NEW! - available soon:
I am currently developing a new tool for motion planning researchers
called PPT, a
collection of C++ classes useful for developing, testing, and
comparing different algorithms and techniques for holonomic path
planning.
Why am I doing this?
It is currently difficult to compare motion planning algorithms for
the following reasons:
- There exists no standard set of models or planning queries for testing.
- Different algorithms are tested on different hardware platforms
(CPU, memory, OS).
- Researchers use different collision checking/distance
computation algorithms whose calculations typically have a large
impact on running time. In addition, collision-free path segments are
often verified in different ways (e.g. discretized sampling, recursive
bisection based on minimum distance, etc.)
- Source code for implementations of known algorithms is rarely
available.
By creating a set of planning algorithms, robot models, planning
queries, and collision-checking/distance computation libraries, I hope
to provide a standard software base for developing new motion planning
algorithms.
Who can use it?
Students, educators, researchers, and anyone else with a
non-profit interest in path planning techniques and algorithms.
My only request is that if you add additional functionality, or create
interesting models or example queries, that you contribute them back
to the PPT project archive. In this way, the usefulness of the
software can continue to evolve. For those interested in using
this library for business or other commercial purposes, commercial
licenses may be available by contacting me by phone or email.
When will it be available?
That depends on how much time I can devote to this project outside of
work during the coming months. If anyone is interested in helping me,
send me an email.
|