Super ROS!

Table of Contents

• Why SuperROS
• Basic Usage: Publishers/Subscribers
• TF Transformations
• 3D Scene Visualization
• RGB Camera

Why SuperROS?

SuperROS is something "super" (above) the Robot Operating System (ROS). First of all, SuperROS allows you to separate your business logic from the ROS framework making more transparent the transition to another middleware (e.g. ROS2 in the years to come). But, more importantly, allows you to Write Less and Do More.

You can use SuperROS as a classical ROS package or just include the scripts folder to your PYTHONPATH. If you use it as a ROS package be sure to catkin_make it. More generally, to make things work you need ROS installed with the default procedure here.

Basic usage: Publishers/Subscribers

In this simple example we want to design a simple Amplifier reading a Float32 data which is then published (doubled) on another topic at a certain frequency. Although in this snippet the code reduction is not astonishing, it is a basic example to introduce the Publisher/Subscriber paradigm in SuperROS.

Super ROS

Plain ROS

from superros.comm import RosNode from std_msgs.msg import Float32 node = RosNode("amplifier") node.setHz(node.setupParameter("hz", 30)) node.createBufferedSubscriber("in", Float32) node.createPublisher("out", Float32) while node.isActive(): data = Float32(node.getBufferedData("in").data * 2.0) node.getPublisher("out").publish(data) node.tick()

import rospy from std_msgs.msg import Float32 current_value = 0.0 def newData(msg): global current_value current_value = msg.data rospy.init_node('amplifier', anonymous=True) rate = rospy.Rate(rospy.get_param("~hz", 30)) sub = rospy.Subscriber("in", Float32, newData) pub = rospy.Publisher("out", Float32) while not rospy.is_shutdown(): data = Float32(current_value*2.0) pub.publish(data) rate.sleep()

You can run directly the Amplifier Example with:

roslaunch superros example_subpub.launch

TF Transforms made easy

In this example (a simple TF Listener/Broadcaster with some 3D transformations) it is clear the benefit of an unified proxy (ie. the RosNode) in order to reduce UserApp<->ROS interactions. Moreover, SuperROS uses the robust PyKDL library as shared model for 3D Transformations, simplifying code readability and portability.

Super ROS

Plain ROS

from superros.comm import RosNode import PyKDL node = RosNode("tf_manipulator") node.setHz(node.setupParameter("hz", 30)) while node.isActive(): # Fetches Frame from TF Tree frame = node.retrieveTransform("base_link", "odom") if frame is not None: # Creates a relative transformation frame t = PyKDL.Frame( PyKDL.Rotation.RotZ(1.57), PyKDL.Vector(0, 0, 0.5) ) # Applies transformation frame = frame * t # Send Frame to TF Tree node.broadcastTransform( frame, "base_link_2", "odom", node.getCurrentTime() ) node.tick()

import rospy import tf import numpy as np rospy.init_node('tf_manipulator') rate = rospy.Rate(rospy.get_param("hz", 30)) listener = tf.TransformListener() br = tf.TransformBroadcaster() while not rospy.is_shutdown(): # Fetches Frame (T,R) from TF Tree. May fail try: (trans1, rot1) = listener.lookupTransform( '/odom', '/base_link', rospy.Time(0) ) except (tf.LookupException): continue # Packs T,R in a single 4x4 matrix trans1_mat = tf.transformations.translation_matrix(trans1) rot1_mat = tf.transformations.quaternion_matrix(rot1) mat1 = np.dot(trans1_mat, rot1_mat) # Creates a relative transformation frame t = tf.transformations.rotation_matrix(1.57, [0, 0, 1]) t[:3, 3] = np.array([0, 0, 0.5]).T # Creates a relative transformation frame mat2 = np.dot(mat1, t) # Unpacks T,R from the 4x4 matrix trans2 = tf.transformations.translation_from_matrix(mat2) rot2 = tf.transformations.quaternion_from_matrix(mat2) # Send Frame to TF Tree br.sendTransform( trans2, rot2, rospy.get_rostime(), "base_link2", "odom" ) rate.sleep()

3D Scene Visualization

In this example an animated 3D Scene was presented using the SuperROS simplified library. The communication between Rviz and your node is -- surely -- made with topics but this behaviour is masked by the VisualizationScene object.

Super ROS

Plain ROS

from superros.comm import RosNode from visualization_msgs.msg import MarkerArray from superros.draw import VisualizationScene, Color import PyKDL import math node = RosNode("example_3dscene") node.setHz(node.setupParameter("hz", 30)) pub = node.createPublisher("test", MarkerArray) # Visualization Scene scene = VisualizationScene(node, "test") scene.createCone(name="cone1", color=Color.pickFromPalette(0)) scene.createCone(name="cone2", color=Color.pickFromPalette(1)) scene.createCube( name="cube1", color=Color.pickFromPalette(2), transform=PyKDL.Frame(PyKDL.Vector(-2.0, 0, 0)) ) # Cone2 Frame cone_2_frame = PyKDL.Frame(PyKDL.Vector(2.0, 0, 0)) while node.isActive(): # Update Cone2 Pose getting stored Object cone_2_frame.M.DoRotZ(0.02) scene.getObjectByName("cone2").setTransform(cone_2_frame) # Update Cone1 Pose overwriting old Object scene.createCone(name="cone1", transform=PyKDL.Frame( PyKDL.Vector( 0.0, 0.0, math.sin(1.0+2*math.pi*1.0*node.getCurrentTimeInSecs()) ) )) # Publish World Frame node.broadcastTransform(PyKDL.Frame(), "world", "base", node.getCurrentTime()) # Update Scene scene.show() # Node Forward node.tick()

Ehm...?

You can run directly the 3DScene Example with:

roslaunch superros example_3dscene.launch

RGB Camera

In this example the simple CameraRGB wrapper is shown. All the original ImageTransport and CvBridge image conversions stuff is hidden behind the simple CameraRGB object. With registerUserCallabck the user can bind its custom callback function called whenever a new FrameRGB is available. The frame contains in rgb_image the off-the-shelf numpy matrix representing the image, compliant with the OpenCV/cv2 library.

from superros.comm import RosNode
from superros.cameras import CameraRGB
import cv2

node = RosNode("example_rgbcamera")
node.setHz(node.setupParameter("hz", 30))
rgb_topic = node.setupParameter("camera_topic","")

# New Frame callback
def newFrame(frame):
    cv2.imshow("image",frame.rgb_image)
    cv2.waitKey(1)

# Camera Object
camera = CameraRGB(
    node,
    rgb_topic=rgb_topic,
    compressed_image='compressed'in rgb_topic
)
camera.registerUserCallabck(newFrame)

while node.isActive():
    node.tick()

You can run directly the 3DScene Example with:

roslaunch superros example_3dscene.launch