Skip to end of metadata
Go to start of metadata




This is the first workshop on MATLAB and Simulink for Robotics and Computer Vision education and research.  This will be a full day  (9:00-17:30) event held on Thursday June 5, 2014. test

Quick link to this page is


  • Peter Corke, School of Electrical Engineering and Computer Science, QUT, Brisbane
  • Yanliang Zhang, The MathWorks Inc., 3 Apple Hill Dr,  Natick, MA
  • Giampiero CampaThe MathWorks Inc., 3 Apple Hill Dr,  Natick, MA



MATLAB is a powerful environment for prototyping algorithms that provides a powerful object-oriented  programming  language, rich  graphics, portability across platforms and an effective code development environment. It  also supports interactive computation using the interactive interpreter, which also allows for incremental software development. MATLAB supports a rich ecosystem of Toolboxes which extend the basic environment for application specific domains such as control or signal processing. MATLAB is also widely used in undergraduate education around the world, particularly for engineering.

Simulink is a block diagram environment for multi-domain simulation.  It supports simulation, automatic code generation, and continuous  test and verification of embedded systems.  Simulink provides a graphical editor, customizable block libraries, and solvers for modeling and simulating dynamic systems. It is integrated with MATLAB, enabling the user to incorporate MATLAB algorithms into models and export simulation results to MATLAB for further analysis.

There is a long  history of MATLAB/Simulink use in robotics, and many contemporary texts (Craig 2004, Siciliano 2010 and Corke 2011). MATLAB/Simulink forms the basis of tutorials and laboratories for many robotics courses around the world.   In more recent times there has been interest in using MATLAB/Simulink for robotics problems beyond the classical arm-manipulator kinematics, dynamics and control - for example computer vision, path planning and connecting the MATLAB environment to real-world robots.

Supported by the RAS Technical Committee on  Education.


The objectives of the workshop are to demonstrate, and share knowledge about, how MATLAB/Simulink is used for:

  • Teaching fundamental principles such as   robot   manipulator   kinematics, dynamics and control; and mobile robot control, localization and planning.
  • Developing sensor processing algorithms such as SLAM, face detection, etc.,  for robotics applications
  • Deploying algorithms to popular platforms such as Raspberry Pi-based robot
  • Connecting MATLAB to real robots

Intended audience

The intended  audience  is a broad  cross  section  of those  attending  ICRA 2014  and would  include  academics  involved  in  teaching  robotics  at  the  undergraduate  and graduate level, as well as graduate research students interested in effective tools to facilitate their research.


We have a great and diverse program with the following confirmed participants.

 Time Activity / TopicSpeaker/PresenterPresentation File 
09:00 - 09:20Introduction to Workshop   
09:20 - 09:40Teaching Forward and Inverse Kinematics of Robotic Manipulators Via MATLABDenise Wong, Philip Dames (Penn)TeachingFKIKofRoboticManipulatorsviaMatlab.pdf 
09:40 - 10:00Robotics Toolbox Code Generation ModuleJörn Malzahn (TU Dortmund)2014_ICRA_Malzahn_MATLAB_WS.pdf
10:00 - 10:20 Teaching robotics and vision with MATLAB, Interface to Lego EV3Peter Corke (QUT)

Corke introduction.pdf

Corke EV3.pdf

10:20 - 10:40Coffee Break  
10:40 - 11:00 Real-time sensor data analytics educationPramod Abichandani etal.MatlabArduino ICRA 2014 Mathworks Workshop.pdf
11:00 - 11:20  Interfacing MATLAB/Simulink with V-REP for an Easy Development of
Sensor-Based Control Algorithms for Robotic Platforms
Paolo Robuffo Giordano etal.Spica_Interfacing_Matlab_and_V-Rep .pdf 
11:20 - 11:40 An Open-source Recipe for Teaching (and Learning) Robotics with a SimulatorRenaud Detry (U Liege)Renaud_Detry_TRS.pdf
11:40 - 12:00 A Simulink Library for the Natural Motion Initiative in Robotics Kamilo Melo 
12:00 - 13:00Lunch Break  
13:00 - 13:20 What is the Best Way to implement algorithms in Simulink?Giampiero Campa (Mathworks)What is the best way to implement my algorithm in Simulink.pdf 
13:20 - 13:40 Interfacing MATLAB and ROSYanliang Zhang (Mathworks)Interfacing MATLAB and ROS.pdf 
13:40 - 14:00 WBI Toolbox (WBI-T): A Simulink Wrapper for Robot Whole Body Control Jorhabib Eljaik Gomez etal.WBI Toolbox Presentation.pdf 
14:00 - 14:20 Hapkit: Open Hardware and Software for Haptics Integrated with MATLABAllison Okamuraokamura-MatlabHapkit.pdf
14:20 - 14:40  Development of the Flight Control Computer Using Embedded Coder of
Hyon Lim etal. 
14:40 - 15:00 Robot Control Using MATLAB/SIMULINKLeon ŽlajpahICRA_2014_Zlajpah.pdf
15:00 - 15:30 Coffee Break  
15:30 - 15:50  Drake: A planning, control, and analysis toolbox for nonlinear dynamical systemsRuss Tedrake and Scott Kuindersmadrake.pdf
15:50 - 16:10 SLAM, SAM, and SFM at Your Fingertips:GTSAM from MATLABVadim Indelman and Frank Dellaert2014-06-05-GTSAM-ICRA-ws.pdf
16:10 - 16:30SynGrasp: a MATLAB Toolbox for Human and Robot Hands with Synergies, Underactuation and ComplianceDomenico Prattichizzo (U Sienna) 
16:30 - 16:50The Surface Patch Library (SPL)Dimitrios Kanoulas and Marsette VonaKanoulas_Vona__2014__The_Surface_Patch_Library.pdf
16:50 - 17:10The MuJoCo physics engine, with applicationsYuval TassatassaICRA14.pdf 
17:10 - 17:30Wrap-up  

Title: Teaching Forward and Inverse Kinematics of Robotic Manipulators Via MATLAB

Denise Wong, Philip Dames, and Katherine J. Kuchenbecker

Bio: Denise Wong received the B.S. and M.Eng. degree in mechanical engineering from Cornell University, Ithaca, NY, in 2009.  She is currently working towards the Ph.D. degree in mechanical engineering and applied mechanics at the University of Pennsylvania under the supervision of Dr. Vijay Kumar.  Her research focuses are in micro bio robotics - combining cell biology for sensing and actuation of microscale robots.

Bio: Philip Dames received his B.S. and M.S. degrees in Mechanical Engineering from Northwestern University in 2010.  Currently he is a Ph.D. candidate in the department of Mechanical Engineering and Applied Mechanics at the University of Pennsylvania working under the supervision of Prof. Vijay Kumar.  Philip works on information-based control strategies for small teams of robots to cooperatively seek out and localize multiple sources of information in an environment.

Philip and Denise were teaching assistants for MEAM 520 in the fall semester of 2012 under the supervision of Prof. Katherine Kuchenbecker.

Abstract: Introduction to Robotics (MEAM 520) at the University of Pennsylvania is a graduate-level course on robotic manipulators currently taught by Professor Katherine Kuchenbecker.  We present three MATLAB projects that give students the opportunity to apply their knowledge to real robotic systems.  The first two projects use a PUMA 260 manipulator with a spherical wrist.  We created a MATLAB simulator, based on the Robotics Toolbox, that allows students to run the same code on either a virtual version of the robot or the real robot in the lab.  In the robotic dancing project, students generated joint-space trajectories to choreograph thirty-second-long dances performed in time with music.  In the robotic light painting project, the PUMA end-effector was fitted with a tri-color LED, and a long-exposure camera system captured the movement of the robot in the dark.  Each team solved the full inverse kinematics of the PUMA and used this to draw an image of their choice.  The third project used a SensAble Phantom Premium to let the students feel several haptic virtual environments created in MATLAB: a box, a sphere, and a damping field.  This presentation includes many examples of student work as well as real-time MATLAB demonstrations of the PUMA simulator.


Title: Robotics Toolbox Code Generation Module

Jörn Malzahn (TU Dortmund)

Bio: TBA

Abstract: The open source Robotics Toolbox for MATLAB by Peter Corke is looking back on a history of almost two decades in providing generic algorithms for the numerical simulation and visualization of spatial transforms, the kinematics and dynamics of serial chain robot arms. This contribution presents the CodeGenerator module added in release 9.8. The module provides commands for the automated generation of symbolic model equations as well as robot specific source code. In education, the symbolic equations provide global structural insights into the arm models, where the generic numerical algorithms only provide output values for individual operating points. Thereby the module improves the value for students seeking to foster and check their understanding of kinematics and dynamics modeling of robot arms. In research, the symbolic equations serve as a basis for model based controllers, sensitivity analyses and system identification. The module relieves the researcher from the error prone and time consuming derivation. The automated code generation functionality comprises robot specific m-code, real-time compatible Simulink blocks, ready-to-use C- and MEX-functions. The C-functions can be interfaced within other software projects. The robot specific MEX-files substantially speed up execution times in MATLAB. The contribution presents application examples from education as well as research and includes a live demo at the workshop.


Title: Real-time Sensor Data Analytics Education

Pramod Abichandani, Vaishali Parikh, Christopher Berry, William Fligor

Bio: Pramod Abichandani serves as the Director of the Second Year Engineering Curriculum at Drexel University. He is a Senior Researcher and an Assistant Teaching Professor at the College of Engineering at Drexel University. He received his Bachelors of Engineering (B.E.) degree in 2005 from Nirma Institute of Technology, Gujarat University, India, and his M.S. and Ph.D. degrees in Electrical and Computer Engineering from Drexel University in 2007 and 2011 respectively. His research interests are centered around optimal, multi-dimensional, data-driven decision-making, through the use of techniques from mathematical programming, linear and nonlinear systems theory, statistics, and machine learning. Specific technical areas include optimal decision-making for multi-robot systems under communication constraints, sensor fusion for naval ship monitoring, cognitive bias effects in handwriting forensics, and embedded systems design for data acquisition and control. Sponsors of his research include the National Science Foundation (NSF), Office of Naval Research (ONR), Mathworks, Drexel ExCITe Center, and Drexel University's College of Engineering.

Abstract: In this talk, we report on the use of this website to create blended learning strategies and flipped classrooms for over 2000 Drexel University students in 2013-2014. The website is being used in an introductory computational problem solving and programming course where 1200+ first year engineering students are asked to utilize MATLAB to perform mathematical modeling and simulations with 2D and 3D animations. Additionally, the website forms the back-bone of an engineering data analytics class for 800+ second year engineering students where they use a 3-axis accelerometer interfaced through the Arduino board to create MATLAB Apps. These applications feature the ability to procure the 3-axis accelerometer data in real-time, filter noise out the data, and apply appropriate signal processing techniques to perform data-driven decision-making and create feature rich 2D and 3D visualizations. Using the Arduino and Accelerometer hardware, we will demonstrate Apps that students have designed in this class.


Title: Interfacing Matlab/Simulink with V-REP for an Easy Development of Sensor-Based Control Algorithms for Robotic Platforms

Paolo Robuffo Giordano, Riccardo Spica, Giovanni Claudio, Fabien Spindler


Bio: Paolo Robuffo Giordano TBA

Bio: Riccardo Spica was born in Rome, Italy in 1987. He received the M.Sc. degree in Electronic Engineering from "La Sapienza" - Università di Roma, Rome, Italy in 2012. He worked first as a Master's Student and later as a Graduate Research Assistant at the Max Planck Institute for Biological Cybernetics in Tuebingen, Germany for one year between 2012 and 2013. In December 2013 he started a Ph.D. in Signal Processing at the Univérsité de Rennes I, Rennes, France within the Lagadic group of Irisa/Inria. His research interests are in planning and control for robotics applications. In particular his current research is in visual servoing and active structure from motion.

Abstract: This presentation will focus on how to interface the matlab/simulink environment with V-REP using the ROS communication libraries (the publisher/subscriber paradigm) for fast prototyping of robot control algorithms. We will first show how to embed ROS nodes in simulink by including custom C S-Functions representing ROS topics to be listened/published. This will make it possible for Simulink to exchange data with V-REP in real-time for obtaining the robot data and computing the needed control actions. Then, we will demonstrate our architecture in two simulated scenarios: (i) visual control of a quadrotor UAV and (ii) visual control of an industrial manipulator. The first scenario will involve a quadrotor UAV equipped with a IMU and a down-looking camera meant to control its pose w.r.t. a ground target by means of a visual servoing law. The second scenario will consider the same situation for a fixed manipulator with an eye-in-hand camera performing a classical visual servoing task.


Title: An Open-source Recipe for Teaching (and Learning) Robotics with a Simulator

Renaud Detry (U Liege)

Bio: Renaud Detry is a senior researcher at the University of Liège, Belgium (Systems and Modeling Group), and a visiting researcher at KTH, Sweden (Computer Vision and Active Perception lab). He earned an engineering degree at the University of Liège in 2006, and a PhD on robot learning from the same university in 2010. From 2010 to 2012, he was a postdoctoral researcher with Danica Kragic at KTH, Sweden. His research interest are in robot grasping, machine learning, and computer vision.

Abstract: We present a cross-platform robot development and simulation environment that can be installed in five minutes and that allows students to write control, navigation, vision or manipulation algorithms in a hundred lines of Matlab or Python code. The environment relies on the V-REP robot simulator, and on the Matlab Robotics Toolbox (RTB). The key feature of this combination is its ease of use. This environment can be used by students to quickly start implementing and testing robot algorithms, and by teachers to organize a master-level robotics project.


A Simulink Library for the Natural Motion Initiative in Robotics

Kamilo Melo, Manolo Garabini, Giorgio Grioli, Manuel Catalano, Lorenzo Malagia and Antonio Bicchi

Bio: Dr. Kamilo Melo, received the B.S. degree in Electronics Engineering from the Pontificia Universidad Javeriana, Bogotá D.C., Colombia, in 2004. The M.Sc. degree in Mechanical Engineering from the Universidad de los Andes, Bogotá D.C., Colombia, in 2005 and the Ph.D. (Dr. Eng.) magna cum laude degree in Engineering (Robotics) from Pontificia Universidad Javeriana, Bogota D.C., Colombia, in 2013. He worked as a research intern under the supervision of Prof. Raja Chatila at ISIR (Institut des Systèmes Intelligents et de Robotique) in the Pierre et Marie Curie University UPMC, Paris, France, in 2012 and as a postdoctoral associate under supervision of Prof. Antonio Bicchi in the Centro di Ricerca "E. Piaggio" in the Universita di Pisa, Pisa, Italy, in 2013-2014. Dr. Melo is CEO/CTO in KM-RoBoTa s.a.s., a R+D+i institution in Bogotá D.C., Colombia, focused in the design and fabrication of Rescue Robot Systems, particularly Modular Snake Robots Platforms.

Abstract: The Natural Motion Initiative (NMI) is an open-source community aiming at the diffusion of technologies that will propel the next generation of robots. NMI fosters the use of new actuation technologies for robotics research and educative purposes. One of the main foci of NMI is on a new generation of actuators called Variable Stiffness Actuators (VSA). To contribute to NMI, we use the "qbmoves" as example of VSA technology. A Matlab/SImulink library has been developed for control the qbmoves, which includes a detailed Simulink model of the main components (mechanical, electronics, communication and control) and allows Matlab users to interconnect qbmoves in arbitrary structures and seamlessly interface with the real hardware through a USB connection. For this workshop, we want to share a case of use of this Matlab/Simulink library in the Master on robotics engineering and automation at the University of Pisa in the course Robot Control (Hardware in the Loop session). Moreover, as a further contribution to the workshop, we will demonstrate the use of the Simulink block to move and change the stiffness of a qbmove at runtime and show how this library is scalable to drive simultaneously a multi-dof robot (Robotic Snake). As an example, we will use the snake robot made out of qbmoves using Simulink to control its gaits.


What Is the Best Way to implement algorithms in Simulink?

Giampiero Campa

Bio: Dr Campa received both his Laurea degree in Electrical Engineering (1996) and his Ph.D. degree in Robotics and Automation (2000), from the University of Pisa, Italy. He has also worked at the Industrial Control Centre, Strathclyde University, UK, (1995) and at the Department of Aerospace Engineering, Georgia Institute of Technology, Atlanta, USA (1999). From 2000 to 2008 he has served as faculty in the Flight Control Group at the Department of Aerospace Engineering, West Virginia University. His research at WVU involved system identification, adaptive and nonlinear control, fault tolerant systems, machine vision, and sensor fusion, especially applied to UAVs (sponsors included NASA, AFOSR, ONR, and local commercial companies). During his stay at WVU Dr. Campa has published around 30 peer-reviewed articles in international journals, about 60 research papers for international conferences, and a couple of book chapters. Since January 2009 he works for MathWorks, where he is currently responsible for the educational sector for southern California as well as for devising strategic activities in the areas of robotics and mechatronics.


There are currently many possible ways of implementing algorithms in Simulink. Apart from the obvious possibility of creating a subsystem by assembling basic and lower-level Simulink blocks, (or Stateflow charts), there are also many options for creating Simulink blocks that rely on MATLAB and/or C code, such as hand-written S-Functions (both in MATLAB and C), the S-Function builder, the Legacy Code Tool, MATLAB Functions, and System Objects. This presentation will describe the main features, advantages and disadvantages of each method. Specifically, using the implementation of an Extended Kalman Filter (EKF) for attitude estimation using data from both GPS and IMU sensors as a case study, all the previously mentioned methods of creating Simulink blocks will then be compared, so that relative strengths and weaknesses can be highlighted. General guidelines will then be given to enable users to choose the method that better suits their need.


Title: MATLAB ROS Interface

Yanliang Zhang

Bio: Yanliang is robotics product and marketing manager at MathWorks. Before joining MathWorks, he did his post-doc in robotics at The Advanced Micro and Nanosystems Laboratory (AMNL) at University of Toronto. Yanliang received his Ph.D. in robotics from Nanyang Technological University, Singapore. Yanliang serves as a distinguished visiting professor at Institute of Soil Science, Chinese Academy of Science, and as an adjunct professor at Shanghai Normal University and Xiangtan University, China. Yanliang is also the sole creator of the most popular Chinese MATLAB forum ( with ~800,000 registered members.  As the sole founder, Yanliang started two spin-off business entities based on his research work in China and Singapore.

Abstract: With ROS support from MATLAB, you can interact with robots and simulators that provide a ROS interface. You can also create a self-contained ROS network directly in MATLAB. These features allow you to develop your robotics algorithms in MATLAB, while giving you the ability to exchange messages with other nodes on the ROS network. This support extends the rosjava API. It includes a new API for creating ROS nodes inside MATLAB based on the same ROS publisher/subscriber mechanism.

Key features allow you to:

Create ROS nodes, publishers, and subscribers directly from MATLAB

  • Create and send ROS messages from MATLAB
  • Enable publishers to publish MATLAB data to their advertised topics
  • Enable subscribers to execute arbitrary user-defined MATLAB functions
  • Enable launching of ROS masters on the local host from MATLAB

In this workshop, we will demonstrate to you how to use this interface.


Title: WBI Toolbox (WBI-T): A Simulink Wrapper for Robot Whole Body Control

Jorhabib Eljaik Gomez, Andrea del Prete, Silvio Traversaro, Marco Randazzo and Francesco Nori

Bio: Jorhabib Eljaik Gomez was born on November 1st, 1988 in Barranquilla, Colombia where he received his B.Sc. degree on Electronics Engineering from Universidad del Norte in 2009. Subsequently, he was awarded an Erasmus Mundus scholarship for the EMARO program (European Masters on Advanced Robotics), through which in 2011 he obtained a dual M.Sc. degree from Warsaw University of Technology (Poland) and Universit_a degli Studidi Genova (Italy). He is currently a Ph.D. candidate at the Italian Institute of Technology (IIT) in Robotics, Cognition and Interaction Technologies with the department of Robotics, Brain and Cognitive Sciences. His current research interests are stabilization strategies in humanoid robots with stiff and compliant actuators and whole-body control.

Abstract: This toolbox was born within the context of the European project CoDyCo (Whole Body Compliant Dynamical Contacts in Cognitive Humanoids), which deals with whole-body motion control of humanoid robots in multiple contacts with the environment. In order to validate the theoretical advances in real-world scenarios, we use the humanoid robotic platform iCub along with the Gazebo simulator. Both the real platform and the simulator use YARP as middleware allowing a seamless transfer of what is done in simulation to the real robot. Nonetheless, YARP is not an iCub-speci_c middleware and can in principle be used with any robotic platform to interface with its sensors, processors and actuators. This being said, our toolbox consists of Simulink blocks wrapping a YARP-based implementation of the Whole-Body Interface (WBI) C++ library. WBI acts as an abstraction layer for any interaction with the robot, making code robot-independent. Moreover, it promotes code reusability and exchange between different teams.


Title: Hapkit: Open Hardware and Software for Haptics Integrated with MATLAB

Allison M. Okamura

Bio: Allison M. Okamura received the BS degree from the University of California at Berkeley in 1994, and the MS and PhD degrees from Stanford University in 1996 and 2000, respectively, all in mechanical engineering. She is currently Associate Professor in the mechanical engineering department at Stanford University. She was previously Professor and Vice Chair of mechanical engineering at Johns Hopkins University. She has been an associate editor of the IEEE Transactions on Haptics, an editor of the IEEE International Conference on Robotics and Automation Conference Editorial Board, and co-chair of the IEEE Haptics Symposium. Her awards include the 2009 IEEE Technical Committee on Haptics Early Career Award, the 2005 IEEE Robotics and Automation Society Early Academic Career Award, and the 2004 NSF CAREER Award. She is an IEEE Fellow. Her academic interests include haptics, teleoperation, virtual environments and simulators, medical robotics, neuromechanics and rehabilitation, prosthetics, and engineering education. Outside academia, she enjoys spending time with her husband and two children, running, and playing ice hockey. For more information, please see the CHARM Lab website:

Abstract: We present a low-cost, open-hardware device, called Hapkit, designed to provide a hands-on experience in environments where traditional laboratories are challenging, particularly online courses and space- and funding-constrained schools. Hapkit is a one-degree-of-freedom kinesthetic haptic robot that allows users to input motions and feel programmed forces. The CAD files, parts list and costs, assembly instructions, and sample software are posted at so the public can both recreate Hapkit and modify its design. In Autumn 2013, we piloted Hapkit in a 5-week online haptics course with videos, quizzes, and hands-on labs. Each student received all the parts needed to assemble and program his or her own Hapkit. In Winter 2014, we used Hapkit to render controllers designed in MATLAB/Simulink in an introductory controls course. This was an engaging, inexpensive, and portable way to provide hands-on control laboratories. We will demonstrate Hapkit with MATLAB/Simulink.


Title: Development of the Flight Control Computer using Embedded Coder of MATLAB/Simulink

Hyon Lim, Chulwoo Park, and Kyunghyun Lee

Bio: Hyon Lim

Hyon Lim (born Oct 8, 1985) is a Ph.D. candidate in School of Mechanical and Aerospace Engineering, Seoul National University in Seoul, Korea ( His research includes aerial robotics, visual simultaneous localization and mapping (VSLAM) and control of dynamic systems.

Abstract: In this presentation, we consider an Embdded Coder which is a MATLAB/Simulink toolbox that generates C-code for our custom-designed flight controller (including Ti C2000 DSP) using existing Simulink model, for Unmanned Aerial Vehicles (UAVs) control. We demonstrate two types of vehicle: quadrotor and fixed-wing aircraft. The detailed methodology is explained step-by-step with flight demonstration.


Title: Robot Control using MATLAB/SIMULINK

Leon Žlajpah

Bio: Leon Žlajpah received the B.Sc. degree in automatics from the University of Ljubljana, Ljubljana, Slovenia, in 1982. From the same University he received in 1985 the M. Sc. degree and in 1989 the Ph.D. degree in electrical science. In 1981 he started as a researcher at the "Jožef Stefan" Institute in Ljubljana, working in different fields of robotics. Since 1998 he is a Senior Researcher at the "Jožef Stefan" Institute. Between 2005 and 2014 he was the Head of the Department for Automation, Biocybernetics and Robotics at the "Jožef Stefan" Institute. Basic research interests of L. Žlajpah include modelling and simulation of robot systems and advanced control systems for robots. His contributions are in the field of control of redundant robots, compliant robot control, sensor systems, applications of robots, and recently control of robots interacting with humans. For his research work in robotics and applications he was awarded several awards. Among them is the National award for research excellence in the field of robotics. Dr. Žlajpah is vice-president of Slovenian Society for Simulation and Modelling, member of Automatics Society of Slovenia and IEEE member.

Abstract: Over 20 years we have been using different simulation tools and many of them have been developed in our lab. In the last decade we have been using MATLAB/Simulink based integrated environment for the control design.The main capabilities supported by our simulation environment are: the simulation of the kinematics and dynamics of manipulators, the integration of different sensor systems, scenarios for complex robot tasks, the visualization of robots and their environment and the integration of real robots in the simulation loop. The advantage of our system is the simplicity which allows easy integration of different robots, sensors and other devices. Due to good experiences with MATLAB/Simulink product family and our Toolboxes, we use this environment as the principal tool for the design of robot control systems. We present special blocks for modeling the robot manipulators, interfaces, sensors and kinematic transformations, etc. The presented control design environment has proved to be a very useful and effective tool for fast and safe development and testing of advanced control schemes and task planning algorithms, including force control and visual feedback. The software can be very easily extended and adapted to different requirements and applied to any types of robotic manipulators.


Title: Drake: A Planning, Control, and Analysis Toolbox for Nonlinear Dynamical Systems

Russ Tedrake and Scott Kuindersma

Bio: Russ Tedrake TBA

Bio: Scott Kuindersma is a Postdoc in the Robot Locomotion Group at the Computer Science and Artificial Intelligence Lab at MIT. He completed his PhD in Computer Science at the University of Massachusetts Amherst in 2012. His mission is to make robots move fluidly and dynamically by advancing the state of the art in optimization-based control. He has had the pleasure of working with a variety of amazing robots including the UMass uBot, Robonaut 2, and Boston Dynamics' Atlas. Currently, he is the planning and control lead for the MIT DARPA Robotics Challenge team, where his research is focused on developing robust and efficient control algorithms for dynamic locomotion and whole-body manipulation.

Abstract: Drake is a toolbox written by the Robot Locomotion Group at MIT CSAIL. It is a collection of tools for analyzing the dynamics of our robots and building control systems for them in MATLAB. It deals with general nonlinear systems (including hybrid systems), but also contains specialized tools for multi-link rigid-body dynamics with contact. You might want to use Drake in your own research in order to analyze the stability of systems, design nonlinear feedback controllers for complicated dynamical systems, perform trajectory and feedback-motion planning, or compute invariant ``funnels'' along trajectories. Drake also contains supporting methods for visualization, identification, estimation, and even hardware interfaces. Drake is implemented using a hierarchy of MATLAB classes which are designed to expose and exploit available structure in input-output dynamical systems. While some algorithms are available for general nonlinear systems, specialized algorithms are available for polynomial dynamical systems, linear dynamical systems, etc.; many of those algorithms operate symbolically on the governing equations. The toolbox also provides a parser for URDF files that makes it easy to define and start working with rigid-body dynamical systems. It has been used by the Robot Locomotion Group, the MIT DARPA Robotics Challenge Team, and a handful of collaborators, but we are now attempting to open up the code to the broader community.


Title: SLAM, SAM, and SFM at Your Fingertips:GTSAM from MATLAB 

Vadim Indelman and Frank Dellaert

Bio: Dr. Vadim Indelman has been a postdoctoral fellow in the Institute of Robotics and Intelligent Machines (IRIM) at Georgia Institute of Technology since 2012. Starting from summer 2014 he will be assuming the position of an assistant professor in the faculty of Aerospace Engineering at the Technion – Israel Institute of Technology, where he also completed his PhD. Dr. Indelman received B.A. and B.Sc. degrees in Computer Science and Aerospace Engineering, respectively, from the Technion in 2002. His research interests include distributed multi-agent systems, SLAM, bundle adjustment and structure from motion, cooperative navigation and tracking, distributed perception, planning under uncertainty and graphical model based inference.

Bio: Frank Dellaert TBA

Abstract: This presentation will describe the Georgia Tech’s Smoothing and Mapping library, GTSAM, and demonstrate rapid algorithm prototyping via its Matlab interface. GTSAM uses factor graphs and Bayes networks to perform computationally efficient inference in large scale Structure from Motion (SfM) and Simultaneous Localization and Mapping (SLAM) problems, as well as other commonly encountered problems in robotics and computer vision. The objective of this tutorial is to introduce the basic components of GTSAM and to present the automatic Matlab wrapping capability, that makes GTSAM functionality directly accessible from Matlab, and can also be used to easily wrap user-implemented new C++ code to maintain computational efficiency. This Matlab interface will be demonstrated in three recently developed approaches: Multi-robot data association and localization from unknown initial poses; A unified estimation and control framework with factor graphs; and Incremental light bundle adjustment.


Title: SynGrasp: a MATLAB Toolbox for Human and Robot Hands with Synergies, Underactuation and Compliance

Domenico Prattichizzo

Bio: Domenico Prattichizzo received the M.S. degree in Electronics Engineering and the Ph.D. degree in Robotics and Automation from the University of Pisa in 1991 and 1995, respectively. Since 2002 Associate Professor of Robotics at the University of Siena. Since 2009 Scientific Consultant at Istituto Italiano di Tecnologia, Genova, Italy. In 1994, Visiting Scientist at the MIT AI Lab. Co-author of the Grasping chapter of Handbook of Robotics Springer, 2008, awarded with two PROSE Awards presented by the American Association of Publishers. Since 2007 Associate Editor in Chief of the IEEE Trans. on Haptics. From 2003 to 2007, Associate Editor of the IEEE Trans. on Robotics and IEEE Trans. on Control Systems Technologies. Vice-chair for Special Issues of the IEEE Technical Committee on Haptics (2006-2010). Chair of the Italian Chapter of the IEEE RAS (2006-2010), awarded with the IEEE 2009 Chapter of the Year Award.Co-editor of two books by STAR, Springer Tracks in Advanced Robotics, Springer (2003, 2005). Research interests are in haptics, grasping, visual servoing, mobile robotics and geometric control. Author of more than 200 papers in these fields.

Abstract: SynGrasp is a MATLAB toolbox developed for the analysis of grasping, suitable both for robotic and human hands. It includes functions for the definition of hand kinematic structure and of the contact points with a grasped object. The coupling between joints induced by an underactuated control can be modeled. The hand modeling allows to define compliance at the contact, joint and actuator levels. The provided analysis functions can be used to investigate the main grasp properties: controllable forces and object displacement, manipulability analysis, grasp quality measures. Functions for the graphical representation of the hand, the object and the main analysis results are provided.

The SynGrasp toolbox has been developed in the context of the EU Project “THE - The Hand Embodied” that aims to study how the embodied characteristics of the human hand and its sensors, the sensorimotor transformations, and the very constraints they impose, affect and determine the learning and control strategies we use for such fundamental cognitive functions as exploring, grasping and manipulating. For this reason, the toolbox provides several functions for human and robot grasping evaluation including specific functions for human hand synergies evaluation.


Title: The Surface Patch Library (SPL)

Dimitrios Kanoulas and Marsette Vona

Bio: Dimitrios Kanoulas is a Ph.D. student at the College of Computer and Information Science at Northeastern University, working with Prof. Marsette Vona in the Geometric and Physical Computing Lab.  His research focuses on 3D perception for robotics, in particular for uncertain natural environments.  Prior to that he was member of the Algorithms and Theory group at Northeastern University working on algorithmic aspects of Game Theory.

Abstract: The Surface Patch Library (SPL) is an open-source Matlab toolbox for prototyping 3D rough terrain perception algorithms.  It includes models of 10 types of curved surface patches, an algorithm to fit patches to potentially noisy range sensor data, and an algorithm to perceptually validate patches.  Uncertainty is quantified throughout using covariance matrices.  The patches are designed to model local contact regions both in the environment (e.g. on rocky surfaces) and on a robot (e.g. foot pads).

Three different live demos will be given.  The first demo starts with a set of 3D points sampled from a simulated elliptic paraboloid.  These data are taken with noise simulated according to a Kinect range error model.  A paraboloid patch with ellipse boundary is automatically fit by our algorithm.  Similarly we will show the results for all 10 patch types.  In the next two demos we will present both an interactive and an automatic segmentation of patches from real data taken with a Kinect, with each patch automatically fit by our algorithm.


Title: The MuJoCo Physics Engine, with Applications

Yuval Tassa

Bio: Yuval Tassa recived his PhD from the Hebrew University of Jerusalem in 2011. In the last several years hehas been a postdoc with Emo Todorov at the University of Washington, Seattle. In the summer of 2014 he will be joining Google Deepmind.

Abstract: I will describe the physics engine MuJoCo ( which stands for Multi-Joint dynamics with Contact. It aims to facilitate research and development in robotics, biomechanics, graphics and animation, and other areas where fast and accurate simulation of complex dynamical systems is needed. MuJoCo is available as a C/C++ library and also accessible from Matlab through a Mex package which provides access to the core features, demos and a tutorial.

Notable features of MuJoCo are:

  • Generalized coordinates combined with impulse contact dynamics
  • Convex, differentiable and analytically-invertible impulse dynamics
  • Tendon geometry
  • General actuation model
  • Reconfigurable computation pipeline
  • Informative visualization
  • Interactive simulation




  • No labels