Jean-Pierre MERLET

Senior scientist, emeritus

Previously head of the SAGA, COPRIN projects and HEPHAISTOS

References database on parallel robots

Address:
INRIA Sophia Antipolis, projet HEPHAISTOS
2004 Route des Lucioles, BP 93
06902 Sophia Antipolis Cedex, France

  • Tel: (33) 4 92 38 77 61
  • Fax: (33) 4 92 38 76 43
  • Email:Jean-Pierre.Merlet@sophia.inria.fr

    IEEE Fellow, IFToMM Award of Merits,member of the Executive Council of IFToMM (2012-2020), doctor honoris causae University Innsbruck, a short resume (to make it short, background: control theory (nowaday called artificial intelligence), mechanical engineering, mathematics and, on the side, computer science)

    Research area:

  • Mechanism theory: I am deeply involved in research on machines and mechanisms theory and I am quite active in IFToMM, the major scientific society for this field. I was Chair of the 12th IFToMM WorldCongress that has taken place in Besancon, April 17-21, 2007 and of IROS 2008, being held at Nice, see here for a photo of the race). I was president of the French Committee of IFToMM, the international federation for the Promotion of the Science of Mechanisms and Machines which regroup about 50 countries and I am member of the Executive Council of IFToMM since 2011. I am a member (and former chair between 1997 and 2005) of the IFToMM technical committee on Computational Kinematics and a member of the IFToMM permanent commission on History. I am also member of the IFToMM Permanent Commission for Communications, Publications and Archiving. As for conferences I am or was member of the Scientific Committee of multiple IFToMM conferences, the next one I will organise being ARK in 2016.
  • Parallel manipulators (Stewart platform and others): ( a state of the art (2002), some open problems (1999) ). Some examples: version 1 of the micro-robot MIPS with 3 DOF for endoscopy (1998): diameter 7mm, length 2.5cm, version 2 of the micro-robot MIPS with 3 DOF for endoscopy (2000): diameter 8.6mm, length 2.5cm. Since 2004 I am working on a new type of parallel wire-driven robot with applications in medical rehabilitation (with force control), service robotics and ultra-fast pick-and place (with velocities expected to be larger than 100m/s). The new prototype first move has been in mid 2007 (see here for the first video, avi format 7.3 Mo). One important feature of this robot is that it will be modular: it's geometry may be adapted to the task to obtain the most appropriate performances (we will not use the same robot for ultra-fast pick-and-place and medical robotics...). We are working on the development of algorithms to determine which geometry is the most appropriate for a given task, while taking into account that the physical instance of a theoretical solution will differ from it, due to unavoidable errors during the manufacturing.
  • Cable-driven parallel robots: Since 2008 I am working on several prototypes of cable-driven robots such as MARIONET-REHAB whose maximal speed may reach 100m/s but is mostly used for rehabilitation. We have built in 2009 the large crane MARIONET-CRANE with a workspace volume larger than 2000 cubic meters, to be used as a rescue device during natural catastrophe (see photos of 5 of our 6 winches and of the manipulation of a victim ). The MARIONET family includes also MARIONET-ASSIST, a robot used for transfer operation in assistance robotics (lifting elderly and improving their mobility), MARIONET-VR that is able to lift a human and will be used in an immersive room for rehabilitation and the MARIONET-SCHOOL's that are used for demontration and teaching teaching of scientific concepts in mathematics, physics, mechanics, control, computer network.. The use of these prototypes have shown that there was severe discrepencies in their theoretical analysis, which differ slightly from the analysis of parallel robots with rigid legs as cables can pull but cannot push. I am currently completely revisiting all aspects of their analysis in order to provide a sound theoretical background. MARIONET-VITRIFICATIONS, a variant of MARIONET-CRANE, has been developed for the artist Anne-Valéie Gasc for the 3D printing of glass micro-beads.

    The MARIONET family of cable-driven parallel robots :REHAB, CRANE, ASSIST, VR, SCHOOLs, VITRIFICATIONS see here for other examples.

  • Assistance robotics: we have started in 2009 a long term effort on this topics. Our purpose is to develop systems fulfilling real needs, with a low intrusitivity, various interfaces and being low cost (guidelines and priorities are given here while developed hardware are presented here). For that purpose we will use only standard hardware, relying on modularity to adapt the device to the situation. We have taken the time to consult several experts in this field (pratician, end-users) to identify the needs and we have built a full scale flat to experiment with our devices. We are currently experimenting with our walking aid ANG which offers mobility help, fall prevention/detection, gait monitoring/rehabilitation, kerb detection (for updating stree maps), is able to vacuum cleaning and many other functionalities. A lighter version ANG-light with encoders on the rear wheels and a 3D accelerometer allows one to monitor the traject of the walking aid. We are also installing a wire-driven crane that will offer assistance for lifting end-users, provides a walking aid and is able to manipulate objects. A larger one will be installed in an immersive room together with a motion base for building a modular, flexible and affordable rehabilitation station. An important point is that all these devices are also smart objects that can communicate with each other in order to collaborate for dealing with distress situations such as a fall. These actions were part of the large INRIA initiative PAL (Personnaly Assisted Living) which regroup 10 INRIA teams working on the topic of assistance for the elderly/handicaped people. For these topics I am developing various assistance devices with embedded computers and hardware such as Arduino and Phidgets which are coupled to various sensors and actuators.An important point is that these objects are multi-functionnal: they may assist but also monitor human health (with a high protection for the data and repecting at best human privacy). Our monitoring goals are to design the assistive device for providing the best data in spite of the unavoidable sensing errors and then to merge/fuse the data to provide synthetic health indicators, usable by doctors to detect trend and possible pathology, taking into account sensing errors and human variability. The methods we are using are a mix of signal processing, deterministic (based on interval analysis) and statistical algorithms both for the model we are using (whenever such models exist) and for managing human variability, while we are also looking at learning algorithms that will be model-free. Our long term goals are to provide medically pertinent indicators, adapted to the subject and the doctors needs, with associated information on their validity, while being able to detect rare events that may be the announciating signs of a pathology.

  • Interval analysis: I am developing both the mathematics and the implementation of this mathematical tool that is often very appropriate to manage in a realistic way problems of robotics and mechanism theory (see here for a tutorial and examples of applications) and more generally the appropriate design of systems that will provide almost all design solutions that are guaranteed to satisfy a given list of requirements in spite of unavoidable differences between the system model and its physical instance. I am the main developer of the ALIAS library, a mixed MAPLE/C++ library, that allow to manage efficiently interval analysis. I am also involved in algebraic geometry which often play a major role in mechanism theory.
    • My main reseach area in mechanism theory are:

  • Forward and Inverse Kinematics
  • Singular configurations
  • Workspace determination
  • Optimal design

    • studied with geometrical and numerical methods.

    Most of the algorithms I have developped in mechanism theory make an intensive use of optimization and systems solving methods developped in the project library

  • ALIAS based on interval arithmetics (see the latex ALIAS-C++ manual in pdf or in html , or the ALIAS-Maple documentation in html, pdf Some pedagogic example of the use of ALIAS are presented here and a complete list of examples here. The official ALIAS page is here. Since 2021 I have investigated the appropriate use of neural networks to solve kinematic problems (appropriate meaning that we are looking for exact solutions and not the very approximate one provided by AI).

    I have also some transversal activities regarding:

  • the famous bibliometry indicators that are now very often used to evaluate the scientific impact although they are very often not well mastered, used and understood (see for example the 2007 analysis document (in French) and its english version of INRIA Evaluation Board on this subject). A small exercise: let FI=m/n where n is a fixed integer and m a measure that is guaranteed to lie in the range [k,7k] where k is known 1: determine the possible range for FI 2: relate this calculation to journal impact factor presented with 3 digits.
  • robotics prospective (see the 2006 prospective document established by INRIA project-teams on the subject)

    Available softwares:

  • Mechanisms
  • xjpdraw, a drawing editor
  • ALIAS (html documentation ). This software is free-of-use for academy. It is available for PC/Linux and is constituted of two parts: a C++ library (to use with g++ 2.5 or 4.1) and a Maple (version 5.5 or 9.5) interface. Currently ALIAS is not downloadable but we are working on it.

    • My bibliography

  • My bibliography (with links, some are here, most recent publications (since 2005/4) are now available through Hal-Inria)
  • Integral Research Reports (query in INRIA research report database)

    Journals

     EJCK (Electronic Journal of Computational Kinematics), 1999-2003 (editor in chief)

    IEEE Transactions on Robotics and Automation (associate editor, 2002-2005)

    Mechanism and Machine Theory (associate editor 2005-2012)

    ASME Journal of Mechanisms and Robotics (associate editor 2008-2011)

    Robotics bibliography

    Robotics bibliography database (15 000 references)

    Information on Parallel manipulators

  • Table of contents of the book "Parallel robots" published by Kluwer in June 2000. A second edition published by Springer is available since the end of 2005 (here is the pdf file of the table of contents)
  • 207 drawings of parallel robots

  • References on parallel manipulators, Copyright J-P. Merlet/INRIA

    Note 1: This is only an informative service. You can get directly references for which a link GET IT is shown.

    Note 2: this database is intended to be exhaustive. Hence the presence of a paper in the database does not provide information on its quality

    Note 3: You may submit a paper for inclusion in the database provided that you provide all the reference data and a copy of the paper. Inclusion time is extremely variable.

    Note 4: this base does not include papers published in journal using Articles Processing Charges (APC) because: 1) such a system is extremely costly for the community while free dissemination is now quite easy 2) the quality of these papers are, in general, low.

    The references database has a search engine allowing to submit queries on authors, title, year, address (for conferences) and keywords.

    4893 references

    Last update:08-28-25

    full bibliography (pdf file 924K)

    Another version of the full bibliography: the references are no more labelled with number (which will change after an update) but with a perennial label that will not change

    full bibliography with perennial index

    References by author

  • Authors A-F
  • Authors G-I
  • Authors J-L
  • Authors M-N
  • Authors O-S
  • Authors T-Z

    This small text explains the various References chapters that follow:

  • A parallel robot has a mechanical structure that is determined through the structural synthesis, an approach that relies on design theory and mobility analysis. See the chapters Architectures, patent and ... DOF Robot. You may also check if the considered structure has not been aleready studied (use the above link to the robots drawings).
  • After having chosen the structure one has to move to the design phase and to optimisation which allow one to determine the dimensioning of the robot, based on analysis that relies on criterion such as isotropy (which involves the inverse jacobian and the jacobian), workspace (authors A-L, authors M-Z) (for which we need the kinematics modeling (authors A-L, authors M-Z)), singularities (authors A-L, authors M-Z), accuracy, stiffness, static, velocity analysis, dynamics (authors A-L, authors M-Z) (may be with vibrations analysis or a l'balancing analysis of the robot), or, in other words, to perform a performance analysis
  • We must also take into account the hardware, such as the actuators and the passive joints, without neglecting the effect of clearances and other uncertainties.
  • For control (authors A-L, authors M-Z) we will need the Forward Kinematics (authors A-L, authors M-Z) to study trajectories and to plan them (motion planning) We may need also to perform a calibration of the robot. Numerical methods that cab be used to sove these problems are diverse (geometry, algebraic geometry, continuation, interval analysis) but has recently appeared the use of neural networks that have still to progress.
  • Some parallel robots are special due to their actuation mode (binary robot or wires-driven robots (authors A-L) (authors M-Z) which require a specific analysis or because of their special mechanical architecture: decoupled robot with translation and orientation that can be independently controled, flexible robot, hybrid robot, mixing serial and parallel architecture, modular robot whose geometry can be changed to adapt the robot to the task at hand, redundant robot with more actuators or sensors than necessary or singular robot that are always in a singular configuration.
  • The applications (authors A-L, authors M-Z) of parallel robots are multiple: force sensor, crane machine-tool, medical, micro-robot, simulator, articulated trusses, joystick, aerial, maritime. We must also consider the fiability of the robot and its energy comsuption. We may also consider an appropriate use of AI in some cases.

    • 2 DOF Robot 3 DOF Robot 4 DOF Robot
    5 DOF Robot 6 DOF Robot Accuracy
    Actuators Aerial Applications
    Architecture Assembly Balancing
    Binary robot CAD Calibration
    Clearance Compliance Control
    Crane Decoupled Robot Design
    Design Theory Direct kinematics Direct kinematics with extra sensors
    Duality Dynamics Energy
    Fiability Flexible Robot Force-feedback
    Force sensor Hardware Hybrid
    Hydraulic Inverse Jacobian Matrix Isotropy
    Jacobian matrix Joystick Kinematics
    Kinetics Machine-tool Maritime
    Maximal workspace Medical Micro-macro robot
    Micro-robot Mobility Modular Robot
    Motion planning Neuron network Optimization
    Orientation workspace Passive joints Patents
    Performance analysis Piezo-electric Planar robot
    Pneumatic Redundant Robot Safety
    Simulation Simulator Singularities
    Singular motion Singular Robot Spherical Robot
    State of the Art Static Stiffness
    Structural synthesis Trajectory Truss
    Uncertainty Vibration Wire Robot
    Workspace Wrist

      références par année/per year

      1813 1897 1906 1908 1931 1942 1957 1962 1965 1967 1972 1973 1975 1976 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025

      Number of References per year

      Journals statistics per year

      Other sites on parallel robots

    1. Parallel Mechanisms Information Center(courtesy I. Bonev)

      2. Applications of Parallel manipulators

      (Pictures)

      (Gough platform Movie,188 Ko)

      (Active wrist Movie,140 Ko)

      more photos/vidéos on the site of the HEPHAISTOS team

      the WEB references of the book "Parallel robots", 2005 edition

      Others servers

    2. Parallel robots in the world:

      LIRMM (Montpellier)|| CERT ONERA|| POSSO (Polynomial solving)|| ETH Zurich|| Laboratoire de robotique de l'Universite Laval|| McGill University || Northwestern University|| Platform 2000 (Spine group, Yale) || Endoscop (Berkeley) || Space System Lab (Maryland Univ.) || Canterbury Univ. (New-Zeland) || National Institute of Standards and Technology (NIST)|| CONSTRUCT project || Joystick (Iwata Lab, Tsukuba, Japon) || Joystick (AEA, Karlsruhe,Germany) || Amsterdam Univ. (Holland) || Frasca Flight Simulator (Urbana, Illinois) || CAE Flight Simulator (Canada) || NASA Simulation System Branch (Langley) || Ingersoll milling machine || || Ultra-light plane application || Drilling machine (Arlington Robotics Research Institute) || || Milling machine (ACROBAT project, Italy) ||

      Other robotics projects at INRIA:

      see ACENTAURI, ACTUS, CAMIN, CHORALE, CHROMA, DEFROST, FLOWERS, LARSEN, RAINBOW on INRIA web site

      Some other links (may be not up to date)

    3. Conferences robotics,ai,control
    4. Robotics on INTERNET (thanks to Sarah Wahlberg)
    5. Robotics on INTERNET (Univ. Indiana)
    6. Students in robotics
    7. CNAM museum
    8. EEVL: reference site in Engineering
    9. For pipe smokers..

      HEPHAISTOS Home Page