Transcript Final Presentation (4/22, 4/24) By Group A3: HMS, Inc

Critical Design Review
Group A3: HMS “Bloodhound”
Matt Morris
Matt Prygoski
Jay Slaggert
Fred Thwaites
Dan Zibton
Tuesday, April 22, 2008
Thursday, April 24, 2008
Project Overview
• Autonomous Hazardous Material Seek-and-Identify
Vehicle
• Use GPS to locate general area of potentially
hazardous material
• Use infrared sensor to detect if material is hazardous
• Return to starting position when detection is complete
Key Features
• Independently Driven
Rear Wheels
– Two 24V Gear Motors
• Front Wheel Casters
• HDPE body
• PVC Boom with Infrared
Sensors
• GPS Navigation System
Abstract
• Purpose: to design a hazardous material seek and
identify vehicle that is capable of autonomously
locating, sampling, and identifying a target that could
possibly be hazardous material
•Important Goal: Simple and Durable
• Target customers:
• Military
• Law enforcement agencies
• HazMat clean-up crews
Design Requirements
• Powered by at most 20V battery
• Move to a location within +/- 1m specified
by provided coordinates
• Travel on grass, dirt, concrete, and asphalt
• Body no larger than 2m in any direction
• Complete search in under 5 minutes
Objectives of Drive System
• Accurate
– Capable of moving to with in +/- 1 m of a target.
• Robust
– Capable of movement over all-terrain
• Maneuverable
– Easily moves in small spaces of 20 square meters.
• Simple
– Minimal parts for easy implementation.
Option 1
Option 2
Option 3
Option 3 was Chosen
Key Advantages Identified in Option 3:
• Zero Degree Turning Radius
• Minimal Wheel Slip
Possible Technical Challenges:
• Account for Minimal Wheel Slip
• Provide Enough Power to Rear Wheels
• Implementation of a Robust Drive System
Technical Challenges
• Control Wheel Slip
–
–
–
–
Position CoG close to Rear Wheels
Acceleration should not exceed a = μsg/2
μs,min,= 0.7
amax =3.5 m/s2
• Provide Enough Power
– ω = 225 rpm
– T = 1 N-m
Robust Drive System
Prototype Considerations
Key Aspects
• Zero degree turning radius capability allows
vehicle to traverse search path
• Sensors in front of vehicle
• Ability to guide a vehicle with GPS
• Direct drive would be simple to control with
feedback loop
Materials
• Purchased large pneumatic casters and
rubber wheels similar to final design
• HDPE used for body – strong, lightweight,
machinable, and cheap
• PVC used for boom
– Strong and stiff
– Easily reconfigurable
Fabrication Limitations
• Rear wheels should have better tread but
the hub was useful
• L-Brackets are not ideal
• Center of gravity should be moved toward
rear of vehicle
– Reduces caster drag
– Increases drive traction
Next Generation Prototype
• Use more powerful motors and battery
– Determine caster performance at high speed
– Analyze repeatability of maneuvers (with a focus on caster
performance)
• Explore better designs for shouldering and
fastening body panels
Circuit Diagram
Circuit Diagram
• 12V Motors, 355 ozin torque failed due to
unpredictable dynamic
forces
• 24V Motors without
specs used in place
• Determined
feasible based on
in-robot testing
Circuit Diagram
• MOSFET transistors
used as a gate to close
circuit, allowing the
motors access to
ground
• Can handle 100V, 12A
Vehicle Operation
• Record Start and Hazard Location
(Sample the GPS Receiver)
GPS Uncertainty
3
Latitude (m)
2.5
2
2.7 m
1.5
1
0.5
0
0
0.5
1
1.5
2
2.5
Longitude (m)
3
3.5
4
4.5
Vehicle Operation
• Two Phases:
- Traverse. Use GPS to guide vehicle within 10m
of hazard
- Search Algorithm. Systematically search area
around hazard
Traversing
• Considerations:
- Speed. Travel fast but allow GPS to acquire
accurate readings
- Position/Orientation. Need to know position
and orientation
- Can’t Rely on Angles. GPS error too great to
accurately calculate angle of orientation
GPS Error vs. Vehicle Speed
Minimum Error at
2.6 m/s
At 1.8 m/s
Only 6% greater error
Traversing
N
• Drives forward
• Record location
• Pivots and reorients if
x or y distance to hazard
is increasing
W
H
• Stops when within 10m
of hazard
S
Search Algorithm
- Square Spiral
- Never necessary to
retrace path
- Covers greater than
90% of area
- 40% faster than back
and forth pattern
Results - Traversing
Results – Search Algorithm
Results
Part 2
• Vehicle can travel at 1.8 m/s (4 mph)
• Video:
• When starting in wrong direction, ‘Bloodhound’ can
correct course
• Locate hazardous material 30-40 m away in less
then two minutes
• Stop within 2.7 m of original target
Feasibility / Recommendations
• Proves GPS technology is feasible for the use
• Without extra technology, reckoning system is crude
but effective
• Electronic compass should be investigated for future
models to refine drive system
• Casters / Rear Driven wheels allow easy navigation
• Motors for this type of application are readily
available
• Both possible motors provided enough torque
• Higher quality motors available to increase speed to
2.0+ m/s
• Stationary boom adequate to search large area
Lessons Learned
• Compass can be combined with GPS to
create useful and accurate system. GPS
on it’s own can be very precise but is not
as repeatable
• Infrared sensors can pick up false
positives based on reflections of the sun
or other objects, including the IR emitter
even when the emitter is out of range
Thank you.
Questions?