YanielCarreno / dhrta

Task Allocation Strategy to distribute goals (in a decentralised manner) in a heterogeneous fleet considering, goals makespan, goals position, robot capabilities.

DHRTA Strategy repository

This repository contains the ROS-DHRTA code, the domain and problems used to prove our task allocation strategy:

Important: This is a simplified version of the algorithm. If you need the full version, please contant the authors.

The implementation is devided in two parts:

1. Goals Distribution

1.1 Compilation:

Select or create a catkin workspace:

     mkdir -p dhrta_ws/src
     cd dhrta_ws/

Get the code:

     cd src/
     git clone https://github.com/YanielCarreno/DHRTA.git

Compile everything:

     catkin build

1.2 To run an example:

     roslaunch dhrta dhrta_mission_execution.launch

1.3 The results:

The allocation_solution.txt file located in the dhrta package shows the allocation's results. The file contains all the goalas and the robot in charge of executing them. For instance, (poi_rock_analysis wp42)[rexrov0] defines the goal (written in PDDL format) is executed by robot rexrov0. These results can be added to the PDDL problem files to generate plans using the steps in the Task Planning section.

2. Task Planning

• The folder "planning_file" shows an example of the PDDL domain and problems used for one of the robots to executed the its part in the whole mission. The DHRTA strategy defines the goals the robot can implement and they are introduced in the robot's problem file using a framework to connect our algorithm and the robot's knowledge base which populate the problem file.

• If the DHRTA is not on place, benchmarh planners generate the task decomposition considering the definition of predicates associated with its capabilities which restrics the implentation to robots that present the capability. The goals are then distributed based on the AI planners task decomposition.

• The "dhrta_example" folder constains the results of DHRTA+TP. The set of goals is decomposed in subsets which are allocated to the problem files of 5 auvs.

  How to run it?

   --> Clone the repository
   --> Compile the planners:
   
            1) TemporAl, TFLAP & OPTIC IPC-2018 (https://ipc2018-temporal.bitbucket.io/)
   
            2) LPG (http://zeus.ing.unibs.it/lpg/) 
            
            3) TFD (http://gki.informatik.uni-freiburg.de/tools/tfd/)
            
            4) POPF (https://nms.kcl.ac.uk/planning/software/popf.html)
            
            
     --> Run the planners using the domain and problem required
  

Publications

If you are using this work for your research, please cite:

@InProceedings{carreno2020a,
  Title  = {A Decentralised Strategy for Heterogeneous AUV Missions via Goal Distribution and Temporal Planning},
  Author = {Carreno, Yaniel and Pairet, {\`E}ric and Petillot, Yvan and Petrick, Ronald PA},
  Booktitle = {Thirteenth International Conference on Automated Planning and Scheduling},
  Year = {2020}
}