in order to test my path-planning algorithm (which is actually a standalone ROS Node communicating with my robot) I would like to build an automated test-platform.
My final goal would be to start the simulation by sending a ros-message (and finally some other information like where an obstacle should be placed in the simulation environment etc.), let the simulation run, store some data (rosbags,...) and send information about the successful or not successful end of the simulation(e.g. when my robot has reached a certain point = success / when the robot hits an obstacle = no success). This last information should lead to sending the "stop" - message via ROS in order to finish the simulation.
Whenever the simulation is stopped this way, the next run should begin automatically (with some other parameters concerning the simulation (environment)-setup) and all the mentioned points from above should be executed again.
When everything works as I imagine, I would like to start the whole simulation-series once (e.g. in the evening) and evaluate the results (stored in rosbags, etc.) the next morning. Doing it that way would give me the possibility to concentrate on developing the algorithm during the day and simulate my algorithms at night.
To summarize what i want to achieve:
- start and stop V-REP simulation several times in a row (ROS or any other approach)
- send new parameters for changing environment setup at every new start(data e.g. stored in excel file)
- send trigger from V-REP to stop simulation when certain condition is reached
My setup: ROS kinetic, Ubuntu 16.04, V-REP PRO EDU V3.4.0
I really hope that I can get some useful information or tips how to achieve the idea of such a test-platform.
Thanks a lot for every kind of help!