Task Spaces and Cost Parameters from Demonstrations Peter Englert Machine Learning and Robotics Lab Universität Stuttgart Germany

More information