ManipulationPlanning-SI-RRT

Combination of Rapidly-Exporing Random Trees (RRT) and Safe Interval Path Planning (SIPP) for high-DOF planning in dynamic environments, i.e. planning a path for a manipulator when moving obstacles are present (and their trajectories are know/accurately predicted)

[Paper]

MVP. Updates will be soon.

Quick start

Download dataset https://disk.yandex.ru/d/-73LOGO5kOSYuA and extract ./tests directory in ./mass_test/tests

Set number of parallel jobs in ./mass_test/do_mass_test.py NUM_CPUS. best practice - set NUM_CPUS as number of cores in CPU - 1

Download docker container, remove tests results, enter the container

make pull_docker
make clean_experiments_result
make enter_debug_docker

build planners and start mass test execution

if [ ! -d ./STRRT_Planner ]
    then
    cd ./app
    fi
make build_planner_debug
make mass_tests

Analyse data

Use ./mass_test/final_analysis.ipynb to analyse planner execution data

Data, that has been used in the paper is available at https://disk.yandex.ru/d/-73LOGO5kOSYuA

Feel free to open issues!