Feature/pm 524 Generalize motor controller #284
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,39 @@ | ||
| // Copyright 2019 Project March. | ||
|
|
||
| #include "actuation_mode.h" | ||
| #include "motor_controller_state.h" | ||
| #include <string> | ||
|
|
||
| namespace march | ||
| { | ||
| class MotorController | ||
| { | ||
| public: | ||
| virtual double getAngleRadAbsolute() = 0; | ||
| virtual double getAngleRadIncremental() = 0; | ||
| virtual double getVelocityRadAbsolute() = 0; | ||
| virtual double getVelocityRadIncremental() = 0; | ||
|
Contributor
There was a problem hiding this comment. What if my motor controller does not have either an absolute or incremental encoder?
Contributor
Author
There was a problem hiding this comment. I thought about this; the incremental/absolute encoders are a characteristic of Joint rather than of the motor controller, so it would make more sense to move such specifications there. That would require me to subclass Joint into different types, which I preferred to avoid, especially since our current three options for motor controllers will probably be connected to both types of encoders. I think such a refactor of the joint class is better postponed to a moment in the future where an encoder is actually connected to something other than a motor controller.
Contributor
Author
There was a problem hiding this comment. Hmm on second thought, I could move the calculations using both encoders downstream to the motor controller. Then these would simplify to getPosition() and getVelocity(). Downside would be that these calculations would be that this code would probably be duplicate in each different MotorController class, but that might again be resolvable by inheriting from a MultipleEncodersMotorControllers class that implements these methods common for motor controllers with both encoders. That was a nice spam of hersenspinsels, let me know if it makes any sense and what you think.
Contributor
Author
There was a problem hiding this comment. Decided that since the 3 motor controllers currently under consideration all have both an incremental and an absolute encoder, this change is not very useful for now. It won't be difficult to implement later on when such a motor controller is actually put to use. |
||
| virtual int16_t getTorque() = 0; | ||
| virtual MotorControllerStates getStates() = 0; | ||
|
|
||
| virtual ActuationMode getActuationMode() const = 0; | ||
| virtual uint16_t getSlaveIndex() const = 0; | ||
|
Contributor
There was a problem hiding this comment. What if a motor controller isn't a slave or doesn't use the master-slave structure.
Contributor
Author
There was a problem hiding this comment. Added a docstring to explain |
||
|
|
||
| virtual float getMotorCurrent() = 0; | ||
| virtual float getMotorControllerVoltage() = 0; | ||
| virtual float getMotorVoltage() = 0; | ||
| virtual bool getIncrementalMorePrecise() const = 0; | ||
|
Contributor
There was a problem hiding this comment. This also seems like a very specific function that should not be inside the motor controller class. When I read this method call I cannot tell what it would do generally on most controllers.
Contributor
Author
There was a problem hiding this comment. Added a docstring with explanation. This method is relevant as long as a motor controller has both an incremental and an absolute encoder. |
||
|
|
||
| virtual void actuateRad(double target_rad) = 0; | ||
| virtual void actuateTorque(int16_t target_torque) = 0; | ||
|
|
||
| virtual void goToOperationEnabled() = 0; | ||
|
Olavhaasie marked this conversation as resolved.
Outdated
|
||
|
|
||
| virtual bool initialize(int cycle_time) = 0; | ||
|
Contributor
There was a problem hiding this comment. What if my motor controller does not need a cycle time, but it does need something else for initialization. That would require adding more parameters and therefore breaking the contract.
Contributor
Author
There was a problem hiding this comment. I think you could alternatively pass something like that to the constructor as well, but I guess that wouldn't be very need since you need it nowhere else. I think I am going to :ostrich: this problem since it is highly dependent on how exactly the startup of another motor controller would work, which is a bit difficult to predict. I will see if I get useful enough insight from the ODrive and the Ingenia to make general statements about this. Otherwise, I think I will leave it up to future generations to redesign this a bit. |
||
| virtual void reset() = 0; | ||
| virtual bool checkState(std::ostringstream& error_msg, std::string joint_name) = 0; | ||
|
|
||
| virtual ~MotorController() noexcept = default; | ||
| }; | ||
|
|
||
| } // namespace march | ||
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,35 @@ | ||
| // Copyright 2019 Project March. | ||
| #ifndef MARCH_HARDWARE_MOTOR_CONTROLLER_STATE_H | ||
| #define MARCH_HARDWARE_MOTOR_CONTROLLER_STATE_H | ||
|
|
||
| #include <string> | ||
| #include "imotioncube_state.h" | ||
|
|
||
| namespace march | ||
| { | ||
| struct MotorControllerStates | ||
| { | ||
| public: | ||
| MotorControllerStates() = default; | ||
|
|
||
| float motorCurrent; | ||
| float controllerVoltage; | ||
| float motorVoltage; | ||
| int absoluteEncoderValue; | ||
| int incrementalEncoderValue; | ||
| double absoluteVelocity; | ||
| double incrementalVelocity; | ||
|
|
||
| std::string statusWord; | ||
| std::string motionError; | ||
| std::string detailedError; | ||
| std::string secondDetailedError; | ||
| IMCState state; | ||
| std::string detailedErrorDescription; | ||
| std::string motionErrorDescription; | ||
| std::string secondDetailedErrorDescription; | ||
|
Roelemans marked this conversation as resolved.
Outdated
|
||
| }; | ||
|
|
||
| } // namespace march | ||
|
|
||
| #endif // MARCH_HARDWARE_IMOTIONCUBE_STATE_H | ||
Uh oh!
There was an error while loading. Please reload this page.