UML graph

The whole software structure is divided into 5 important parts:

1. Python nodes
2. Cpp nodes
3. Helper scripts
4. URDF
5. Packages

1. Python nodes

• Finite State Machine(FSM):

  FSM is a mother node of the whole architecture, it uses the build_ontology_map.py to import a class defined to ADD, MANIPULATION, and get QUERY from the information. Moreover imports three important services defined in set_object_state.py to set the state of the battery, base movement, and arm movement. This node subscribes to marker_publisher.cpp node to get /image_id detected by the robot camera and this id number is communicated with marke_server.cpp node to get information about id founded. When the robot wants to move after detecting all markers id published a Move Base Action to the move_base package to move base to the robot. Finally uses a service response to get Pose, Base movement state, and arm movement state from the robot_states.py node.

• Robote States:

  Robot states subscribe to the /odom topic from robot urdf to get information odometry information. All services do set and get here, the services are: base_movement_state and battery_level_state. Additionally Get_pose service done in this node and the I/O of the service communicates via <response & requist> with FSM node and helper scripts.

2. Cpp nodes

• Marker Server:

  All room proprieties are defined in this node like room id and room details (position, connection, door, last visit time). This node communicates with the FSM node via service (Req/Res) to send room information.

• My Moveit:

  This node contains all necessary joint configurations to find boxes in the environment. Via service communicates with a helper script set_object_state.py to changes state of robot arm between True and False each time the arm state is True this node becomes active and sends joint position information to robot_assignment.urdf. Otherwise, if it is * False* the robot arm stop working.

3. Helper scripts

These scripts don’t start any node they were used only to import essential functions used in node FSM and robot_state

• Build Ontology Map:

  This script uses a service response generated from the robot_state.py node to get the current position of the robot in the map and update the is_in method. This method indicates the current location of the robot. Furthermore, this script is useful to load ontology_map defined in the armor package ad sync with the resoner.

• Set Object State:

  Three important services for the state of the battery_state, base_movement_state, and arm_movement_state are defined here. This script sends a service request to the robot_state.py node to set the base movement state and set battery level state. Additionally sends a service request also to my_moveit.cpp node for set arm movement state.

4. URDF

• Robot Assignment:

  Its generated from * Moveit Setup Assistant* and contain a robot model and various topic useful for moving the base and controlling the arm of the robot. Robot urdf subscribes to move_base to get velocity uses /cmd_vel topic and publishes odometry information uses /odom topic to robot_state.py node to get the current position of the robot in the ontology map, furthermore publishes /robot_camera/image_box to marker_publisher.cpp node to send image detected via camera.

  Finally publishes base information using the /scan topic and publishes tf transform tree [1]_ uses /tf topic to slam_gmapping to update robot situation in map.

  [1]_ tf is a package that lets the user keep track of multiple coordinate frames over time. tf maintains the relationship between coordinate frames in a tree structure buffered in time, and lets the user transform points, vectors, etc between any two coordinate frames at any desired point in time.

5. Packages

• Aruco Ros:

  Aruco library provides real-time marker-based 3D pose estimation using AR markers. In this assignment used marker_publisher.cpp node to publishes /image_id detected from /robot_camera/image_box and its published to FSM node.

  – For more information on Aruco package then refer to this link.

• SLAM Gmapping:

  The gmapping package provides laser-based SLAM (Simultaneous Localization and Mapping), as a ROS node called slam_gmapping. Using slam_gmapping, you can create a 2-D occupancy grid map (like a building floorplan) from the laser and pose data collected by a mobile robot. Slam_gmapping used /scan and tf topic data generated by robot urdf in order to update the position of the base_link and arm_link in the map.

  – For more information on the Slam package then refer to this link.

• Move Base:

  The move_base package provides an implementation of an action that, given a goal in the world, will attempt to reach it with a mobile base. The move_base node links together a global and local planner to accomplish its global navigation task. Move base to get an Action from FSM node and publish velocity via /cmd_vel to robot urdf.

  – For more information on the Move base package then refer to this link.

• Assignment Moveit:

  After building the urdf or xacro file you can proceed in your terminal with the command roslaunch moveit_setup_assistant setup_assistant.launch and run Moveit setup assistant. Now there is a possibility to define various properties for robot models and generate new urdf file plus package where there is an essential configuration with a launch file to be able to use a robot model.

  – For more information on the Moveit package then refer to this link.

• aRMOR:

  aRMOR is a powerful and versatile management system for single and multi-ontology architectures under ROS. It allows loading, querying, and modify ontology map. aRMOR add and manipulate room and door in ontological_map.owl and send the query as feedback to the FSM node. Additionally, it works parallel with helper script build_ontology_map.py to load the ontology map and synchronize with armor_client.py each time ontology is updated.

  – For more information on the armor package then refer to this link.