Are your parameters configured correctly?
- The test duration is a guard against tests not terminating. It should be measured in wallclock time, i.e. "if my test hasn't finished after this many seconds, kill it"
Is you test using simulated time?
- If so, is your time publisher publishing time at a faster rate than your sensor data? If not, this can cause an error of one period in determining the gap between
Are you setting the /use_simtime property to prevent time from jumping?