EGM is active, but not in running mode (i.e. listening for command references)!

[dave992]

I am trying to automate starting an EGM session using the abb_robot_driver. I basically follow the steps in the example provided here, using a state machine in Python. Here is a short summary of the calls I perform via the state machine:

rosservice call /rws/stop_rapid "{}"
rosservice call /rws/pp_to_main "{}"
rosservice call /rws/start_rapid "{}"
rosservice call /rws/sm_addin/start_egm_joint "{}"
rosservice call /egm/controller_manager/switch_controller "start_controllers: ......"

There are some race conditions that I take care of by monitoring /rws/system_states, /rws/sm_addin/runtime_states and /egm/egm_states/ before I execute the next service call.

For this issue specifically, I monitor the active field of the message received from the /egm/egm_states topic after executing the service call to /rws/sm_addin/start_egm_joint to check if my request has been processed. If I then try to activate a controller via the /egm/controller_manager/switch_controller service, I sometimes get the following error from the /egm_hardware_interface node:

EGM is active, but not in running mode (i.e. listening for command references)! Refusing to start controller(s): [joint_position_trajectory_controller]

This is another race condition (if I add a sleep, it works), but I wonder how I can make sure EGM is ready to start a controller. How can I monitor if EGM is 'in running mode'?

[gavanderhoorn]

This one I wouldn't know immediately.

I'd have to check the various state topics to see whether we're already publishing a piece of information which could be used here.