GRTFramework for Python, for event-based robotics programming.
Beyond basic framework and abstractions, GRTPyFramework provides a method by which more sophisticated logic and abstractions can be developed, provided a particular code structure is followed in line with design methodology.
User code is split into two parts: mechanisms, and controllers.
Mechanisms are collections of sensors and actuators that represent a specific robot assembly, providing an API by which that assembly can be controlled globally, as opposed to in terms of its individual parts. Mechanisms should not operate on their own, simply translating abstract commands into specific hardware directives. Mechanisms may be nested, and also may provide closed loop operation (limit switch logic, etc.)
Controllers are governing objects that provide instructions for the various mechanisms on a robot. A controller contains the logic necessary to tell a mechanism what to do. A controller may operate over one or more mechanisms, defining the logic by which each assembly is used. Controllers may be fully autonomous or draw on sensor or user feedback to command Mechanisms in the appropriate way. As a matter of abstraction, Controllers should never have direct access to Sensors or Actuators. Mechanism controllers, drive controllers, and macros are the different types of controllers.
Flow Diagram:
Controller (Mechanism controller, drive controller, macro)
|
'--- Mechanism
|
|--- Actuator(s)
'--- Sensor(s)
Sensors help the robot to sense its state and environment. For example, a potentiometer may be used to determine the rotation undergone by an axle (robot state), while a distance sensor may be used to sense the robot's location in relation to a field (environment).
Sensors have listeners; listeners are methods that are called whenever a sensor's state changes. When a listener is called, some statements are executed. For example, a joystick listener (see mechcontroller.py) would be called whenever a joystick state changes; when a button is pressed. The joystick listener would execute different commands depending on the id of the button that is pressed.
A macro is a fundamental action that can be executed autonomously by the robot, more accurately and more reliably than a human driver. For example, driving forward an exact distance, or turning by a certain angle. These two macros come built in, along with macros for executing other macros sequentially, concurrently, and conditionally. Macros can take in mechanisms and sensors as parameters, as they will make mechanisms do certain things based on sensor inputs.
Different autonomous modes can be created in the autonomous module as instances of AutonomousMode. If an autonomous mode is a sequence of macros, it will be implemented as an instance of MacroSequence (A subclass of AutonomouseMode specifically for executing macros.)