Suggestion to streamline the implementation of new sensors

[AdelGuiot]

Depth sensors can only measure the distance up to a sample. As such, this information must be translated in the robot's workspace to be usable. This is currently done within the get_data function of the sensor class. More precisely, the get_rectified_data function, which itself is called out of the class (which, as a sidenote, probably hinders modularity).

I see two ways to go about improving this:

• Only return the raw distance measured by the sensor from get_data, and leave anything to do with the robot outside of the sensor class. Positions will only be set ad-hoc or within the robot class in that case.
• Or change the template by adding a get_rectified_data function, or such, which will use data from measured from get_data to send a new position for the robot.

What do you think?

[pchaillo]

Hi Adel,

Thanks for the feedback.

The actual structure follow this logic : the get_data function returns directly the sample height, to directly get the useful data. As get rectified data is used inside the get_data function, I think it does'nt hinder modularity.

But I get your point, I think putting the translation of the data into robot referential, to deduct the sample height from sensor data and robot position, directly in the template could be useful. But as all sensor have different size and sensing properties, I think each sensor could have a different position in robot referential, so the translation from one space to one another will not be exactly the same. I think it will need to be cautious, to find a generic way to do this referential transition.

Tell me what do you think !