| Research Projects | OSCAR Version 2.0 : Performance Criteria Tutorials |
|
|
| Excercises |
Tutorial 1 - Introduction To Generate Options and Performance Criteria
Use the IK object RRIKJGenerateOptions to generate a null space for specified hand position
on the 10 DOF Fault Tolerant Manipulator. Implement Measure of Transmissibility and output
the criteria properties to the screen along with the null space solutions. Use MOT to choose
the optimal solution and send it to RoboWorks. Vary the joint perturb size to see how it effects
the null space and criterion values. Try setting different critical limits and normalization schemes
to see how the criteria properties are effected.
10 DOF Roboworks Model
PC Tutorial XML
Solution
Tutorial 2 - Performance Criteria on Simple Manipulators
Modify the RRIKJGenerateOptions object to handle a 4R Planar manipulator. Implement a criterion
and use it to choose the optimal solution. Vary the input and output space active/inactive coordinates
to achieve enhanced redundancy.
PC Tutorial XML
4R Planar RoboWorks Model
4R Planar Inertia Matrices
4R Planar Mass Data
Solution
Tutorial 3 - Performance Criteria Fusion
Implement a RRPCFusion object on the 10 DOF Fault Tolerant Manipulator. Use at least
two criteria. Output the criteria and fusion values to the screen and try using different
weights and critical limits. Pick the best solution and send it to RoboWorks
PC Tutorial XML
10 DOF Roboworks Model
10 DOF Inertia Matrices
10 DOF Mass Data
Solution
Tutorial 4 – Advanced Redundancy Resolution Techniques
Create RRPCFilter and RRPCFusion objects for the 10 DOF and combine them in a sequential filters
scheme. Set the critical limits appropriately.
PC Tutorial XML
10 DOF Inertia Matrices
10 DOF Mass Data
Solution
Tutorial 5 – Create a New Performance Criteria Class
Create a class that does some simple EEF obstacle avoidance and derive it from RRPerformanceCriteria.
The manipulator should attempt to follow a path specified by the user, but also avoid obstacles that
are in the path of the EEF. Implement this scheme on a Puma manipulator. The path is [75,-100,0]
to [75,100,0] and the obstacles should be moveable. There are many possible ways this application can
be setup but here are some hints on one approach.
Hints:
- To generate an EEF null space the RRPerturb class can be used.
- A criterion that attempts to minimize or maximize the distance of the EEF from a point
is the simplest implementation.
- Create a custom fusion scheme that ignores obstacles until the EEF reaches a threshold.
- It is helpful to locally normalize the criteria before fusion.
- A simple pass/fail technique cannot be used for the obstacles using RRPerturb. RRPerturb
generates a cube of options so the EEF will get stuck if the obstacle is directly in the
path.
PC Tutorial XML
Puma Roboworks Model
New Criteria Class Header Example
New Criteria Class Source Example
Application Example
|
|
