The University of Sydney Page 1William Reid | MATLAB Conference Sydney | 1
Using MATLAB/Simulink on Mars*
*+/- 401 million km
William Reid
May 24, 2016MATLAB Conference SydneyThe Museum of Applied Arts and Sciences, Sydney
The University of Sydney Page 2William Reid | MATLAB Conference Sydney | 2
• Engineer/Educator/Space Enthusiast
• The Mars Lab• The MAMMOTH Rover• MATLAB/Simulink• MAMMOTH Software• Other MAMMOTH adventures
• Engineer/Educator/Space Enthusiast
• The Mars Lab• The MAMMOTH Rover• MATLAB/Simulink• MAMMOTH Software• Other MAMMOTH adventures
Introduction/Outline
The University of Sydney Page 3William Reid | MATLAB Conference Sydney | 3
Schools
The University of Sydney Page 4William Reid | MATLAB Conference Sydney | 4
The University of Sydney Page 5
The University of Sydney Page 6
The University of Sydney Page 7William Reid | MATLAB Conference Sydney | 7
The MAMMOTH Rover
12 x 100 W Rotary
Actuators
12 x 100 W Rotary
Actuators
4 x 1.5 kNLinear Actuators
4 x 1.5 kNLinear Actuators
Mass = 75 kgMax Footprint: 1800 mm x 1800 mmMin Footprint: 650 mm x 650 mmMax Flat Plane Clearance: 800 mm
Mass = 75 kgMax Footprint: 1800 mm x 1800 mmMin Footprint: 650 mm x 650 mmMax Flat Plane Clearance: 800 mm
12 x 97 WhLi-Ion Batteries12 x 97 Wh
Li-Ion Batteries
Vision/IMU Fused Localization
Vision/IMU Fused Localization Front and Side
Haz CamsFront and Side
Haz Cams
4 x Leg Force
Sensors
4 x Leg Force
Sensors
RBG-D Sensor (Not shown)
RBG-D Sensor (Not shown)
Abs. Encoders for each actuator
Abs. Encoders for each actuator
The University of Sydney Page 8
Hip
Thigh
Ankle
Wheel
Body
Leg 4
Leg 3
Leg 2
Leg 1
The University of Sydney Page 9William Reid | MATLAB Conference Sydney | 9
MATLAB/Simulink Software Development
Algorithmic Development
Software-In-The-Loop
Hardware-In-The-Loop
C++ Generation for Deployment
The University of Sydney Page 10William Reid | MATLAB Conference Sydney | 10
Body pose vector:
Body velocity vector:
Wheel contact point velocity:
Hip and thigh joints:
Single leg kinematic model:
Actively Articulated Suspension – Algorithmic Development
The University of Sydney Page 11William Reid | MATLAB Conference Sydney | 11
Actively Articulated Suspension – Algorithmic DevelopmentSteering Joint Drive Joint Wheel contact-point velocity estimation
Kinematic equation reduction
Re-arrange
Solve for the joint actuation rates
The University of Sydney Page 12William Reid | MATLAB Conference Sydney | 12
Simulink Model
The University of Sydney Page 13William Reid | MATLAB Conference Sydney | 13
Simulink Model
Common interface for SIL and HIL blocks
SIL uses Simulink numerical ODE solver to run native simulation
SIL may be interfaced with external simulation applications:– Gazebo, V-REP
The University of Sydney Page 14
The University of Sydney Page 15William Reid | MATLAB Conference Sydney | 15
Simulink Model
Common interface for SIL and HIL blocks
HIL receives and transmits inputs and outputs over LAN
Robotics System Toolbox used to interface directly with ROS
The University of Sydney Page 16
The University of Sydney Page 17
RGB-D SensorRGB-D Sensor
Input Body VelocityInput Body Velocity
Joint Articulation
Rates
Joint Articulation
Rates
The University of Sydney Page 18
The University of Sydney Page 19William Reid | MATLAB Conference Sydney | 19
Analytical Tools
The University of Sydney Page 20William Reid | MATLAB Conference Sydney | 20
Other MAMMOTH Adventures – Digging Mission
MAMMOTH is used to dig a series of trenches along an operator-defined path
The Mawson Rover follows MAMMOTH to inspect each of the trenches
The University of Sydney Page 21William Reid | MATLAB Conference Sydney | 21
Multi-rover digging mission video
The University of Sydney Page 22William Reid | MATLAB Conference Sydney | 22
Digging Mission
MAMMOTHMawson
The University of Sydney Page 23
MAMMOTH the Farmer
The University of Sydney Page 24
The University of Sydney Page 25William Reid | MATLAB Conference Sydney | 25
• Novel wheel-on-leg platform• Model-based design for rapid
software development• Interfacing with various
sensors/actuators/software-in-the-loop applications
• Variety of applications
• Novel wheel-on-leg platform• Model-based design for rapid
software development• Interfacing with various
sensors/actuators/software-in-the-loop applications
• Variety of applications
Conclusions