TemPl is a planning system for reconfigurable multi-robot systems which is described in the PhD Thesis "Autonomous Operation of a Reconfigurable Multi-Robot System for Planetary Space Missions" (Roehr 2019) To solve the highly combinatorial optimization problem it combines the use of knowledge-based engineering, constraint-based programming and linear programming.
The focus of TemPl is to allow planning and optimization of (currently limited to medium-sized) missions of reconfigurable robotic agents.
TemPl comes with multiple utility programs such as a graphical interface to inspect solutions, a mission editor to design solutions (for those not wanting to deal with the XML format), and a mission generator to automatically generate simple as well as complex test missions.
Please use autoproj to bootstrap and include the package-set rock in your manifest. Afterwards you can install TemPl via:
amake planning/templ
This software remains under development, so if you plan to use it and experience difficulties, you are encouraged to contact the author of this library by opening an issue on the repository.
For testing the library Boost testing is used, along with the definition of a number of so-called test-scenarios, i.e. sample missions.
$> autoproj test enable planning/templ
$> amake planning/templ
$> cd planning/templ
$> ./build/test/templ-test --log_level=all
The folder test/data/scenarios contains a set of missions that cover/validate different aspects of the mission planner.
After building the library (and if following the instructions given in this document), the executables are typically available in the corresponding build/ folder and in PATH, so that you can easily access them via the 'templ-' prefix.
All executables support the '--help' argument to describe their current usage.
The actual planner: templ-transport_network_planner
allowed options:
--help describe arguments
--mission arg Path to the mission specification
--configuration arg Path to the search configuration file
--om arg IRI of the organization model (optional)
--min_solutions arg Minimum number of solutions (optional)
The default configuration for templ-transport_network_planner can be found under: test/data/configuration/default-configuration.xml, further details on configuration options can be found here.
Every run of the planner will result in the creation of a log folder under /tmp: for the overall mission a 'spec' subfolder will be created and successful epochs (also referred to as sessions) will be logged - identified by a counter. Each epoch folder contains solutions and results to intermediate planning steps. So check /tmp/_temp/ for the corresponding folder of the session that you would like to inspect.
$>./build/src/templ-transport_network_planner --mission test/data/scenarios/should_succeed/0.xml --min_solution 1
...
Session 0: remaining flaws: 0
Solution found:
# session id 0
# flaws: 0
# cost: 0
Saving stats in: /tmp/20201111_102602+0100-templ/search-statistics.log
Solution Search (epoch: 1)
was stopped (e.g. timeout): no
found # solutions: 1
minimum # requested: 1
TemPl:
# of solutions found: 1
Check log directory: /tmp/20201111_102602+0100-templ/
The graphical user interface templ-gui allows you to load solutions, so that you can inspect them and achieve a better understanding of the solutions. The graphical interface will also permit the creation of missions via the MissionEditor.
Details can be found here.
Solution analysis allows to recompute the cost analysis that is performed for each solution. The command line interface can compute and output the text format of the solution:
$>./build/src/templ-solution_analysis --mission test/data/scenarios/should_succeed/0.xml --solution /tmp/20201111_102602+0100-templ/0/final_solution_network.gexf --report
Processing: /tmp/20201111_102602+0100-templ/0/final_solution_network.gexf
Report:
SolutionAnalysis:
Resulting plan:
plan:
- role: AtomicAgent: Payload_0 (Payload)
Tuple:
a:
lander
b:
t0
roles (assigned):
Payload:
0
Sherpa:
0
roles (required):
Payload:
0
Sherpa:
0
Tuple:
a:
lander
b:
t1
roles (assigned):
Payload:
0
Sherpa:
0
roles (required):
Payload:
0
Sherpa:
0
Tuple:
a:
base1
b:
t2
roles (assigned):
Payload:
0
Sherpa:
0
roles (required):
Payload:
0
Tuple:
a:
base1
b:
t3
roles (assigned):
Payload:
0
roles (required):
Payload:
0
Tuple:
a:
base1
b:
t4
roles (assigned):
Payload:
0
Tuple:
a:
base1
b:
t5
roles (assigned):
Payload:
0
- role: AtomicAgent: Sherpa_0 (Sherpa)
Tuple:
a:
lander
b:
t0
roles (assigned):
Payload:
0
Sherpa:
0
roles (required):
Payload:
0
Sherpa:
0
Tuple:
a:
lander
b:
t1
roles (assigned):
Payload:
0
Sherpa:
0
roles (required):
Payload:
0
Sherpa:
0
Tuple:
a:
base1
b:
t2
roles (assigned):
Payload:
0
Sherpa:
0
roles (required):
Payload:
0
Tuple:
a:
lander
b:
t3
roles (assigned):
Sherpa:
0
Tuple:
a:
base2
b:
t4
roles (assigned):
Sherpa:
0
roles (required):
Sherpa:
0
Tuple:
a:
base2
b:
t5
roles (assigned):
Sherpa:
0
roles (required):
Sherpa:
0
Timepoints:
t0
t1
t2
t3
t4
t5
time horizon: 1.89795e+06
t0: 0
t1: 0
t2: 96452.1
t3: 194464
t4: 1.89795e+06
t5: 1.89795e+06
cost: 77.9083
efficacy: 1
efficiency: 76.4451
reconfiguration: 1560
safety: 0.463291
travel distance: 948193
# of agents: 2
# of mobile agents: 1
ModelPool:
http://www.rock-robotics.org/2014/01/om-schema#Payload : 1
http://www.rock-robotics.org/2014/01/om-schema#Sherpa : 1
If you wish to report bugs, please use the issue tracker at Github. If you would like to contribute to TemPl, open an issue or a pull request. Please assign the current maintainer to the reviewer list of the issue or pull request to ensure notification. The target branch for pull requests is the 'main' branch.
Releases will be made available as tags of the main branch.
The recommended workflow to add a new feature, add documentation, or fix a bug is the following:
- push your changes to a branch (e.g. feature/x, doc/y, or fix/z) of your fork this repository
- open a pull request to main branch
All new features need to be accompanied by a corresponding unit test.
If you use TemPl for a scientific publication, please cite the following paper:
@article{Roehr:2018:IBERAMIA,
author = {Roehr, Thomas M.},
doi = {10.4114/intartif.vol21iss62pp25-39},
journal = {Inteligencia Artificial, Revista Iberoamericana de Inteligencia Artificial},
mendeley-groups = {all},
number = {62},
pages = {25--39},
title = {{A Constraint-based Mission Planning Approach for Reconfigurable Multi-Robot Systems}},
volume = {21},
year = {2018}
}Copyright (c) 2015-2021 Thomas M. Roehr, DFKI GmbH Robotics Innovation Center