casey dill – team lead arthur connors andrew torkelsonedge.rit.edu/content/p09029/public/poster...

Casey Dill – Team LeadArthur Connors

Andrew Torkelson

Introduction

� Fourth project in artificial limb track� First three projects focused on hand

DOF

Project Goals

� To create a computer simulation of the kinematics of an air muscle-controlled human joint (the elbow)

� Build a prototype joint to compare to and improve the computer model

� Make end product adaptable and useful for future � Make end product adaptable and useful for future iterations

Customer Needs

Engineering Specs

Team Roles

� Casey� Math Theory� Pneumatics� Arm Design� Arm

� Arthur� Controls

System� Electronics� Simulation

� Andrew• Test Stand

Design• Test Stand

Fabrication� Arm

Fabrication� EDGE

� Simulation� Integration

Fabrication• Testing• Data Analysis

System Architecture

System Architecure

Physical Prototype

ComponentsAir Muscle

Artificial Limb

Test Stand Relays

Elbow joint moved by

pneumatic actuators

Holds arm and sensors

in place Provide power to

Pressure Gauge

w/ Flow Restrictor

Measures and/or slows

airflow

Solenoid Vacuum

Pump

Air Tank

Strain Gauge

in place

Controls airflowSucks air out of

muscles

Provide power to

and control

solenoid

2 count: one holds

pressurized air, one

is vacuumed

Produces voltage

change with

displacement

Math Model

( ) ( )t

t

e σ

0muscle

TankTank0muscle

28.PP

:Hold

PPPP

:Fill

−=

+−=

Calculating Air Pressure1) LabView first calculates what

air pressure should be in the air muscle given the amount of time that has passed.

2) LabView calculates the force the air muscles should be exerting based on the

( )te σ0muscle

0muscle

PP

:Drain

=

( )

−

−+−= 2muscle

2muscle2

sin

1sin2P1cos3

4

Pkk tDt

DF

θθπθπ

Apply Calculation of Force Out of McKibben Air Muscles*

*”Measurement and Modeling of McKibben Pneumatic Artificial Muscles” by Ching-Ping Chou and Blake Hannaford

exerting based on the pressure in the air muscle given a few characteristics of the air muscle.

Testing

� Components� Pressure Tests� Elastic Cord/Strain Gauge

Components

� Air muscles – contract the way they should

� Relays and Solenoid – Power works� Vacuum Pump – Failed to power up; � Vacuum Pump – Failed to power up;

worked after rewiring

Pressure Tests

� Muscles filled with air at different increments (0.05, 0.125, 0.25 seconds)

� Data fitted to function

Test Data

60

70

80

90

Test Results

-10

0

10

20

30

40

50

0 0.5 1 1.5 2 2.5 3 3.5

Pre

ssu

re

Time

Test 1Test 2Test 3Test 1

Fitted Curve

60.0

70.0

80.0

90.0

Case Fill

Fitted

Theory

0.0

10.0

20.0

30.0

40.0

50.0

60.0

0.00 0.50 1.00 1.50 2.00 2.50 3.00

Pre

ssu

re (

psi

)

Time (sec)

Elastic Cord/Strain Gauge

� 6 strings, different sizes� Weights 50g-2kg added� Displacement measured� Linear Regression to find k

� Cord tied onto gauge� Cord displaced set amounts,

voltage measured� LRA to find Constant

Scatterplot of % Deformation vs Weight (lbs)

0.12150

Scatterplot of Voltage vs Force

Elastic Cordk≈0.212 lbs/in

Strain GaugeConstant≈2347 lbs/Volt

543210

1.0

0.8

0.6

0.4

0.2

0.0

Weight (lbs)

% Deformation

2.52.01.51.00.5

0.12150

0.12125

0.12100

0.12075

0.12050

Force

Voltage

Simulation

Communication is Key!!!

� Mechatronics Toolkit is a linking tool between LabVIEW and COSMOS Motion

� Each program receives, translates, and � Each program receives, translates, and passes data to the next program

Lessons Learned

•Using a single 3-position, 5-way solenoid is cheaper than four 2-way, 2-position solenoids, and takes half of the relays.

•Find local suppliers:•Find local suppliers:•Roessel has many pneumatic parts on hand.•Cross Bros has sprockets and chain on hand.

•Don’t rely on non-team members for mission critical parts.

•Start programming earlier.

Future Work/Recommendations

� Expanding to the shoulder� Refining and generalizing models� Scaling� Don’t reinvent the wheel� Don’t reinvent the wheel� Rely on past BOMs� Use updated version of SolidWorks