dynamixel-client 0.1

Dora Node client for dynamixel motors.

DynamixelClient for XL motors

This node is a client for the Dynamixel motors. It is based on the Dynamixel SDK and is used to control the motors. It is a Python node that communicates with the motors via the USB port.

YAML Configuration

nodes:
  - id: dynamixel_client
    path: client.py # modify this to the relative path from the graph file to the client script
    inputs:
      pull_position: dora/timer/millis/10 # pull the present position every 10ms
      pull_velocity: dora/timer/millis/10 # pull the present velocity every 10ms
      pull_current: dora/timer/millis/10 # pull the present current every 10ms

      # write_goal_position: some goal position from other node
      # write_goal_current: some goal current from other node

      # end: some end signal from other node
    outputs:
      - position # regarding 'pull_position' input, it will output the position every 10ms
      - velocity # regarding 'pull_velocity' input, it will output the velocity every 10ms
      - current # regarding 'pull_current' input, it will output the current every 10ms

    env:
      PORT: COM9 # e.g. /dev/ttyUSB0 or COM9
      CONFIG: config.json # the configuration file for the motors

Arrow format

Outputs

Arrow StructArray with two fields, joints and values:

import pyarrow as pa

# Create a StructArray from a list of joints (py_list, numpy_array or pyarrow_array) and a list of values (py_list, numpy_array or pyarrow_array)
arrow_struct = pa.StructArray.from_arrays(
    arrays=[joints, values],
    names=["joints", "values"]
)

# Send the StructArray to the dataflow
node.send_output("output_name", arrow_struct, None)

# Receive the StructArray from the dataflow
event = node.next()
arrow_struct = event["value"]
joints = arrow_struct.field("joints")  # PyArrow Array of Strings
values = arrow_struct.field("values")  # PyArrow Array of Int32/Uint32/Float32...

Inputs

Arrow StructArray with two fields, joints and values:

import pyarrow as pa

# Create a StructArray from a list of joints (py_list, numpy_array or pyarrow_array) and a list of values (py_list, numpy_array or pyarrow_array)
arrow_struct = pa.StructArray.from_arrays(
    arrays=[joints, values],
    names=["joints", "values"]
)

# Send the StructArray to the dataflow
node.send_output("output_name", arrow_struct, None)

# Receive the StructArray from the dataflow
event = node.next()
arrow_struct = event["value"]
joints = arrow_struct.field("joints") # PyArrow Array of Strings
values = arrow_struct.field("values") # PyArrow Array of Int32/Uint32/Float32...

Note: The zero-copy is available for numpy arrays (with no None values) and pyarrow arrays.

Configuration

The configuration file that should be passed to the node is a JSON file that contains the configuration for the motors:

{
  "shoulder_pan": {
    "id": 1,
    "model": "x_series",
    "torque": true,
    "P": 800,
    "I": 0,
    "D": 0,
    "goal_current": null
  }
}

The configuration file starts by the joint name of the servo. id: the id of the motor in the bus, model: the model of the motor, torque: whether the motor should be in torque mode or not (at the beginning), P: the proportional gain for position control mode, I: the integral gain for position control mode, D: the derivative gain for position control mode, goal_current: the goal current for the motor at the beginning, null if you don't want to set it.

License

This library is licensed under the Apache License 2.0.