Transcript Project OSCAR - ECpE Senior Design

Project OSCAR
Octagonal Speech-Controlled
Autonomous Robot
ONGO-01
Project OSCAR
Fall 2005

Client: Iowa State University
Department of Electrical and
Computer Engineering

EE Team Members





Faculty Advisor: Ralph E.
Patterson III






EE 492
EE 492
EE 491
EE 491
CprE Team Members

Presentation Date: December 6,
2005
Kevin Cantu
Jawad Haider
Robert Dunkin
Nicholas Hoch
Jeff Parent
Peter Gaughan
Andrew Levisay
Mike Mikulecky
CprE 492
CprE 491
CprE 491
CprE 491
ME Team Members



Lynn Tweed
Michael Snodgrass
David Brownmiller
ME 466
ME 466
ME 466
Project OSCAR
Presentation Overview



Initial Information
Project Introduction
Description of Activities







Tachometer
Software
End-effector construction
End-effector electronics
Documentation: Wiki
Jeff
Jeff
Jawad & Bob
Mike & Peter
Dave & Michael
Nick
Andy
Resources, Schedules, Summary Kevin
Closing
Jeff
Project OSCAR
List of Definitions












OSCAR
BasicX-24
CVS
Cybot
Drive train
Octagonal Speech-Controlled Autonomous Robot
Microcontroller used to interface with SONAR system
Concurrent versions system
The predecessor to OSCAR
The assembly of electrically controlled motion elements,
including the robot’s wheels, gears, belts, and
tachometers
End effector The electrically controlled mechanical arm and gripper
GUI
Graphical user interface
I/O
Input and output to a device
PEEL
Programmable Electrically Erasable Logic
SONAR
Sound navigation and ranging
Tachometer A device for indicating speed of rotation
Wiki
An internet based content management system for many
users
Project Introduction
Jeff Parent
Project Introduction
Problem Statement

General Problem
Develop a robot and perform demonstrations to
generate interest in the field and in the
department.

General Solution Approach
An ongoing project was started to design a
modular, autonomous robot which incorporates
speech control, sonar sensors, and an end
effector to interact with its surroundings and
audience.
Project Introduction
Operating Environment

Indoors

Flat surfaces, no
downward stairs
or drop-offs

Obstacles must
be 2.5 feet high
Project Introduction
Intended Users and Uses

Users



Project OSCAR team members
Supervised non-technical users
Use: Demonstration to raise interest in the field and
the department





Autonomous navigation of a hallway
Ability to pick up and place objects via the end effector
Ability to speak
Manual movement via wireless control software
Control via spoken commands
Project Introduction
Assumptions and Limitations

Assumptions
 Demonstrations last less than one hour
 Technical supervisors present during operation
 Operators speak English and are familiar with control software
 Remote PC for robot control has the appropriate software and
hardware

Limitations
 Software must run in Mandrake Linux
 Speech commands are issued less than 15 feet away
 Sonar range is 15 inches – 35 feet
 Wireless Ethernet within 328 feet
 Must fit through a standard 30-inch doorway
 End effector must fit within top module
Project Introduction
End Product & Deliverables

A robot with working systems






Power
Drive
Sensors
Software
End effector
Documentation
Tachometer
Jawad Haider
Bob Dunkin
Tachometer
Electromechanical Design

Problem




Interface of Motor Controller and Optical Encoder
Optical encoder outputs digital pulse train
Motor controller needs analog 5V with direction
Solution

Build a Wheel Tachometer circuit and interface the motor
and encoder
Computer
Serial
Motor
Controller
???
Optical
Encoder
Tachometer
Electromechanical Design
t=0
Channel A
Input voltage
Channel B
forward
+ 5.0 V
Channel B
backward
Optical encoder digital output
+ 2.5 V
Rotation
backward
forward
Needed analog signal
Optical Encoder
Tachometer
Proposed Design
Tachometer
Parts Used and Schematic
Optical encoder Channel A
Input From Phase Decoder
F-V
U1A
C2
1k
LM2907/DIP14
-
LM324
1n
R1
Output from F-V
-V CC
+V CC
D
S2
S1
IN
1k
A DG41 9
4
IN
OUT
3
3
1
V IN
V OUT
2
OUT
R1
2
1k
A DJ
LM7812C/TO220
-12V Source
U4+5V
1
IN
OUT
LM7805C/TO
-Voltage Source
Analog Out to Motor Controler
2
-
LM324
R1
LM317/CY L
R1
+5V Optical Encoder
R1
1
1k
-
LM324
3
R1
R1
1k
1k
1k
U1A
OUT
V-
1
+
U1A
11
U5 +Var iab le
+12V Source
V+
4
3
V+
U5 +12V
+
OUT
2
7
4
1k
1
-
1k
6
4
3
LM324
1
+
OUT
R1
2
-
1k
R1
R1
R1
1k
1k
1k
From 2.5V Regulator
3
+
R1
U5 SPDT Sw itch
1
8
2
R1
1k
3
V+
4
Output from Switch
1
V-
2
U1A
U1A
OUT
V+
+
R1
11
1k
3
9
12
V-
9
12
11
2
3
5
R1
V+
2
3
5
1n
8
V-
C1
8
4
1
4
10
11
11
1
4
10
11
V-

11

GND

2

GND

Switch: ADG419
Frequency-to-voltage converters: LM2907 and/or AD650KN
Phase decoder: LS7184 LSI sheet/LS7184 USD sheet
Op-amps: LM324
Charge pumps (providing negative voltage): ADM660
Adjustable voltage regulator: LM117
2

LM324
1
Tachometer
Accomplishments
Tested the phase decoder



We look at the UP/DN
output
Signal flips between +5V
and 0V with the change
in the direction of shaft
motion
Signal level stays there
until direction changes
again
Tachometer
Testing

Charge Pump



Two capacitors of 10uF are used for charge
storage
The voltage inversion operation is obtained using
ADM 660
Voltage Regulators

Two types of voltage regulators are used (5V and
12V)
Tachometer
Frequency to Voltage

LM 2907


Unknown chip malfunction
AD 650KN




MATLAB analysis
Ripple voltage too high
Used for higher frequency motors
Range (100Hz—1MHz)
Tachometer
Average and Ripple Voltage
Tachometer
Future




Need to put more research into chips
TC 9402 chips seems more feasible up to
100Hz
Design new circuit, with new chips
Create and test circuit components
Software
Mike Mikulecky
Peter Gaughan
Software
Past Accomplishments

Design process



Software controls hardware
Software extends in all directions to all levels
Main software system
Software
Software Languages




All ported to Linux
Java
Pearl
C#
Software
Current Problems

Code



Voice recognition
Documentation of code
Computer hardware



Inconsistent power supply performance
Defective power button
Motherboard battery dead
Software
Java



Improve Java code
Reorganize
Add support for debugging
Software
Prototyping



Rapid evaluation of ideas
Wireless motion control via Xbox controller
Prototyping framework
Software
Perl



Prototyping language
Flexible and fast
Modular
Software
Miscellaneous


New brain for OSCAR
No change in voice synthesis
Software
Future




Continue modularization of Java
Finish and extend prototyping framework
Use framework to test motion algorithms
Integrate better voice synthesis
End Effector
Mechanical
David Brownmiller
Michael Snodgrass
End Effector
Previous Design
3
1.
1
2.
3.
2
4
4.
Design was only 50%
Complete
Slide mechanism had
binding issues
Gears and motors
were not modeled to
scale
Structural issues on
wrist rotational motor
End Effector
Current Design
1
2
1.
2.
3.
4.
Remodel Gears and Motors
Design rotational joint to eliminate stress on the rotation motor
A completed arm with slide and base rotation for spring 06
Selected materials for structural integrity and aesthetics
End Effector
Current Status




Acquisition of materials
Physical manufacture of the arm
Manufacturing limitations on campus
Machine shop in Nevada
End Effector
Control
Nick Hoch
End Effector Control
Overview
Functionality
 Computer control for five motors in the new end
effector
 H-bridges for power
 Controlled by microcontroller(s)
 Communication with the PC
Goals
 To fully design the system
 To build the system without significant design
revisions
End Effector Control
Original Technology Selection




BasicX-24 top level
Multiplexers
LM629 motorcontrollers (1 per motor)
H-bridges (1 per motor)
End Effector Control
Questions




Too complex
Serial PC <-> BasicX
Serial BasicX <-> LM629
Skills requred: Java, Basic, LM629 codes,
hardware programming
End Effector Control
Possible Improvements




USB connection (PC <-> microprocessor)
Fewer parts (possibly only 1 microcontroller +
5 H-bridges)
More software, less hardware (faster
implementation)
C instead of BASIC as a primary language
(students have experience)
End Effector Control
Possible Solutions

LabVIEW board and software


PIC like the PIC18F4550


USB capable
Specialized PIC or a DSP chip like the
dsPIC30F4011



previously discarded because of PC and Linux issues
6 PWM outputs
1 optical encoder input
FPGA with programmed logic to replace entire
circuit.
Documentation
Andy Levisay
Documentation
Previous Problems




Incomplete
No central repository
Decision process not documented
Design and testing not well documented
Documentation
Solution: The OSCAR Wiki
•
•
•
•
•
Well organized
Carries from semester to semester
Easy sharing of documents and pictures
Also provides a place for making
announcements and meeting times
Useful in document collaboration
Documentation
The OSCAR Wiki
Documentation
The OSCAR Wiki
Documentation
Documentation Activities




Software
Tachometer testing
Sonar maintenance
End Effector
Documentation
Future Activities


Dedicated server for the WIKI
Adding more back data to the WIKI
Resources and
Summary
Kevin Cantu
Resources and Schedules: Fall 2005
Material Requirements

End effector





Software





Operating system – free
OSCAR PC – $10
Documentation


Structural materials, machining – donated
Motors – salvaged
Electronics – $99.90
Workstation PC - donated
Wiki – free, donated
Wiki PC – $10
Projected semester cost: ~$700
Actual semester cost: $119.90
Resources and Schedules: Fall 2005
Personnel Effort Requirements





Visitor demonstrations
End effector control circuit design
Tachometer implementation
Software
Documentation project
Senior Design reporting
160
140
120
100
Hours

80
60
40
20
Je
ff
Pa
re
nt
iku
le
ck
y
M
M
ik
e
Le
vi
sa
y
oc
h
H
ic
ho
la
s
N
H
ad
Ja
w
An
dr
ew
ai
de
r
n
au
gh
a
D
Bo
b
Pe
te
rG
an
tu
un
ki
n
0
C

Projected total hours: 1013
Actual hours: 622
Ke
vi
n

Resources and Schedules: Fall 2005
Financial Requirements

Fall 2005







Previous Semesters







Projected cost of materials: $700
Actual cost of materials: $119.90
Projected cost of labor at $10.50 per hour: $10,636.50
Actual cost of labor: $6,131.00
Fall 2005 Projected Total: $11,336.50
Fall 2005 Actual Total: $6,650.90
Spring 2005: $6,000-9,000
Fall 2004: $9,000-13,000
Spring 2004: $12,000
Fall 2003: $15,000
Spring 2002: $10,000-16,000
Fall 2001: $11,000-17,000
Estimated Overall Total, Spring 2001- Fall 2006: $115 thousand
Resources and Schedules: Fall 2005
Project Schedule
Project OSCAR: Summary
Lessons Learned

What went well


What did not go well



New team member orientation to complex system
Implementing tachometer design
Initial team progress: late start this semester
What technical knowledge was gained


Electronic, mechatronic and control systems
Linux software development
Project OSCAR: Summary
Lessons Learned

What non-technical knowledge was gained





Project management experience
Documentation methods, skills, and importance
Presentation skills
Interdisciplinary engineering interaction
What would be done differently


Better teaching of new team members
Better completed and organized documentation
Project OSCAR: Summary
Risks and Risk Management

Anticipated potential risks





Part ordering delays
Documentation problems
Personal injury
Loss of a member
Anticipated risks encountered


Part ordering delays
Documentation problems
Project OSCAR: Summary
Risks and Risk Management

Unanticipated risks encountered




Long term loss of faculty advisor
Software malfunction
Lost knowledge
Resultant changes in risk management


More sophisticated documentation
Emphasis on shared knowledge
Closing
Jeff Parent
Project OSCAR: Summary
Closing

Still in overall implementation stage – autonomy is
incomplete

Continued demonstrations have been effective in
developing team member abilities

Future should involve


Finalizing OSCAR system
Satisfying department needs through further robot
development projects
Project OSCAR
Questions?
http://seniord.ee.iastate.edu/ongo01