TAMPEST

Task and Motion Planning (TAMP) addresses finding solutions to automated planning problems that involve discrete actions executed by low-level continuous motions. This field is increasingly important in robotics, as it greatly enhances robot autonomy in real-world applications. Although various solutions and formulations exist, no standard representation has yet emerged. In this project, we propose a general, open-source framework for modeling and benchmarking TAMP problems. We introduce an innovative meta-technique for solving TAMP challenges involving moving agents and multiple task-state-dependent obstacles. This approach allows the use of any off-the-shelf task and motion planner, while leveraging geometric analysis to prune the task planner's search space, improving efficiency. Additionally, we specialize this meta-engine for an incremental SMT-based planner, TAMPEST (Task and Motion Planning by Encoding into Satisfiability Testing). Our framework includes a suite of benchmark problems of increasing complexity, where robots navigate environments with movable obstacles, and integrates the state-of-the-art PDDLStream solver.

Citation

E. Tosello, A. Valentini, A. Micheli. A Meta-Engine Framework for Interleaved Task and Motion Planning using Topological Refinements. ECAI, 2024 (to appear).

Info

The code has been tested on Ubuntu 22.04

Prerequisites

The first requirement is Python3 since the framework is written in python:

apt-get install python3-dev python3-pip

For the motion planning part OMPL is needed:

wget https://ompl.kavrakilab.org/install-ompl-ubuntu.sh
chmod u+x install-ompl-ubuntu.sh
./install-ompl-ubuntu.sh --python

Other Python requirements:

pip3 install -r requirements.txt

Requirements needed by the SMT-based task planner:

apt-get install libgmp3-dev
pysmt-install --msat --confirm-agreement

Requirements needed by the ENHSP task planner:

apt-get install openjdk-17-jdk

For the PDDLStream integration, PDDLStream is needed:

cd up-pddl-stream
git clone --recursive --branch main https://github.com/caelan/pddlstream.git
cd pddlstream
git submodule update --init --recursive
./downward/build.py

Install

Tempest

cd tempest
pip3 install .

Tampest

cd tampest
pip3 install .

Commands

1. To run a Tampest test the command is:

   python3 tampest/test.py --domain <domain-name> --dim <dim> --d <n> --c <m> [--capacity <robot_capacity>] --tp <task-planner-name> --mp <motion-planner> --tr <topological-refinement>
   

   The possible domains are: doors, maze, delivery, rover.

   The possible dimensions are: 2D, 3D. 3D setups are available only for maze and rover.

   The possible set of benchmark options are:

   • doors: --d [1 2 4 6 8 10] --c [0 1 2 3], with d the number of doors to be open and c = {0 = [(0,0)], 1 = [(10,0)], 2 = [(0,10)], 3 = [(5,5)]} the number of extra configurations sampled within the initially reachable space, the unreachable space, or equally splitted between both;

   • maze: --d [1 2 3 4 5 6 7 8 9 10] --c [0 1 2 3 4 5 6 7 8 9 10], with d the number of closed doors and c the number of locations to be visited;

   • deliery: --d [1 2 4 6 8 10] --c [0 2 4 ... 48] --capacity [4 3 2 1], with d the number of closed doors, capacity the loading capacity of the robot, and c the configuration of the parcels. In detail, c = n = [nr, ng, dr, dg], where nr is the total number of red parcels, ng is the total number of green parcels, dr is the number of red parcels already at their delivery stations, and dg is the number of green parcels already at their delivery stations;

   • rover: --d [2 4 6 8 10] --c [0 1 2 3 4], with d the number of samples (equally splitted between soils and rocks, each one separated from the robot by a closed door) and c the number of objectives to be photographed around the sample.

   The possible task planners are: tampest, tamer, enhsp, fast-downward.

   The possible motion planners are: LazyRRT, RRT.

   The possible topological refinements are: none, unreach, obs, all.

2. To run a PDDLStream test the command is:

   python3 up-pddl-stream/test.py --domain <domain-name> --dim <dim> --d <n> --c <m> --tp <task-planner-name> --mp <motion-planner>
   

   The possible domains are: doors, maze, rover.

   The possible dimensions are: 2D, 3D. 3D setups are available only for maze and rover.

   The possible set of benchmark options are:

   • doors: --d [1 2 4 6 8 10] --c [0 1 2 3]

   • maze: --d [1 2 3 4 5 6 7 8 9 10] --c [0 1 2 3 4 5 6 7 8 9 10]

   • rover: --d [2 4 6 8 10] --c [0 1 2 3 4]

   The possible task planners are: incremental, binding, focused, adaptive.

   The possible motion planners are: LazyRRT, RRT.