Cool Robot Mechanical Design of a Solar-Powered Antarctic Robot

1
Cool RobotMechanical Design of a Solar-Powered
Antarctic Robot

• Alex Price
• Advisor Dr. Laura Ray
• Thayer School of Engineering at Dartmouth College

2
Project Goals

• Traverse the Antarctic south polar plateau
  autonomously on renewable energy
• Relatively cheap (about 20,000)
• Travel 500 kilometers in 2 weeks
• Easy to handle, transport, and maintain
• As lightweight as possible (also for energy
  reasons)
• Small enough to fit inside the Twin Otter
  aircraft.
• Easily assembled and tested after delivery
• Scientific instruments easily added and integrated

3
Antarctic Plateau

• Large central flat plateau
• High altitude (2800 meters)
• Cold (-20 to -40 C in summer)
• Dry and sunny, but windy
• Firm, clean snow
• Flat, but with wind-sculpted sastrugi snow
  drifts
• Possible Robot Missions
• Automated distributed sensing
• Magnetometers
• Ionosphere studies
• Ground-penetrating Radar
• Traverse team support
• Ecological Studies

Sastrugi
4
Specifications and Solutions

• Specifications
• Average Speed of 0.4 m/s, top speed at least
  twice that
• Maximum dimensions to fit in Otter
• 1.5 m long
• 1.2 m wide
• 1.2 m tall
• Less than 75 kg empty 15 kg payload capacity.
• Maximum ground pressure of 3 psi
• Design to achieve those goals
• Specialized lightweight construction
• Optimized dimensions
• Careful component selection (tires, bearings,
  etc.)
• Custom wheels, hubs, and drive train components

5
Overall Robot Design

• Solar panels attached over chassisand wheels by
  support arms
• Tube on top of chassis box may be required to
  support center of top panel
• Insulation is likely not required.

6
Solar Power in the Antarctic

• In summer, sun never sets, but is always at a low
  angle
• Sun is brighter in high, dry climate
• As bright as 1200 W/m2 on a clear day
• Few cloudy days in the central plateau
• Significant reflected light from snowfield
• Proportional to sun azimuth
• Snow albedo of as high as 0.95
• Diffuse component of insolation as large as 100
  W/m2 from atmospheric scattering
• Sunny day insolation fairly constant, but
  scattering and cloud cover varies with the time
  of year.

Variation in azimuth between max and min
decreases to zero at 90, at the pole.
7
Solar Power in the Antarctic
Top (direct sun only) 34
Back (in shadow) 11
Front 128
Sides 34 (reflected light only)

• Available Power in Average Summer Sun
• 1000 W/m2 of solar power available on an average
  sunny day
• Sun azimuth angle 20 from horizon (average for
  November-February)
• Robot facing front towards sun (worst case)
  Snow albedo 90

Panel capacities are based on nominal 1-sun (1000
W/m2) input 100 200 W/m2 energy output (20
efficient cell in direct sun)
8
Scaling Capability

• Design can be scaled well to a variety of sizes
  for different mission goals.

9
Tire Selection
ATV tires
Custom cut tire
Russian Snow Bug tire
Apollo 17 rover mesh wheel
Roleez ballon tire
Mars Rover solid wheel

• Ideal tire would be lightweight and would have
  good traction, low ground pressure, and low
  rolling resistance but no such tires are
  available within budget.

10
Tire Selection

• Best tire of available selection was Carlisles
  16x6-8 knobby ATV tire
• About 6.5 pounds, very stiff, good tread pattern

11
Wheel Design
ITP aluminum
Carlisle steel standard
1st design iteration

• Commercially available wheel options are not
  suitable.
• Aluminum racing wheels are all too large
• Available 8x5.5 wheels are too heavy (gt 2.3 kg)
• Require the use of heavy bolts and hubs
• Thus, a custom wheel had to be designed to meet
  the requirements of the design

12
Wheel Design

• Factor of Safety of 3 against static failure in
  worst-case loading
• Factor of Safety of at least 2 against fatigue
  failure in worst-case driving conditions
• Only 0.9 kg, and uses smaller bolts hub
• Tubeless if 2 halves are sealed

13
Hub Design

• Standard 4-inch bolt circle
• Welds to drive shaft, bolts to wheel tabs
• Factor of safety of at least 2.5 against fatigue
  failure in worst-case loading

14
Assembled Wheels

• Wheel hub nuts and bolts 1.1 kg
• Far better than the commercially available 3 kg
• Total assembly (with tire and covers) 4 kg
• Total weight savings on robot 8 to 9 kg

15
Drive Train
Option 2 Bearing pair to carry load, motor
mounted loosely so bearings will support the
bending loads.
Option 1 Cantilevered support tube with
press-fit bearing, minimizes loads on gearbox.

• Very efficient motor and gearbox
• Custom hollow aluminum shaft and supports

16
Integration and Assembly

• Heaviest components mounted in the center
• Motors, controllers, power electronics, and
  scientific instruments mounted symmetrically on
  chassis

17
Future Plans and Goals

• Complete Design and Test Components
• Wheels and Hubs NC machined
• Drive Train design completion
• Assemble and test drive train
• Assemble and test solar panels
• July - Chassis operational on batteries
• August - Solar power systems tested and
  operational
• September - Robot operational on solar power
• Next year - Testing in Greenland and in
  Antarctica!

18
Conclusions

• Design has been optimized within the strict
  parameters
• Robot should easily meet the mission goals
• Future versions could be lighter and faster.

• Autonomous navigation at the south pole is a
  daunting task, but we are well on our way to
  achieving that goal.
• Building a robot is a lot of work, but has been
  and will continue to be a great experience.

19
Acknowledgements

• Laura Ray
• Alex Streeter
• ENGS 190/290 group
• Guido Gravenkötter
• Gunnar Hamann
• Mike Ibey
• Kevin Baron
• Pete Fontaine
• Leonard Parker
• Paula Berg
• Cathy Follensbee

• Jim Lever
• Dan Denton
• CRREL
• Marc Lessard
• Gus Moore 99
• Michael at Wilson Tire
• Don Kishi at Carlisle Tire
• National Science Foundation
• Everyone at Thayer School who has made this
  possible

• Full reference and bibliography information is
  included in the report.