optimizing robotic systems with simscape
TRANSCRIPT
![Page 1: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/1.jpg)
1© 2015 The MathWorks, Inc.
Optimizing Robotic Systems
with Simscape
Veer Alakshendra
![Page 2: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/2.jpg)
2© 2018 The MathWorks, Inc.
Optimizing Robotic Systems with Simscape
![Page 3: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/3.jpg)
3
In this session
▪ Simscape and MATLAB enable
engineers to combine CAD models
with multidomain, dynamic simulation
MATLAB
![Page 4: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/4.jpg)
4
MATLAB
Computer
Vision
Real-Time
Testing
Image
Processing
Statistics &
Optimization
Control
Systems
Signal
Processing
Machine
Learning
In this session
▪ Simscape and MATLAB enable
engineers to combine CAD models
with multidomain, dynamic simulation
Simulink
StateflowSimscape
![Page 5: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/5.jpg)
5
In this session:
▪ Simscape and MATLAB enable
engineers to combine CAD models
with multidomain, dynamic simulation
▪ Results you can achieve:
1. Optimized mechatronic systems
2. Improved quality of overall system
3. Shortened development cycle
![Page 6: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/6.jpg)
6
Why Combine CAD and
Multidomain Dynamic Simulation?
▪ Fewer iterations on mechanical design
because requirements are refined
▪ Fewer mechanical prototypes
because mistakes are caught earlier
▪ Reduced system cost
because components are not oversized
▪ Less system downtime
because system is debugged
using virtual commissioning
![Page 7: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/7.jpg)
7
SimscapeOverview
▪ Enables physical modeling (acausal)
of multidomain physical systems
– Assemble a schematic
– Equations derived automatically
– Leverage MATLAB and Simulink
V+
V-
![Page 8: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/8.jpg)
8
Simscape MultibodyOverview
▪ Enables multibody simulation
of 3D mechanical systems
– Assemble bodies and joints
including import from CAD
– No need to derive and program equations
![Page 9: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/9.jpg)
9
Design Challenge
Challenge: Select motors and define
controls for robot and conveyor belts.
Solution: Import CAD model into
Simscape; use simulation to define
actuator requirements and control logic
System: 1. Import CAD Model
2. Determine Motor Requirements
3. Integrate Electrical Actuators
4. Minimize Power Consumption
5. Develop Control Logic
![Page 10: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/10.jpg)
10
System ModelSystem Model
Control Logic
Electrical Actuation
Conveyors
CAD Assembly
![Page 11: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/11.jpg)
11
Kuka Robot
▪ 5 degrees of freedom,
and a gripper
▪ Key advantage of Onshape:
Ability to directly define joints
– Exact mapping to constraints
used in multibody simulation
▪ System engineer
reuses mechanical design
in dynamic simulation
![Page 12: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/12.jpg)
12
1. Import Model from CAD
▪ Convert CAD assembly to
dynamic simulation model
for use within Simulink
– Mass, inertia, geometry, and joints
![Page 13: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/13.jpg)
13
Simscape Multibody CAD Import
▪ Import CAD assemblies
– Part definitions
– Converts mate definitions to joints
– SOLIDWORKS, Inventor, Onshape,
and PTC Creo® (Pro/ENGINEER®)
▪ Import CAD Parts
– CATIA, NX, SolidEdge, and others
– STEP files
![Page 14: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/14.jpg)
14
2. Determine Motor Requirements
▪ Define and run a set of tests
– Maximum payload, speed
– Worst case friction levels
– Full range of movement
▪ Use dynamic simulations
to calculate required torque
and bearing forces
▪ If design changes,
automatically rerun tests
and re-evaluate results
![Page 15: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/15.jpg)
15
3. Integrate Electrical Actuators
▪ Add motors, drive circuitry,
gears, and friction
▪ Choose motors based on
torque requirements
▪ Assign parameters
directly from data sheets
![Page 16: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/16.jpg)
16
4. Minimize Power Consumption
Challenge: Identify arm trajectory that
minimizes power consumption.
Solution: Use dynamic simulation to
calculate power consumption, and use
optimization algorithms to tune trajectory.
Model:
![Page 17: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/17.jpg)
17
…Computer Cluster
Workers
… …
Running simulations in parallel
speeds up your testing process.
Desktop System
Workers
… …
Simulation 1
Simulation 2
Accelerate Design Iterations
Using Parallel Computing
Robot Arm Model
This optimization task required
nearly 2000 simulations.
![Page 18: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/18.jpg)
18
5. Design Control Logic for
Arm and Conveyor Belts
▪ Sense quantities
within model that
govern system events
▪ Design logic
using a state chart
▪ Use outputs of logic
to control models
of system components
Belt Empty
Belt On
Box Waiting
Box on Belt?
Light blocked?
Curtain clear?
Control Logic
![Page 19: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/19.jpg)
19
Home
5. Design Control Logic for
Arm and Conveyor Belts
Belt
In
Belt
Out
Closed Grip
Open
State charts depend
on each other
![Page 20: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/20.jpg)
20
5. Design Control Logic for
Arm and Conveyor Belts
▪ Parallel state charts for
system-level logic
– Logic for arm depends upon
state of conveyor belt
▪ Define algorithm with flow charts,
state transition diagrams, truth tables,
and state transition tables
▪ Animation to observe behavior
and debug algorithm
![Page 21: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/21.jpg)
21
Test Production Control Software
▪ Automatically convert algorithms
to production code
– C Code, IEC 61131-3 Code
▪ Incrementally test the effect
of each conversion step
– Fixed-point math
– Latency on production controller
▪ Use the same plant model
– Test without expensive
hardware prototypes
Hardware-in-
the-Loop (HIL)
Processor-in-
the-Loop (PIL)
Convert to
C Code
Convert to
C Code
![Page 22: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/22.jpg)
23
How we did it
▪ Convert CAD assemblies
into dynamic simulation models
with Simscape Multibody
▪ Add electric actuators with Simscape
and control logic using Stateflow
▪ Optimize system using MATLAB
▪ Perform dynamic simulation
in Simulink
Simscape Multibody
Stateflow
CAD
MATLAB
Simscape
Simulink
![Page 23: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/23.jpg)
24
Summary
▪ Simscape and MATLAB enable
engineers to combine CAD models
with multidomain, dynamic simulation
▪ Results:
1. Optimized mechatronic systems
2. Improved quality of overall system
3. Shortened development cycle
![Page 24: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/24.jpg)
25
Modeling Physical Systems with Simscape
▪ Create models in various physical domains
▪ Combine Simulink and Simscape models
▪ Model energy transfer between different physical domains
▪ Create user-defined Simscape components
![Page 25: Optimizing Robotic Systems with Simscape](https://reader031.vdocuments.mx/reader031/viewer/2022012506/618131489c337a447f14d045/html5/thumbnails/25.jpg)
26
Thank YouEmail: [email protected]
LinkedIn: https://www.linkedin.com/in/veer-alakshendra-b5324222/