Joint control

NAOqi Motion - Overview | API | Tutorial


What it does

These API are dedicated to control directly the position of the robot joints.
Each joint can be controlled individually, or in parallel with other joints.
There are two ways of controlling a joint or a group of joints:

How it works

These API just create a higher level of the DCM actuator control and provide interpolation to have more smooth behavior.

Getting started

“Body”, chains and joint names

Joints can be controlled:

  • individually, using a joint name, or
  • in parallel, using a chain of joints or a group of joints like “Body”.

For further details, see: NAO’s Chains.

It is possible to get the list of joints available on your robot using the ALMotionProxy::getBodyNames() method, described in the section: Case 2: Programmatic access to Joint Names.

Use Cases

Case 1: Controlling Joints

To control a joint, you need to specify the name of the joint, the target angle in radians, and how fast you want to go to the target angle.

almotion_controllingJoints.py


# -*- encoding: UTF-8 -*-

import time
import almath
import argparse
from naoqi import ALProxy

def main(robotIP, PORT=9559):
    motionProxy = ALProxy("ALMotion", robotIP, PORT)

    motionProxy.setStiffnesses("Head", 1.0)

    # Simple command for the HeadYaw joint at 10% max speed
    names            = "HeadYaw"
    angles           = 30.0*almath.TO_RAD
    fractionMaxSpeed = 0.1
    motionProxy.setAngles(names,angles,fractionMaxSpeed)

    time.sleep(3.0)
    motionProxy.setStiffnesses("Head", 0.0)

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--ip", type=str, default="127.0.0.1",
                        help="Robot ip address")
    parser.add_argument("--port", type=int, default=9559,
                        help="Robot port number")

    args = parser.parse_args()
    main(args.ip, args.port)

/sites/default/files/repository/69_html_nao/_images/motion_headyaw_position.png /sites/default/files/repository/69_html_nao/_images/motion_headyaw_speed.png

Case 2: Timed Interpolations

When you know in advance the trajectory that you want to follow, the ALMotionProxy::angleInterpolation() and ALMotionProxy::angleInterpolationWithSpeed() methods can be used to set up an interpolation.


# Example showing a joint trajectory with a single destination
names = "HeadYaw"
angleLists = 1.0
times = 1.0
isAbsolute = True
proxy.angleInterpolation(names, angleLists, times, isAbsolute)

You can command multiple joints in one command, by using a single time, and a number of target angles equal to the number of joints.


# Example showing a command for the two joints in the 'Head' alias
# 'Head' is expanded to ['HeadYaw','HeadPitch']
names = "Head"
angleLists = [-1.0,-1.0]
times = 1.0
isAbsolute = True
proxy.angleInterpolation(names, angleLists, times, isAbsolute)

The same command can take a list of angles with corresponding times.


# Shake the head from side to side
names = "HeadYaw"
angleLists = [1.0, -1.0, 1.0, -1.0, 0.0]
times      = [1.0,  2.0, 3.0,  4.0, 5.0]
isAbsolute = True
proxy.angleInterpolation(names, angleLists, times, isAbsolute)

Similarly, trajectories can be specified for multiple joints.


# Two trajectories in one command. Each trajectory must have a
# corresponding number of times
names      = ["HeadYaw", "HeadPitch"]
angleLists = [[1.0, -1.0, 1.0, -1.0], [-1.0]]
times      = [[1.0,  2.0, 3.0,  4.0], [ 5.0]]
isAbsolute = True
proxy.angleInterpolation(names, angleLists, times, isAbsolute)

Case 3: Reactive Control

The commands ALMotionProxy::setAngles() and ALMotionProxy::changeAngles(), do not block the calling thread. This makes them ideal for being called often in reactive control loops, such as head tracking. You can call them often, with contradictory commands, and motion will ensure that the trajectory is smooth in position and continuous in velocity.

almotion_reactiveControl.py


# -*- encoding: UTF-8 -*-

import time
import argparse
from naoqi import ALProxy

def main(robotIP, PORT=9559):
    motionProxy = ALProxy("ALMotion", robotIP, PORT)

    motionProxy.setStiffnesses("Head", 1.0)

    # Example simulating reactive control
    names = "HeadYaw"
    angles = 0.3
    fractionMaxSpeed = 0.1
    motionProxy.setAngles(names,angles,fractionMaxSpeed)
    # wait half a second
    time.sleep(0.5)
    # change target
    angles = 0.0
    motionProxy.setAngles(names,angles,fractionMaxSpeed)
    # wait half a second
    time.sleep(0.5)
    # change target
    angles = 0.1
    motionProxy.setAngles(names,angles,fractionMaxSpeed)

    time.sleep(3.0)
    motionProxy.setStiffnesses("Head", 0.0)

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--ip", type=str, default="127.0.0.1",
                        help="Robot ip address")
    parser.add_argument("--port", type=int, default=9559,
                        help="Robot port number")

    args = parser.parse_args()
    main(args.ip, args.port)

/sites/default/files/repository/69_html_nao/_images/motion_headyaw_reactive_position.png /sites/default/files/repository/69_html_nao/_images/motion_headyaw_reactive_speed.png

It is also possible to use time with angleInterpolation to do reactive control.

almotion_angleInterpolationReactif.py


# -*- encoding: UTF-8 -*-

import almath
import time
import argparse
from naoqi import ALProxy

def main(robotIP, PORT=9559):
    motionProxy = ALProxy("ALMotion", robotIP, PORT)

    motionProxy.setStiffnesses("Head", 1.0)

    # Head Start to zeros
    names             = "Head"
    targetAngles      = [0.0, 0.0]
    maxSpeedFraction  = 0.2 # Using 20% of maximum joint speed
    motionProxy.angleInterpolationWithSpeed(names, targetAngles, maxSpeedFraction)

    # Example showing a reactive control with time function angleInterpolation
    # Goal: after 1.0 second, retarget from 40 to 0 degree: smooth transition

    # Interpolate the head yaw to 20 degrees in 2.0 seconds
    # With post, angleInterpolation become non-blocking
    names      = "HeadYaw"
    angleLists = 20.0*almath.TO_RAD
    timeLists  = 2.0
    isAbsolute = True
    motionProxy.post.angleInterpolation(names, angleLists, timeLists, isAbsolute)
    time.sleep(0.5)

    # Call getTaskList to have the previous angleInterpolation task number
    taskList = motionProxy.getTaskList()

    # Prepare the next target to 50.0 degrees in 1.0 second
    angleLists = 50.0*almath.TO_RAD
    timeLists  = 1.0
    motionProxy.post.angleInterpolation(names, angleLists, timeLists, isAbsolute)
    time.sleep(0.5)

    # Kill the first angleInterpolation (go to 40.0 degrees), the second start
    # smoothly from the current joint position and velocity (go to 0 degrees)
    motionProxy.killTask(taskList[0][1])

    time.sleep(2.0)
    motionProxy.setStiffnesses("Head", 0.0)


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--ip", type=str, default="127.0.0.1",
                        help="Robot ip address")
    parser.add_argument("--port", type=int, default=9559,
                        help="Robot port number")

    args = parser.parse_args()
    main(args.ip, args.port)