Newsgroups: comp.robotics.research
Path: cantaloupe.srv.cs.cmu.edu!europa.chnt.gtegsc.com!news.mathworks.com!hookup!nstn.ns.ca!newsflash.concordia.ca!sunqbc.risq.net!hobbit.ireq.hydro.qc.ca!NetNews.IREQ.Hydro.QC.CA!mboyer
From: "Victor B. Naumov" <nau@iias.spb.su>
Subject: Russian Robotics 1/8: [LONG] Summary of Research
Message-ID: <MBOYER.95Jul5164119@pellan.ireq-robot.hydro.qc.ca>
Lines: 481
Sender: news@ireq.hydro.qc.ca (Netnews Admin)
Organization: St.-Petersburg Institute for Informatics and Automation, RAN
Date: Wed, 5 Jul 1995 20:41:19 GMT
Approved: mboyer@ireq-robot.hydro.qc.ca, crr@ireq-robot.hydro.qc.ca

 RUSSIAN ROBOTICS RESEARCH AND ACTIVITY OF ROBOTICS LAB.
 ======================================================

  Prof. F.M.KULAKOV

  St.Petersburg Institute for Informatics and
  Automation of Russia Academy of Sciences,
  Robotics Laboratory
  199178 St.Petersburg, Russia
  39, 14-th Line, St.Petersburg, Russia, 199178

  phone: 007(812) 218 10 88;
  fax: 007(812) 217 51 05:
  e-mail: kul@lrob.iias.spb.su

========================================================
  Table of Contents For Research Summary
   
    1. HISTORICAL INFORMATION

    2. PROJECTS: SPACE AND SPECIAL ROBOTS

     2.1. THE SYSTEM OF BOARD MANIPULATORS (SBM)
     2.3. MOBILE ROBOT ON WHEELS
     2.4. SPECIAL CATERPILLAR ROBOTS WITH A MANIPULATOR
     2.5. THE RESONANCE MANIPULATOR
     2.6. THE ARTIFICIAL MUSCLE
     2.7. MOBILE ROBOT FOR MARS RECONNAISSANCE
 
    3. ACTIVITY OF ROBOTICS LAB
 
     3.1. FUNDAMENTAL RESEARCH OF ROBOTICS LAB
      3.1.1. GENERAL RESEARCH OF ROBOTICS LAB
      3.1.2. METHOD OF AUTOMATIC FORMATION OF DESIRED TRAJECTORIES
      3.1.3. THE THEORY OF SUPERVISORY ROBOT CONTROL
      3.1.4. METHODS OF RECURRENT FORMATION OF FUNCTIONAL MODELS
              OF COMPLEX CONTROLLABLE AND UNCONTROLLABLE MECHANISMS
              AND CONSTRUCTIONS.  ALSO FUNCTIONAL MODEL TO IMITATE 
              HUMAN MOVEMENT
     3.2. APPLIED RESEARCH
      3.2.1. PRACTICAL REALIZATION OF SUPERVISORY CONTROL
      3.2.2. SOFTWARE-HARDWARE OF 6D FORCE/TORQUE SENSOR
              AND 6D JOYSTICK
      3.2.3. THE SYSTEM CONTROL FOR UNTRADITIONAL USE OF ROBOTS
      3.2.4. PACKAGE COMPUTER AIDED ENGINEERING FOR INDUSTRIAL ROBOTS
              (CAEIR) DESCRIPTION OF FUNCTIONS

    REFERENCES

====================================================================
1. SOME HISTORICAL INFORMATION

The year 1948 can be considered the starting point of robot
manufacturing in the former USSR.  The first master-slave
anthropomorphic manipulator was created that year.  Soon, other
similar manipulators were built.  Their construction was motivated by
the acute need for carrying out work in areas with high radioactivity;
one of many new challenges encountered during the USSR's efforts to
develop an atomic bomb.

In the period from 1948 through 1960 only the USA and USSR had similar
devices; in the USA the first master-slave manipulator was created a year
earlier and used in the National Arizona Laboratory for the same purposes
as in the USSR.  However, if in the USA, the first industrial robot
was created in 1960 and put to use in 1961, in the USSR this was done
only in 1968 (Editor Beljanin P.N., 1973).  By that time, the
manufacturing of industrial robots had been started in the U.K.
(since 1966) and in Japan (since 1968) though on licenses bought in
the USA.

The first experimental robot with tactile sensing and computer control (first
home control computer of type UMNCH) was built in the USSR in 1968
(Ignattev, Kulakov et al., 1970).  The manipulator of this robot had a
load capacity 200 Kg, six degrees of freedom and was provided with a
hydraulic drive which was capable of functioning under water.  This
project was the first attempt to create an underwater robot with
computer control to carry out the undersea work.  In particular, one
task, for this robot was the following:  to carry out the scanning of
a surface with any profile and to collect objects on it.

In 1969 the first experimental robot with supervisory control was
built (man was the supervisor) (Kulakov, 1976, 1977).  It was two
anthropomorphic arms.  Each grip of the arms had 48 tactile sensors .
They covered all surfaces of grip and was used as an artificial skin
(Kulakov, Uspenskiy et al., 1973).  In addition, there were 12
proximity sensors.  The man-robot interface was composed of a TV
screen.  The TV-screen showed the environment of the robot and 
the operator indicated targets with a light pen (which must be round).

Control computer software provided an opportunity to plan the
trajectories of the robot motion in its environment using knowledge of
obstacles which were indicated by light pen on the TV screen and by
tactile and proximity sensors.  This project was the first attempt to
create a control system for a space robot.  These projects were on the
cutting edge of technology for world wide robot development of that
period and in some areas they exceeded this level.

In 1971, the well known scientist Prof. Nevins from Driper lab moved
from the USA to the USSR.  He was surprised by the level of research 
being conducted in Russian robotics and asked for film documenting the
function of the supervisor controlled robot.  In accordance with our
request, he wrote an appendix about teleoperators in the first
monograph on robotics (Ignattev, Kulakov et al., 1972) published in
the USSR in 1972.  It was quickly translated into English (Ignattev,
Kulakov et al., 1973). 

However, beginning from 70's the rate of robot development in the USSR
started to lag behind the world level, especially in the field of
production of industrial robots.  The first cause of this can be
explained by cheap labor in the USSR, which made the use of separate
robots in the production process uneconomical even in FMS.  During
this period robotics developed largely due to the enthusiasm of teams
of scientists and engineers who managed to get financial backing for
their projects and promised good perspective in the future.

Due to this fact, by 1989 large number of industrial manufacturers, in
the framework of the governmental program UIntensification '90, had
been equipped with IR's, and more advanced factories had been equipped
with FMS.  By that time, robotics had been introduced as a new
speciality at technical Universities.

At present the level of robotics research in our country continues to
sink; the main reason is reduction of financial backing.  In spite of
this, in Russia a number of interesting research projects continue to
develop, largely due to the enthusiasm of teams of scientists and
engineers.  Some of these projects are very promising for the further
development of robotics.

2.  PROJECTS: SPACE AND SPECIAL ROBOTS

2.1. THE SYSTEM OF BOARD MANIPULATORS  (SBM)

SBM for the space system of multiple using (SSMU) has been developed
by Central Institute of Robotics and Technical Cybernetics.  SBM has
too identical anthropomorphic manipulators 15 meters long.  The main
task of SBM is unloading and loading of loads, which are furnished by
SSMU on the orbital station.  It is capable of transferring mass up to
30 tons.  The mass of the manipulator is 360 kg.

SBM is controlled by a reliable board computer.  Control of SBM has
both autonomous mode and master-slave mode.  Master-slave mode
involves the robot being controlled locally by a cosmonaut or remotely
from a ground station on Earth.  Master-slave mode control is realized
by the operator with the help of a camera on the manipulator, on SSMV
and on the orbital station. 

There is a control input language for the manipulators to form a
flight program.  At present SBM has been built and tested.  In the
process of creating the SBM, researchers accumulated great experience
in carrying out different tests of space manipulators and their parts,
and of training cosmonauts by using a unique test complex to imitate
work under space conditions.  This complex is located in a building 70
m high and 30 m in diameter.  It is hollow inside and is equipped with
different systems for simulating the behavior of bodies under
weightless conditions.  This complex allows us to test all the steps
of flight program for SBM.  It is possible to use this complex for
international cooperation.

Our institute has begun work on a new project creating a space
manipulator for the Russian module of the International space station.
The project name for this manipulator is compasses.  It is a two
legged robot with each leg ending in a gripper.  The robot moves by
grasping special handles placed on the surface of the space station.
If one of the legs has securely grasped to a handle, it is possible
to use the other leg as an arm for useful work.

2.2.   THE LEGGED ROBOT

The legged robot has been developed by Institute of Applied
Mathematics RAS (Ohozimskiy et al., 1984; Golubev et al., 1989).  The
robot has six legs, each of which has 3 degrees of freedom.  It is
equipped with tactile and optical scanner sensors; the latter is
intended to map out the topography of the place in which the robot has
to go.  The robot is controlled by a board computer, which plans the
movement of the robot in its environment based on the given goals and
visual based world model.  At present there are experimental
prototypes of the robot.

2.3.   MOBILE ROBOT ON WHEELS

Mobile robots on wheels have been created for the traditional tasks
encountered in FMS.  At present, wheeled robots for FMS are in
production.  The main characteristics of these robot are approximately
the same as characteristics of mobile robots of other manufacturers.

2.4.   SPECIAL CATERPILLAR ROBOTS WITH A MANIPULATOR

Special caterpillar robots with a manipulator were developed for tasks
involving reconnaissance and emergency repairs in dangerous
conditions, especially in the case of damage to nuclear power
stations.  Now this robot is used for the visual inspection of rooms
and equipment of nuclear power stations in zones with high nuclear
radiation.  They are also used to measure nuclear radiation and the
temperature of air, in addition to, carrying out auxiliary operations
using its manipulator.

A prototype of this robot has carried out different tasks in 
processing the of damage of the Chernobyl Nuclear Station.

Other areas of mobile robot development include:  floor cleaning in
the Metro (Subway), territory guarding, and using against terrorists
in airplanes.

2.5.   THE RESONANCE MANIPULATOR

(Developed by the Institute of Machine knowledge RAS)

The resonance manipulator's principle of work is based on the
resonance of natural frequency vibrations of the mechanical arm with
the repeating rhythm of simple operations performed by the robot
manipulator.  This manipulator can be described as a quasi -
conservative system with minimal dissipation.  It is possible to
change its natural frequency at the expense of the stiffness change of
special springs intended to store potential energy in the system.
This principle ensures high energy savings and super high productivity
in carrying out simple operations if need be.

At present there are experimental prototypes of such manipulators.

2.6.   THE ARTIFICIAL MUSCLE

(Developed by St. Petersburg Technical University)

The artificial muscle uses a quickly progressing chemical reaction
accompanied by intensive gas emittence.  The reaction arises when
electric current flows through a special substance, with which a cord
is impregnated - an artificial muscle - because it is porous.  Due to 
the chemical reaction, the pores of the artificial muscle widen in the
transverse direction and narrow in the longitudinal direction.  This
leads to the constriction of the cord.

The reaction stops when the current is cut and the pores and cord
return to their normal shape.

At present there are experimental samples of artifice muscles in
production.  Now the main task of this project is to improve the
coefficient of useful action [expansion] of the artificial muscle.

2.7.  MOBILE ROBOT FOR MARS RECONNAISSANCE

(Developed by the Institute of Transport Mechanical Engineering)

Before this project the Institute had developed a mobile robot for
lunar exploration, in Russian - uLunohodf.  The mobile robot for mars
has improved the moon based chassis and perfected the control system
and sensor system which uses a supersonic scanner sonar.  These
systems allow the mapping of topography, planning of the movement
trajectory, and avoiding obstacles.

At present a prototype of this robot exists.


3.    ACTIVITY  OF  ROBOTICS  LAB.

The robotics lab which I am leading is a very old robotics lab.  All
the peculiarities of the history of robotics in this country is
reflected in the processes of its activity.  The first experimental
underwater robot and first experimental robot with supervisory control
were realized in this lab.  The head of this lab is one of author of
the first monograph on robotics, published in the USSR [Ignattev,
Kulakov , 1973].  It is necessary to note that in 1976 this lab was a
part of the Technical Cybernetic Institute.  Now this Institute is
called the Central Institute of Robotics and Technical Cybernetic.
The word Robotics in its title appeared thanks to the successful
activity of our robotics lab.

After 1976 the lab became part of the St. Petersburg Institute for
Informatics and Automation of Russia Academy of Sciences.  The title 
of Central Institute Robotics and Technical Cybernetics was retained.

The lab carries out work in 4 scientific directions, two of them can 
be considered fundamental research.

1.  Development of innovative methods of control of complex spatial
    multidimensional mechanisms.

2.  Development of methods of CAS CAD of controllable and
    uncontrollable machines, robots in particular.

The rest of the work has the character of applied research. The theoretical
basis for this work is the research in the above mentioned directions:

1)  The first is software-hardware implementation of advanced methods of
    control of complex multidimensional mechanisms and their
    exploration with the help of computer and physical simulation.

2)  The second direction is designing of packages of applied programs for
    CAD.

In the lab we use the following technological sequence:  development
of a theory - computer experiment - physical experiment - correction
of the theory - new experiment.

We pay special attention to the physical experiment.  For this purpose
we have built a flexible experimental complex, on which we can
investigate methods of control covering a wide range of possibilities.


3.1.    FUNDAMENTAL RESEARCH OF ROBOTICS LAB.

3.1.1. GENERALIZATION RESEARCH OF ROBOTICS LAB.

This research seeks to perfect the lowest level of the robot control
system.  This includes components which are directly connected with
drives and produce the control signals for them.  We have produced
continual active methods of force control for real robots which have
elasticity in the transmission from the drives to the links of the
manipulator, wrist elasticity and elasticity of links (Kulakov, 1993;
Kulakov, 1994).  New control laws of force control were also
developed.  They include the generalization of the known force control
laws (Stiffness Control, Damping Control, Explicit Force Control,
Hybrid Position/Force Control) and provide stability of perturbed
motion of the closed force control system.  The computer aided
simulation of dynamic robot motion with the new force control
illustrated advantages this method of force control.

3.1.2.  METHOD OF AUTOMATIC FORMATION OF THE DESIRED TRAJECTORIES

Some research being conducted here is targeted at perfecting the
tactical level of robot control; this level forms the goals (desired
program trajectories) for the lowest level.  These problems have been
investigated in our country for a long time.  One of the first methods
of automatic formation of the desired trajectories was developed
here [Kulakov,1980; Kulakov, 1982].

This method is based on the methods of nonlinear programming. The
desired trajectory is formed in the space G of the joint coordinates g as the
discrete sequence of the arguments g1,g2,..., gn of some function, which
provides the global minimum of this functional; in the case where this
minimum is found, the task is executed.  A more in depth discussion of 
this method is provided in:

Russian Robotics 3/8: Automatic Path Generation

3.1.3.  THE THEORY OF SUPERVISORY ROBOT CONTROL

In our lab, the highest level of robot control provides the
characteristics of a highly autonomous system and universality of
robot.  Achieving these characteristics simultaneously is a very
complex, contradictory problem.  For the practical use of robots,
especially in extreme, variable, nondeterminate environments, the most
important property is universality.  This approach to the construction
of a robot controller for operation in a complex environment allows an
increase in the degree of universality of the robot at the expense of
its fully autonomous nature.  We call this supervisory robot control,
with man as the supervisor.


3.1.4.   METHODS OF THE RECURRENT FORMATION OF FUNCTIONAL MODELS
         OF COMPLEX CONTROLLABLE AND UNCONTROLLABLE MECHANISMS
         AND CONSTRUCTIONS AND ALSO FUNCTIONAL MODEL TO IMITATE
         MAN MOVEMENT

The essence of this work is the creation of mathematical methods, the
basis of which are recurrent calculations and corresponding
algorithms.  These algorithms make it possible to construct model
equations of any concrete mechanism with a very economical set of
parameters used to characterize the mechanism (Kulakov 1982, Heiman
1987).


3.2.  APPLIED RESEARCH

3.2.1.  PRACTICAL REALIZATION OF SUPERVISORY CONTROL

For research in area of supervisory control a special experimental
robot with supervisory control was developed.  It is used for the
verification the new theoretical results.  An in depth discussion of 
this area of research is provided in:

 Russian Robotics 4/7: Supervisory Control

3.2.2.  SOFTWARE-HARDWARE OF 6D FORCE/TORQUE SENSOR AND 6D JOYSTICK

6-D Force-Torque Sensor 
This sensor realizes the measuring of forces and torques, acting on
different objects.  The mechanical part of sensor consists of two
flanges connected by a special elastic tubular construction.  A more 
in depth discussion of this hardware, including technical specs are 
given in:

Russian Robotics 5/7: 6D Force/Torque Sensor

6D joystick
Inputs to the joystick are force and torque, applied to it.  They are
transformed in 6 analog electrical signals by strain gauge sensors
mounted on the joysticks construction.  A more in depth discussion of
this hardware, including technical specs are given in:

Russian Robotics 6/7: 6D Joystick

3.2.3.  THE SYSTEM CONTROL FOR UNTRADITIONAL USE OF ROBOTS

For the simulation of a body's motion in weightless conditions
a special device has been developed.  It is similar to a robot with 
six DOF equipped with a special gripper.  The "weightless" body is
placed into this gripper.  The robot uses a six dimensional wrist
force/torque sensor to define the forces and torques applied to the
body.  A more in depth discussion of this project is provided in:

Russian Robotics 7/8: Weightless Robotics

3.2.4.   PACKAGE COMPUTER AIDED ENGINEERING FOR INDUSTRIAL ROBOTS (CAEIR)
         DESCRIPTION OF FUNCTIONS


Package CAEIR realizes the initial stage of design for controlled
mechanisms and building constructions, formed of systems of bodies by
using computer simulation.  Typical computer simulation involves
numerical solution of direct and inverse dynamic, kinematic, and
geometric tasks for mechanisms and buildings.  For more detailed 
information see article:

Russian Robotics 8/8: CAE for Industrial Robots


REFERENCES

Golubev, U.F., Degtjarjeva, E.V., Phairullin, R.Z. (1989).
Optimal control of leg of leged robot. Preprint  N59  Institute
of Applied Mathematics named M.V.Keldish of Academy of Sciences
of USSR, Moscow, 1983.

Ignattev, M.B., Kulakov, F.M., Pokrovskiy, A.M (1972). The Robot
manipulator Control algorithms. Leningrad, Mashinostroenie, 1972,
248 p.

Ignattev, M.B., Kulakov, F.M., Pokrovskiy, A.M.  (1970). To problem of
development underwater manipulator with automatic control.
Okeanologia, 1970, vol. V, pp. 37-54.

Kulakov, F.M.(1976, 1977).  Method supervisory Robot Control.
Izvestija of Academy of Sciences of USSR, Tehnicheskaja Kibernetika,
1976 N 5, N 6; 1977 N 1.

Kulakov, F.M. (1980). Supervisory Robot Control. Nauka, Moscow,
1980, 448 p.

Kulakov, F.M., Uspenskiy, V.N., Trubnikov, G.N.(1973). Two-finger
grusp of manipulator. Patent N 423623 from 02.01.73.

Test - Construction Development and Researches in Industrial
Robots area. (1973). Proc. of  Institute, 1973 U. (Editor Beljanin
P.N.).

Ohozimskiy, D.E., Golubev, U.F.(1984). Mechanics and Control of
automatic apparat.  Moscow, Nauka, 1984.

Ignattev, M.B., Kulakov, F.M., Pokrovskiy, A.M. (1973). The robot
Manipulator Control  algorithms. Virginia, Joint Publication Research
Arlington, 1973 and Rep. NN 59717 NTIS,  Springfield Va, Aug. 1973.

Kulakov, F.M. (1993). About new aspects of compliance control. Proc. of the
Intern. Conf. on  CAD/CAM, Robotics and Factories of the Future,
St.Petersburg, Russia, May, 17-19, 1993.

Kulakov, F.M. (1994). Generalization of Robot Force Control Theory. Proc.
IROS'94,  Munchen, September 12-16, 1994.

Kulakov, F.M. (1982). Modeling Robot Control in Assembly Operations.
Modern Robot  Engineering, Moscow, MIR Publishers, 1982, pp.100-116.

Kulakov, F.M. (1991). Problems of Force Adaptation of Assembly Robots.
China.  Proceedings Intern. Conference on Information Technology   for
Advanced Manufacturing  Systems, Nanjing, Jiawgsy,  P.R.China,  1991.

Whitney,  D.E. (1985).  Historical Perspective and State of the Art
in Robot Force Control.  Trans.IEEE Int. Conf. on Robot, 1985.

Kulakov F.M. Automated research and design of robot. Proc. 5th Int.
IFIP/IFAC Conference PROLAMAT - 82, Leningrad, USSR, 16-18 May, pp.
507-522.

Heimann B., Kulakov F.M., Loose H. Nollau R. Maschinen-Bautechnic. N 11,
1987.


--
 *********************** (moderated) ***************************
 Submissions:                 Meta-discussions/information:
 crr@ireq-robot.hydro.qc.ca   crr-request@ireq-robot.hydro.qc.ca
