Upload
maurice-martin
View
213
Download
0
Tags:
Embed Size (px)
Citation preview
Life in the Atacama Carnegie Mellon
Hyperion Mobility Testing
July 28, 2003Dimi Apostolopoulos
Michael WagnerKevin Peterson
James TezaStuart Heys
Life in the Atacama Carnegie Mellon
Mobility Characterization Project
Atacama, April 2003
Hard Surface, July 2003
Sand Simulant, July 2003
Life in the Atacama Carnegie Mellon
Mobility Approach
Atacama 2003 Hard Surface Soil Simulant
Drive/Steer/ClimbTorque/Power/Energy
Evaluate Performanceand Rethink Design-Locomotion-Mechanical Driveline-Traction/Steering Control-Mobility Sensing-Vehicle Electronics -Overall configuration-Payload accommodation
Life in the Atacama Carnegie Mellon
Weighing Hyperion
Scale placed under each wheelTotal weight in desert: 181.1
kgLab experiments: 140 kg
Hyperion
39.78 kg
40.00 kg
51.54 kg
49.78 kg
Life in the Atacama Carnegie Mellon
Mobility Studies
Driving
Steering
Slope Grading
Discrete Obstacle Climbing
Drawbar Pull
Combined Feature Negotiation
Life in the Atacama Carnegie Mellon
Measured & Computed Variables
Power is measured into amplifiers
Torque is calculated through the following equation:
T =
Torque measurements are most accurate when robot driving straight and at higher velocities
Compute consumed locomotion power as the product of resistive force (T/r) and vehicle speed
Compute losses due to soil work
I Viv / r
= 0.35-0.45 (total efficiency)
Life in the Atacama Carnegie Mellon
Atacama Flat Ground – Power
Peaks occur near moment of ~10 deg
pitch
Life in the Atacama Carnegie Mellon
Atacama Flat Ground – Torque
15 Nm
20 Nm
Life in the Atacama Carnegie Mellon
Atacama Flat Ground – Rolling Resistance
Life in the Atacama Carnegie Mellon
Flat Concrete Floor – Power
Life in the Atacama Carnegie Mellon
Flat Concrete Floor – Torque
Life in the Atacama Carnegie Mellon
Sandbox – Power
Life in the Atacama Carnegie Mellon
Sandbox – Torque
Life in the Atacama Carnegie Mellon
Flat Ground – Preliminary Summary
Max wheel power, considering only flat portions of terrain
Concrete floor: 40 W (rear right wheel)
Sandbox: 35 W (right wheels)
Atacama: 45 W (front right wheel)
Average wheel power
Concrete floor: 15 W
Sandbox: 20 W
Atacama: 30 W
Life in the Atacama Carnegie Mellon
Slope Climbing in the Atacama
Life in the Atacama Carnegie Mellon
Slope Climbing in the Atacama: 14 deg
25 Nm
35 Nm
Life in the Atacama Carnegie Mellon
Slope Climb Testing in Lab
Life in the Atacama Carnegie Mellon
Slope Climbing in the Lab: 30 deg
Entire robotclimbing ramp
Entire robotclimbing ramp
50 Nm
65 Nm
Life in the Atacama Carnegie Mellon
Slope Climbing – Preliminary SummaryMax wheel power
Ramp: 85 W (30-deg slope)
Atacama: 70 W (14-deg slope)
Average wheel power
Ramp: 70 W (30-deg slope)
Atacama: 50 W (14-deg slope)
Life in the Atacama Carnegie Mellon
Obstacle Climbing
Life in the Atacama Carnegie Mellon
Reconstructing Terrain Features
Twist = steering roll – body roll
Robot width x sin(twist) = height of object
Can be used to quantify surface roughness
Negative twist:Front left wheel
climbing
Positive twist:Rear left wheel
climbing
Vehicle length
Height: 20 cm
Life in the Atacama Carnegie Mellon
20-cm Obstacle on Concrete Floor
ba dc
a. Driving forward, front left wheel climbs obstacle
b. Driving forward, rear left wheel climbs obstacle
c. Driving backward, rear left wheel climbs obstacle
d. Driving backward, front left wheel climbs obstacle
Life in the Atacama Carnegie Mellon
20-cm Obstacle in Sandbox
a. Driving forward, front right wheel climbs obstacle
b. Driving forward, rear right wheel climbs obstacle
c. Driving backward, rear right wheel climbs obstacle
d. Driving backward, front right wheel climbs obstacle
ba dc
Life in the Atacama Carnegie Mellon
12-cm Obstacle in Atacama
Life in the Atacama Carnegie Mellon
12-cm Obstacle in the Atacama
a. Robot’s front left wheel climbs obstacle
b. Robot stops, fails to climb obstacle
c. After a second command, rear left wheel climbs obstacle
ba c
Life in the Atacama Carnegie Mellon
Obstacle Climbing - Preliminary SummaryMax instantaneous wheel power
Lab: 150 W (20-cm block)
Atacama: 120 W (14-cm rock)
Life in the Atacama Carnegie Mellon
Drawbar Pull Tests
Drawbar pull is the force a vehicle can pull on a given soil
The drawbar pull is measured by attaching the robot to a load cell and steel cableThe robot is driven until the cable is tensioned and its wheels begin to slip
The drawbar pull is the maximum force sensed by the load cell
Direction of travel
Steel cable
Load cell
Wall
Life in the Atacama Carnegie Mellon
Drawbar Pull Tests
Drawbar pull is a useful metric because it can be used to find the maximum climbable slope for a given soil type:
Max slope = atan(DP / weight)Performed tests in cohesionless sand
This soil provides very little traction, similar to regions of loose sand seen in the Atacama
Life in the Atacama Carnegie Mellon
Drawbar Pull Tests – Results
c
a
b
d
e
a. Robot driving normally
b. Cable tension rapidly increases
c. Wheels slipping
d. Motion controller fault, at least one wheel stops servoing
e. Robot reverses, cable goes slack
Max DP: 550 NMax slope: 22 deg
Life in the Atacama Carnegie Mellon
Discussion (1)
Key total locomotion power results
Driving power: 100-150 W
Steering: ~1.5 x Driving
Slope: 250-350 W to climb 20-25 sandy slope
Slope continuous; battery thermal limit
Obstacle: 150-200 W per wheel
Multiple obstacles is the worst case
Life in the Atacama Carnegie Mellon
Discussion (2)
Poor slope climbing and need for traction optimization motivate new locomotion system
Need for better and more payload accommodation, and new solar array layout motivate new chassis configuration
Must devise more precise/repeatable mobility to aid close-up science.