code for article: Multi-robot Trajectory Generation with Feasibility Gurantee and Deadlock Resolution： An obstacle-dense Environment

It is a online replanning method for multi-robot which can achieve the recursive feasibility when solving optimization problems during replanning in addition to its capibility that resolving deadlock problem in obstacle-dense enviroment.

Here we provide a python3-based program which can simulate the trasition for multi-robot systems.

We have tested our code with: ubuntu 20.0.4 in python3

This APP need following dependencies:

numpy          1.21.2
scipy          1.6.3
cvxopt         1.3.0
matplotlib     3.4.3

IMPC-OB depends on the package openGJK to apply polytope-related calculation. As openGJK is written in C++, a conversion from C++ to python is needed. This can be done by generating a Python-Binding file called opengjkc.so.

PS: This file must be generated specifically for each different running environment. The one provided in IMPC-OB source code is generated based on python 3.9, thus it might not be available for importing in your environment. Be sure to manually generate it yourself!

The process is shown below:

git clone https://github.com/MattiaMontanari/openGJK  

cmake -E make_directory build  
cmake -E chdir build cmake -DCMAKE_BUILD_TYPE=Release ..   
cmake --build build  
cmake -E chdir build/examples/c ./example_lib_opengjk_ce  

Move the 'openGJK.c' file from the root directory to the 'src' directory, then open it and delete all keywords named restrict.

cd to the 'examples/python' directory and run source build.sh.
This should generate a file called opengjkc.cpython-38-x86_64-linux-gnu.so.
Rename it as opengjkc.so and move it to the 'IMPC-OB/src/opengjk' directory to replace the existing opengjkc.so file.

cd src
python3 test.py