Add MBD-based FT FW

[triccyx]

Used MBD generated code from Simulink as new FT sensor code.
Check Simulink model in robotology/icub-firmware-models#2.

[pattacini]

I took the chance to trace back the pristine memory areas from which we do the memcpy:

icub-firmware/emBODY/eBcode/arch-arm/embot/app/embot_app_application_theSTRAIN.cpp

Lines 270 to 271 in b6c0a71

	embot::dsp::q15::matrix transfQ15matrix[numOfSets];
	embot::dsp::q15::matrix transfQ15tare[numOfSets];

Then, it'd be sufficient to move the memcpy into those code lines where transfQ15matrix and transfQ15tare are filled in to obtain the correct asynchronous update.

However, I see (and I remember now) that the FW deals with a set of diverse matrices and offsets; therefore, it might be convenient to keep things as they are now, for the sake of simplicity.

After all, the continuous copy does not take long.

[triccyx]

... sake of simplicity

Agree

[pattacini]

@simeonedussoni and @triccyx tested the MBD FW today successfully 🚀

Hence, @marcoaccame you can now review the code.

[marcoaccame]

Hence, @marcoaccame you can now review the code.

Hi @pattacini and @triccyx, I'm having a look at it.

[marcoaccame]

Well done: the changes are OK and can be merged.

There are only some minor comments reported in the following.

• Inside bool embot::app::application::theSTRAIN::Impl::processing() you moved the calibration inside if(false == runtimedata.data.adcsaturation) { .. } . It was originally outside. Does this movement of code change the behaviour? I have checked. The behaviour is the same. I am telling that only because at first I thought there was a mistake. But it there is not any.

• When you produce a new binary, you will need to advance the application number so that we can identify the new binary vs current released version which now is:

  constexpr embot::app::theCANboardInfo::applicationInfo applInfo 
  { 
      embot::prot::can::versionOfAPPLICATION {2, 0, 10},
      embot::prot::can::versionOfCANPROTOCOL {2, 0}    
  };

• In file embot_app_application_theSTRAIN.h I see an include of a .h file w/ the entire relative path. That compiles and is fine, but it is better to remove the path in inclusions and to add it to the environment variable of the project.

• Your fork is not rebased vs latest robotology:devel. However, the changes of the missing PRs are not related to any of the code you have touched and automatic rebase will work fine.

[pattacini]

In file embot_app_application_theSTRAIN.h I see an include of a .h file w/ the entire relative path. That compiles and is fine, but it is better to remove the path in inclusions and to add it to the environment variable of the project.

I agree w/ @marcoaccame.
@simeonedussoni @triccyx could you address this, please?

[simeonedussoni]

fixed few minor issues:

• version number is now 2.0.13 (note: it is already higher than the IMU remap)
• tabs are replaced with spaces
• included file is without explicit path, which is in the configuration tab

[pattacini]

Thanks @triccyx @simeonedussoni 🚀
Let's proceed with yielding the binaries then!