Add low pass filter for force torque sensor

[JohnTGZ]

Overview

This PR waits on #346 to be merged.

This PR adds the low pass filter to the force torque sensor broadcaster fts_broadcaster.
The low_pass_filter is implemented in control_toolbox.

We provide users with 2 different F/T sensor topics they can use as input to their control policies:

• They could read raw values from fts_broadcaster/wrench topic
• They could read values filtered with the low pass filter from the fts_broadcaster/wrench_filtered topic.

The following parameters were chosen for the low-pass filter:

          sampling_frequency: 50.0 # [Hz]
          damping_frequency: 10.0 # Cut-off frequency [Hz]
          damping_intensity: 0.707

Testing

Before testing, make sure to:

• install dependencies using rosdep install --from-paths src --ignore-src --rosdistro kilted -yr --skip-keys "gz-cmake3 DART libogre-dev libogre-next-2.3-dev rosetta"
• Clean the workspace and then re-build from source.

Results

The top plot shows the z-axis linear force values before filtering and the bottom plot are the values after allowing the low pass filter. You can observe that the readings are smoothed and that the magnitude of the spikes are attenuated.

[Yadunund]

Thanks for working on filtering the senor reading.

1. Do we expect users to subscribe to the filtered topic for their policy training and action? if so we'll need to update the aic_adapter subscription to subscribe to this topic.

• we should also update aic_interfaces.md to inform users of the availability of this topic Also clarify when to subscribe to /axia80_m20/wrench

[JohnTGZ]

@Yadunund This is still part of an ongoing discussion as to whether we need to filter these values for accurate scoring. Will update when we reach a conclusion.

[JohnTGZ]

Hi @luca-della-vedova Could I also have your help in reviewing this? This PR mainly adds a low-pass filter so that the already existing fts_broadcaster/wrench_filtered topic has a low-pass filter applied to it. It does not completely attentuate the spikes in magnitude but smooths the noise. Parameters could be tuned further with some knowledge of what we should look out for

[luca-della-vedova]

Is this change relevant to this PR or was it left by accident?

[JohnTGZ]

oops, unintended. Removed it.