DTIC AD0836039: HARDIMAN I PROTOTYPE PROJECT

UNCLASSIFIED

_ AD NUMBER _

AD836039

LIMITATION CHANGES
TO:

Approved for public release; distribution is
unlimited.

FROM:

Distribution authorized to U.S. Gov't, agencies
only; Administrative/Operational Use; 10 JUN
1968. Other requests shall be referred to Army
Mobility Equipment Research and Development
Center, Attn: SMEFB-HM, Fort Belvoir, VA 22060.

_ AUTHORITY

USAMERDCOM ltr, 9 Sep 1982

THIS PAGE IS UNCLASSIFIED

AD 836039

Prepared by

Specialty Materials Handling Products Operation
General Electric Company
Schenectady, New York 12305

10 June 1968

—»

Supported Jointly by ^

Engineering Psychology Branch (Code 455)

Office of Naval Research
Washington, D.C.

Contract Authorization Identification No. NR 196-049

and

Army Mobility Equipment Research and Development Center

Fold Belvoir, Virginia

United States Army Project No. IM624101050702

Each transmittal of this document outside the agencies
of the U. S. Government must have the approval of the
Army Mobility Equipment Research and Development
Center (Attn: SMEFB-HM), Fort Belvoir, Virginia 22060,
or the Engineering Psychology Branch of the Office of
Naval Research (Code 455), Washington, D.C. 20360.

Contract N00014-66-C0051

S-68-1081

HARDIMAN I PROTOTYPE PROJECT

SPECIAL TECHNICAL REPORT
ON JOINTS IN SERIES

Prepared by

Specialty Materials Handling Products Operation
General Electric Company
Schenectady, New York 12305

10 June 1968

Supported Jointly by

Engineering Psychology Branch (Code 455)

Office of Naval Research
Washington, D. C.

Contract Authorization Identification No. NR196-049

and

Army Mobility Equipment Research and Development Center

Fort Belvoir, Virginia

United States Army Project No. IM624101050702

Each transmittal of this document outside the agencies
of the U. S. Government must have the approval of the
Army Mobility Equipment Research and Development
Center (Attn: SMEFB-HM), Fort Belvoir, Virginia 22060,
or the Engineering Psychology Branch of the Office of
Naval Research (Code 455), Washington, D. C. 20360.

Contract N00014-66-C0051

f

FOREWORD

This study has established, a method for
stabilizing high performance servos that are in
series. The Hardiman servos can be stabilized
and made to perform as required using this tech¬
nical report as a guide.

4

This report represents the successful
completion of the servo analysis work initiated
during the Special Interim Study (see Special
Interim Study . S-68-1060, 19 April 1968)

ii

ACKNOWLEDGMENT

We wish to thank the following scientific officers and
project monitors for their efforts in the guidance and di¬
rection of this program during the period covered by this
report:

Mr. J.W. Beaudet
U.S. Army Mobility Equipment
Research and Development Center
Fort Belvoir, Virginia 22060

Dr. J.W. Miller
Office of Naval Research
Code 455

Washington, D.C. 20360

Dr. M.J. Farr
Office of Naval Research
Code 455

Washington, D.C. 20360

We also appreciate the interest and constructive
comments of Mr. Geyer and members of his staff.

Mr. G.J. Geyer
Naval Air Systems Command
Code Air-5344
Washington, D.C. 20360

•....

TABLE OF CONTENTS

Se ction

FOREWORD .

ACKNOWLEDGMENT .

1 INTRODUCTION .

2 EXOSKELETON STABILITY ANALYSIS

Introduction to Analysis .

Scope of Analysis • •••••••

Results and Conclusions from Analysis

Analysis Procedure .

Recommendations .

Appendix I -- DERIVATION OF SYSTEM

diagrams

Appendix n - COMPUTER RESULTS

LIST OF ILLUSTRATIONS

Figure Page

1 Single Joint Block Diagram . 15

2 Typical Frequency Response for a Single Joint --

Proportional Gain Only .. 16

3 Typical Frequency Response for a Single Jdiht --

Proportional and Rate Compensation . 17

4 Typical Frequency Response for Single Joint --

Proportional and Rate and Lag Compensation . . 18

5 Signal Flow Diagram for Three Joints in Series

Tickler Control-- Unilateral . 19

6 Two Joints in Series Block Diagram -- Unilateral . 20

7 Typical Frequency Response of Two Joints in

Series -- Proportional Gain Only . 21

8 Typical Frequency Response of Two Joints in

Series -- Proportional and Rate Compensation „ 22

9 General Bilateral Position Servo . 23

10 Block Diagram of Bilateral Servo with G^s) =

G g (s) = G(s) 24

11 Block Diagram of Single Joint-- Bilateral ... 25

12 Signal Flow Diagram of Bilateral Servos .... 26

13 Analog Computer Diagram of Exoskeleton of

Three Joints in Series . 27

14 Recorder Trace of Analog Simulation of Exoskeleton

Arm -- Light Mass . 28

15 Mechanical Parameters . 29

16A Partial Signal Flow Diagram Based on Derivation

of Dynamic Intercoupling Between Links .... 30

16B Same Diagram Rearranged to More Convenient Form.

This is the Form Used in the Complete Signal Flow
Diagram (Figure 12) 30

17 Formula for Each \ . 31

18 Recorder Trace of Analog Simulation of Exoskeleton

Arm -- Heavy Mass . 32

vij.

^f^TrAS.

THE POWERED EXOSKELETON PROJECT

The Powered Exoskeleton concept is that of a material handling
machine under intimate control of the operator.

"Worn as an outer mechanical garment, the exoskeletal structure
will be powered to dramatically amplify the wearer's strength and endur¬
ance by a factor of approximately 25 to one, i. e., when the exoskeleton
wearer lifts 25 pounds, he will 'feel' as if he is lifting only one pound.

The device will provide him with a set of 'mechanical muscles' that
enables him to lift and handle loads in excess of 1000 pounds. The human
operator will 'feel' the objects and forces he is working with almost as
if he were in direct body and muscle contact. This feature, called force
feedback, will provide the operator with sensitive control of the structure
and will act as a safeguard against the application of excessive force.

"The exoskeleton, called 'Hardiman,' mimics the movements of its
wearer, presenting a literal union of man and machine. Thus, the human's
flexibility, intellect, and versatility are combined with the machine's
strength and endurance. "*

* Naval Research Reviews . July 1967

Section 1

INTRODUCTION

During the development of Exoskeleton servos, it became apparent
that the effect of interactions between servos in the series of joints in the
arms or legs must be considered. These interactions have for a long time
been a question in the design or manipulators. The past history has been
to wonder about them, ignore them in design, and find no obvious ill
effects in operation.

The Exoskeleton presented a different situation than previously encoun¬
tered. The needeo servo performance was higher than ever before in
manipulators; the force ratio (25:1) was high; the load (1500 lbs) was much
greater than previous experience. The first look into joints in series
analysis was occasioned by the hope that the intercoupling effects might
make the solution to the Exoskeleton servo problem easier. This was not
to be, however.

It was found that for the particular Exoskeleton parameters, the
intercoupling reduced the stability of the servos rather than helped as had
been hoped. With the help of consultants, Dr. H. Chestnut and P. C. Callan,
a solution to these problems was found. The change to electrohydVaulic
from hydromechanical to provide a more feasible method for servo
compensation has been previously reported. * There remained the need
to show that it is feasible to stabilize the Exoskeleton high performance
bilateral servos in series. This report documents a successful engineering
solution to this feasibility question. It provides technical guidelines for
the stabilization of the general joints in series problem.

The conclusion drawn from this work is that it is possible to stabilize
the Exoskeleton arm servos at a level of performance well above that
needed for satisfactory operation.

*Special Interim Study S-68-1060, 19 April 1968.

BLANK PAGE

Section 2

EXOSKELETON STABILITY ANALYSIS

INTRODUCTION TO ANALYSIS

This report is based on studies made of the tickler activated Exo¬
skeleton.. The hands, arms, and back are assumed to have provisions for
force reflection. The legs and girdle do not have this provision.

Three primary considerations are necessary in examining the stability
aspects of such a system. They are:

• The stability of each joint alone: This requires analysis of the
parameter variations in the joint model due to load changes on the
system.

• The stability of a number of joints connected in series: This re¬
quires analysisof parameter variationsdue to load changes in the model
representatingthe cross coupling terms between individual joints.

Also, the stabilization methods used in individual joints must be
compatible with inter-joint stability requirements. This is because
some terms appear in both places.

• The stability of a number of joints connected in series with force
reflection from slave to master: This case is the most difficult,
and the wrist joint is the worst case since inertia changes of
250 to 1 are experienced from no load to full load.

SCOPE OF ANALYSIS

This section outlines the analysis performed to obtain data needed to
construct the Exoskeleton. A simulation of three joints in series has been
achieved on an analog computer. Both the unilateral (leg system) and bi¬
lateral (arms) have been studied. A program has also been written for the
digital computer using ADA.

The computer models have been used as tools in conjuction with control
theory calculations to determine the compensation networks necessary to
implement the system and have it stable under all operating conditions. .

RESULTS AND CONCLUSIONS FROM ANALYSIS

As a result of the work performed, the following conclusions have been
reached:

• Compensation networks in the slave must consist of a proportional net¬
work, a rate network, and in some cases a lag network. This results

from the fact that a high steady state gain is needed to achieve
required compliance.

This high gain causes inter-joint instability which can be
eliminated by supplying a rate signal. The rate signal, however,
caused the bandwidth of the individual joints to be high, resulting
in possible oscillations due to mechanical resonance.

The addition of a lag term, to each joint, reduces this band¬
width and makes it adjustable. When the lag is supplied in para¬
llel with the proportional and rate terms, inter-loop stability is
still maintained. If placed in series, inter-loop instabilities
result. This is explained in the following subsection, "Analysis
Procedure. "

• In addition to series rate stabilization velocity feedback must be
provided on the master for the bilateral case. The sensor for this
feedback will measure the velocity difference between adjacent mem¬
bers, e.g. , the actuator at the elbow joint will have a velocity trans¬
ducer to measure its motion. The actuators at the wrist, shoulder,
and back will need the same feature.

ANALYSIS PROCEDURE

The analysis procedure is presented in four parts below. First, con¬
sideration is given to a single unilateral joint. Then three unilateral joints
in series are considered. Next, a single bilateral joint is examined, and
finally three bilateral joints in series are analyzed.

Single Joint - Unilateral

A block diagram of a single unilateral joint for the tickler system is shown
in Figure 1. A typical frequency response plot is shown in Figure 2 for the
case where G c is a constant. In general, because of the static compliance
limit of ±1 inch between master and slave, the single joint approaches dyna¬
mic instability. In the paragraphs under the topic "Three Joints in Series -
Unilateral," it is shown that unstable operation results at all times forthree
joints in series at the gains required unless additional compensation is provided.

A method of decreasing the crossover phase shift is to make G (s) a
rate as well as proportional gain. This can be done to produce a lead at
the same frequency as the first lag break. The resultant frequency re¬
sponse plot is shown in Figure 3. It is noted that the crossover frequency
(and thus the bandwidth) is increased by this method. In any event, the
crossover frequency of 100 to 200 radians per second is higher than neces¬
sary for dynamic response of the system, and it is possibly higher that the
resonant frequency of the mechanical structure.

4

A means of decreasing bandwidth while maintaining the same unity
frequency gain and about the same phase shift is shown in Figure 4. This
frequency-gain profile is obtained by putting a lag circuit in parallel with the
proportional and rate circuit. With diis configuration all of the requirements
of steady state and dynamic compliance, stability, and response time cam be
met for a single loop. The general equation for G^s) is:

G (s) = K + K s
c g V

K,

1 + T t s

Jj

( 1 )

The two lag breaks due to the hydraulics and geometrical conf iguration
of the system are functions of A, v, ’t', V, Kv, and B; therefore, they
will vary from joint to joint. The compensation used in any one joint should
be calculated given these lag break values.

A value of K of 0. 05 cis/psi has been found to be practical in valves,
and produces lag v breaks of the order shown in Figure 2 for the slave joints
examined for this study. The master joints are examined in the latter part
of this section.

Three Joints in Series - Unilateral

A signal flow diagram of three joints in series is shown in Figure 5.

The large number of inter-loop connections shows the complexity of the
system. Analysis of this system was achieved by considering only two
joints in series, and then extending to three by simulation on the analog
computer.

A block diagram of two joints in series is shown in Figure 6. The trans¬
fer functions F x (s), F a (s), H la (s), and Hj! (s) are frequency sensitive and
are derived by reduction of two joints from the signal flow diagram of
Figure 5. The a and 8 are constants and do not enter into stability calculations.

The loop of interest for examining stability is from F x through H la , F 3 ,
and Hgx back to F x . This loop has positive feedback which means that the
gain must be less than unity for all frequencies for stability to exist.

Therefore, the condition for a stable system is given by:

F x (j«i) H ls F 2 (juu) H^jtu) < 1 for all w (2)

Since the product in Equation 2 must be kept less than unity, it is
important that lags occur at as low a frequency as possible to keep the gain
going down as frequency increases. Of course, if the zero frequency gain
is greater than unity, stability cannot exist. In the Exoskeleton joints con¬
sidered, this zero frequency gain is always less than one.

5

In the transfer function H la (s) and (s), the terms are of the form:
1+Tj s

H(s) iq

G s
c

(3)

and F x (s) and F a (s) are, in simplified form:

1

F(s) =

(1 + T a sHi + Tds)

(4)

In general, T x is larger than T a and T3, and leads occur at a low fre¬
quency tending to increase the gain from Equation 2 as frequency increases.

The G (s) is the compensation network transfer function from Figure 1.
c

A typical frequency response plot for a proportional C^. (s) is shown in
Figure 7. Since the gain becomes greater than unity (dB), instability is
indicated. This was, in fact, observed on the analog computer simulation.

In addition of rate to G c (s) results in a frequency response plot typically
represented by Figure 8. Two additional lags are created because of the
lead in each of the G (s) terms since G (s) appears in the denominator of
Equation 3.

The addition of a lag term to reduce single joint bandwidth as indicated

in subsection, "Single Joint - Unilateral, " does not affect inter-loop stability

adversely since another zero is created in G (s) as shown by Equation 1.

In general, for inter-loop stability there should be one more zero than pole

in G (s). Hence, there is the necessity for rate feedback in all cases,
c

It is noted that the introduction of a lag to reduce single joint bandwidth
must be in parallel with the proportional and rate terms and not in series
with them. Otherwise, the additional lead is not created and inter-loop
instability will again result.

This analysis was extended to three joints ir series on the analog com¬
puter simulation, and the system was found to be stable.

Single Joint - Bilateral

A block diagram of a general bilateral servo is shown in Figure 9. The
open loop transfer function of loop 2 with loop 1 held fixed is given by:

G c (e) 0,(8) (5)

When loop 1 is free to move, the open loop transfer function of loop 2 is
given by

6

(6)

G (s) G a (s)

c ___

1 + G (s) Gi (s)
c

If G x (s) = G 3 (s) = G{s), then Equation 6 looks like the closed loop trans¬
fer function of a loop with forward gain of G(s) G (s) and unity feedback.
Replacing Figure 9 with this representation resufts in a block diagram as
in Figure 10. This servo has an open loop gain equal to twice that of the
unilateral servo, a change which can often be tolerated because only a 6 db
gain change is present.

With the frequency relationships in mind, the bilateral single joint case
can be designed using the concepts described for the unilateral servo as long
as the forward transfer function of both the master and slave loops are made
to be nearly equal. "Nearly" is a relative term, but it is the approach used,
and the computer simulation showed it to work.

A block diagram of a bilateral joint for the Exoskeleton is shown in
Figure 11. In the arm system the value of Y can be as much as 75 times
larger than ^ . For a given G (s), the unity Trequency gain of the master

loop would be as much as 75 times larger than that of the slave, and the
lag breaks in the master loop are quite different from the slave loop.

It is possible to sense velocity on the master and feedback, a strong
enough signal in parallel with ^ to make the sum of that signal and ¥ be
essentially equal to ♦ . While r ?his does not make the transfer function of
the master equal to that of the slave, it does bring them much closer to¬
gether, and the bilateral servo transfer function approaches that given by
Figure 10.

Simulation of the single bilateral joint on the analog computer showed
that this approach works for that case.

Three Joints in Series - Bilateral

A signal flow graph of three joints in series - bilateral is shown in
Figure 12. The complexity of the system is evident from the diagram.
Because of this complexity, analysis of the entire system uneconomic in
time; therefore, an approach is used which extrapolates the findings of the
unilateral three joints in series and the single joint bilateral cases.

An extensive analog computer simulation was set up for the bilateral
three joints in series to try various approaches. The computer diagram is
shown in Figure 13.

Using the approaches outlined above, i. e., compensation networks used
for the unilateral case and velocity feedback used in the bilateral single joint

7

case, it was found that the three joints in series - bilateral could indeed be
stabilized for the range of changed inertia called for in the design specifi¬
cations, 0 to 1500 pound load.

Computer runs shown have only proportional and rate components as
reflected in Figure 12, showing the signal flow graph. The system worked
very well for this case, but it must be kept in mind that mechanical res¬
onance in the actual system will possibly require the parallel lag circuit.

Computer runs showing operation of this simulation are shown in
Figure 14. Values of the compensation and feedback parameters to achieve
these results are shown on the signal flow graph. Figure 12.

It is noted that the values obtained by this study are not optimized. It
is probable that other values could be found which would improve perfor¬
mance. However, the sensitivity of the parameters was found to not be
critical when changes were tried on the computer.

Therefore, the feasibility of stabilizing three joints in series - bilat¬
eral has been shown and the approach to use in design is spelled out. Pro¬
visions should be made for adjusting the proportional, rate, lag, and veloc¬
ity feedback terms over a large range when building equipment.

RE COMMENDA TIONS

It has been shown that there is a method whereby the Exoskeleton arm
system can be made stable without being overly sensitive to parameter vari¬
ations. There are, however, other aspects of operation which should be
better understood such as dynamic compliance characteristics and actual
frequency response of a joint when connected in series with other joints.

During the actual arm design it is recommended that:

• Optimization runs be made with dynamic compliance constraints
to determine the best range of compensation for operation. Now
that the dynamic form of the compensation networks is specified,
optimization can be performed on the parameters. A suitable
digital computer program (ADA) is available for this analysis.

• Frequency response be run on the existing analog computer pro¬
gram to give more information on the system. Work with the com¬
puter simulations in conjunction with the planned hardware devel¬
opment will allow design to proceed with insight into operation pro¬
gressing at the same time.

8

Appendix I

DERIVATION OF SYSTEM DIAGRAMS

The dynamic behavior of the manipulator linkwork was analyzed to yield
terms in a form that lead directly to the construction of servo signal flow dia¬
grams, Figures 6 and 12. The diagram shows three bilateral servos with a
number of cross connections between them. Each of the horizontally dis¬
played patterns is in itself the diagram for a single joint; the cross conne¬
ction, which are displayed more or less vertically, represent the dynamic
interaction between joints in series as determined by the analysis.

The analysis was based on the link diagram Figure 15. The following
assumptions were made:

• The linkage rest position is a straight line.

• Excursions are small: sin 0 = P and cos 9 - 1,

• Centrifugal forces are negligible

None of the assumptions are true for the re*', device. The justification for
making them is that they greatly reduce the lauor of the derivation and sub¬
sequent stability analysis while not invalidating the stability analysis. This
is a customary and well established procedure for stability analysis which
is the purpose of this work; it is not satisfactory for predicting exact forces,
moments, etc. which, of course, was not intended here.

The dynamic analysis (which was done by JA. Bain of the Engineering
Mechanics Unit) proceeded to find force, acceleration equations (Newton's
first law equations), for the three links. Since it was convenient to have
torque and linear displacement as variables, the general form of these
equations is: :

T = Xy (7)

where t = torque (in lbs).

y = linear acceleration (in/sec? )

and X is a mass and link length term of dimensions: lbs. - sec 3 .

The results of deriving the various X terms are shown in Figure 16
and 17. Figure 16A is an extract from the signal flow diagram which iden¬
tifies the terms, and Figure 16B shows the diagram manipulated by block
diagram algebra to the same form as shown on Figure 12. Figure 17 shows
a table giving the formula for each X in terms of mechanical parameters.

9

The numerical values for these are shown in Figure 12. The distinc¬
tion between the heavy versus light case implies the load inertia (w ) is
present or absent on the slave. The operator's hand was always assumed
present on the master.

These are some other intercouplings as shown on the signal flow dia- .
gram. These feed from velocity (y) to flow. These are necessary to relate
the linear velocity to angular velocity at the joints, and thus to flow in the
actuators. The derivation follows from the geometry.

For small angles
9i = yi /r x

0 a = <y a - yi)/ r a - 6i

e 3 = <y 3 - y a )/r 3 - <e a + e,)
ex = y t / r x

e 8 = % /*, - <r- + fj-> ^

e a = yJ r 3 - <r7 + r7 ) y 3 + ^

Since actuator flow is related to angular velocity by the area and crank
radius of the actuator we have, finally

Qi = <yi / rj) \ rSi

qa = (^/ r 3 -(J-+_i_) 5^ ) Aars a

r x

% = $ 3 lr 3 + 7 _) * y 3 + y /r * ) A 3 rS 3

3 S

for the three actuator flow terms, which are then converted to paths on the
signal flow diagram.

10

_ j« aaram Figure 13, is made directly from the signal
_ u’singttte customary methods for simulation on

an^imlog 1 ^compute r*. Note that computer time is ten times real time to
present a slower moving display.

-Hrr H O^tartTlzgtemB Engineering Tools, Chapter 4, Wiley and
Sons, NYC,

11

Appendix II

COMPUTER RESULTS

The purpose of this work was to demonstrate the feasibility of operating
three bilateral servos on a series of three adjacent joints of a particular
design, namely the Exoskeleton arm. This was done by setting up the analog
computer according to the diagram Figure 13 and adjusting the available
parameters, loop gain and velocity feedback gain, until stable operation was
achieved at sufficiently high performance levels to satisfy the Exoskeleton
servo design requirements. Operation should, ideally, be satisfactory at
both heavy and light slave mass loading without the requirement to change
the gain parameters with load changes.

Satisfactory gain settings were found. The recorder traces Figure Hand
18 show the system operating stably while responding to a sinsusoidal torque
input to the wrist. The following table gives the system properties at these
settings:

Loop Gain

Compliance

Lead Corner Freq,

Wrist

3

65. 0 dB

1. 47%

27. 2 R/S

Elbow

2

56. 5'dB

1. 91%

15. 5 R/S

Shoulder

1

56.. 9 dB

1. 32%

16.1 R/S

The performance represented here is fully satisfactory for the Exoskeleton.

13

16

Figure 2. Typical Frequency Response for a Single Joint -- Proportional Gain Only

Figure 3. Typical Frequency Response for a Single Joint — Proportional and Rate Compensation

18

Figure 4. Typical Frequency Response for a Single Joint -- Proportional and Rate and Lag Compensation

Figure 5. Signal Flow Diagram for Three Joints in Series Tickler Control -- Unilateral

Figure 6. Two Joints in

20

22

Loop 1

Loop 2

Figure 9. General Bilateral Position Servo

G c (s) G(s)

Z

Figure 10. Block Diagram of Bilateral Servo with G 1 (s) = G a (s) = G(s)

24

For Computer Simulation
But Real Time

26

Input Torque
to Master Wrist

pr**-i4! f l.

fPw

Wrist-error Signal

Wrist-master Position

Wrist-slave Position

Elbow-error Signal

Shoulder-error Signal

Elbow-master Position

Shoulder-master Position

..Hr#** : f‘«4

Figure 14. Recorder Trace of Analog Simulation of
Exoskeleton Arm - Light Mass

/ V

: i . i 1

* v Ss\ / \ N

s

^ \ '

! !.■' 1

':t\

, • j •

A

i

L T *

•

||i;

* I

,

28

1

3

M (lb sec s

’/in)

r (in)

Master

Slave

Shoulder

1

19. 5

0.0145

0. 056

Elbow

2

10. 0

0.0202

0. 130

Wrist

3

13 0

0.0174

0.037

Load

s

0. 026

1. 94

Figure 15. Mechanical Parameters

29

*••'■**

Figure 16A. Partial Signal Flow Diagram Based on Derivation
of Dynamic Intercoupling Between Links

Figure 16B. Same Diagram Rearranged to More Convenient Form.

This is the Form Used in the Complete Signal Flow
Diagram (Figure 12)

30

1

1

. M, ^ M_.

r i { Y + Y’

2

r i 1 2 2 '

3

r, (^+ M a )

Figure 17.

Formula for Each X

Input Torque
to Master Wrist

Wrist-error Signal

Wrist-master Position

Wrist-slave Position

Elbow-error Signal

Shoulder-error Signal

Elbow-master Position

Shoulder-master Position

Figure 18. Recorder Trace of Analog Simulation of
Exoskeleton Arm - Heavy Mass

32

Unclassified

j DOCUMENT CONTROL DATA - R & D |j

| (Security classification of title, body of abstract and indexing annotation must be entered when the overall report 1a claeellled) j;

l. originating activity (Corporate author)

Specialty Materials Handling Products Operation
General Electric Company

1 River Road. Schenectady. New York

REPORT SECURITY CLASSIFICATION

Unclassified

26. GROUP

3 REPORT TITLE

Hardiman I Prototype Project

4 nr-coiprivE NOTES (Type of report and Indueive datee)

Special Technical Report on Joints in Series

8 au tmoriSI (Firel name, middle Initial, laet name)

Kendall E. Gilbert

Patrick C. Callan

r, PI PO‘U »A T E

10 June 1968

7«. TOTAL NO. OF PACES lb. NO. OF REFS

3

•a. contract or grant no.

Contract N00014-66-C0051

h. rifO j EC t NO.

Project No. IM62101050702

Contract Auth. Ident. No. NR196-049

d.

9a. ORIGINATOR*9 REPORT NUMBER!*) ]l

S-68-1081

96. OTHER REPORT NO(S) (Any other numbere that may be aealgned
thia report)

Government must have the approval of the Army Mobility Equipment Research and DevelopJ-
ment Center (Attn: SMEFB-HM), Fort Belvoir, Virginia 22060, or the EngineeringPsycholj-
ogy Branch of the Office of Naval Research (Code 455), Washington, D.C. 20360.

11 SUHPv n T A R Y notes

12. SPONSORING MILITARY ACTIVITY

Office of Naval Research and U. S. Army
Mobility Equipment Research and
Development Center_

M A P«. ' w A C '

Stable operation of unilateral and bilateral servos on a series of manipulator links
for Hardiman was analyzed and demonstrated on an analog computer.

DD

FORM

1473

Unclassified

Security Classification

Security Classification

Manipulator
Exoskeleton
Bilateral Servos
Joints in Series
Hardiman I

THIS hEPORT HAS SEEN DEllMITt.
rvHj> CLEARED FOR PU1LSC RElIASE

inter doc directive 5200.20 Af ,r
N ; HESTR'ct ions are imposed jpoi

it:. iSE and DISCLOSURE

J>/;,TK l BUT i ON STATEiiENT A

APPROVED -OR PU*L|C hE-EASF
.P/5TRIBUTiON JNLlMl t ED