[Spinal][Config] Additional features and new configuration for Spinal

[sugihara-16]

What is this

This PR is an integrated PR with #624 and #594 . The structure of the STM32 project was also changed to accommodate all available boards.

Details

This PR provides following modifications:

• Supporting direct servo control.
• Supporting DShot protocol.
• Supporting board-by-board independent configuration setup.

TODO

I have checked this branch works in Spinal F7, H7, and H7_v2 without problems. So, after the review by contributes, it can be merged immediately.

[tongtybj]

@sugihara-16

Have you test this branch with any multiliinked robot, like Hydrus ?

[sugihara-16]

@tongtybj
Not yet.
This PR provides direct servo control mode and Dshot protocol mode, but they do nothing about robots that use Neuron because direct servo control mode cannot be mixed with NERVE_COMM mode currently (https://github.com/sugihara-16/aerial_robot/blob/e73d86c035ecf7634a171f48faa8a5ec1edcca3a/aerial_robot_nerve/spinal/mcu_project/boards/stm32H7/Src/main.c#L258-L263). Additionally, regarding thrust signal, Spinal doesn't transmit thrust signal by itself when we use Neuron.

[tongtybj]

@sugihara-16

I understand that theoretically this PR does not effect the spinal-neuron system, but the real word is much stricky than what we predict or image.

For a very large modification such as this PR, it is indispensible to test with real platform as much as possible (mini_quadrotor, hydrus, dragon, delta, etc.).
However, I also realize it is an exhausted work. So please only try with hydrus.

[sugihara-16]

OK, I got it.
I gonna try Hydrus in a few days.

[sugihara-16]

@tongtybj
Thank you for your insightful suggestion!
I tested this branch with Hydrus Xi real machine (equipped with Spinal H7) and found several potential problems. After some work, I confirmed Hydrus Xi can perform hovering and aerial transformation.
Additionally, although I guess it's also better to test with F7, now we don't have multilinked robots equipped with clean F7.

[tongtybj]

I tested this branch with Hydrus Xi real machine (equipped with Spinal H7) and found several potential problems. After some work, I confirmed Hydrus Xi can perform hovering and aerial transformation.

Excellent!

[sugihara-16]

@tongtybj
Hi, regarding the problem of the encoder initialization in F7, I found this problem is due to my commit sugihara-16@af3f910 . I remember this modification was done to make use of the same Dynamixel driver as Neuron, but the init func for external encoder is now disenabled in this line. I will check around this part more detail so please wait to merge this PR for a bit.

[sugihara-16]

I solved the above problem by using temporary directives.