ROS2 Robot Stacking (Palletizing)

Overview

This project demonstrates controlling Shanghai Agilebot collaborative robots for palletizing tasks using ROS2 software packages. It integrates vision processing, stacking algorithms, and robotic arm control functions, supporting both physical robots and virtual simulation environments. Readers should have basic knowledge of ROS2 and robot programming.

click to view

System Requirements

• Operating System: Ubuntu 22.04
• Core Framework: ROS2 Humble
• Robot Drivers:
  • Shanghai Agilebot ROS-driver package (version >= v0.0.1)
  • Python SDK (version >= v1.6.3.2)
• Vision System:
  • AgileGaze (Agilebot proprietary vision software) or compatible simulation node
• Robot Requirements:
  • Shanghai Agilebot collaborative robot (physical or virtual simulation)
  • Software: Copper ≥ v7.6.0.1

Virtual Environment Note: Simulation environment support available. Contact Agilebot for virtual controller license.

---

System Architecture

Node Topology Diagram

Legend:

1. Node Types (Green Rectangles):

   • robot_Stacking : Main control node (system coordinator)
   • gbt_agilegaze : Vision recognition node (image processing)
   • stacking_visualizer : Visualization node (result display)
   • service_server : Service interface (control commands)
   • robot_bridge : Communication bridge (hardware interface)
   • RViz : ROS visualization tool (3D display)

2. Non-Node Entities:

   • 🔶 Entity (Orange):
     • Robot Controller : Physical robot controller
   • 🔷 Service (Blue):
     • GetAgileGaze : Vision recognition service
     • PalletVisualizer : Visualization service
     • send_script : Script transmission service

---

Core Module Functions

Module Name	Function Description
gbt_agilegaze	Interfaces with AgileGaze system, provides object center coordinates (x,y) and rotation angle (c)
robot_stacking	Main stacking controller, coordinates vision, algorithms, and execution
stacking_visualizer	Visualizes grasp/placement points in real-time
service_server	ROS-driver gateway, forwards control commands
robot_bridge	Synchronizes robot state to RViz
Robot Controller	Physical robot controller

---

Workflow

Legend

Element Type	Color/Marker	Example
ROS Nodes	💚 Green background	robot_Stacking, gbt_agilegaze
Physical Entities	🧡 Orange background	Robot Controller
Service Calls	🔷 Blue diamond	GetAgileGaze.srv()
External Interfaces	💜 Purple background	Operator
Data Feedback	Solid lines	xycList, execution results

Workflow Steps

1. Start Trigger
   Operator sends start command with parameters to main node

2. Vision Recognition
   Main node calls AgileGaze via GetAgileGaze.srv() to get xyc object data

3. Path Planning
   Calculates grasp/placement points and motion trajectories

4. Visualization Update
   Syncs points via PalletVisualizer.srv() for RViz rendering

5. Robot Execution
   Sends commands through send_script.srv() to robot controller

6. Status Synchronization
   Real-time robot state sync to RViz, returns task results to operator

---

Configuration Files

File	Path	Contents
Robot Config	../gbt_driver/config/robot_config.yaml	Robot IP address, connection parameters
Stacking Params	config/stacking_params.yaml	Algorithm parameters, workspace settings
Visualization Params	config/pallet_viz_params.yaml	Visualization display parameters

Modify according to actual application. See file comments for details.

---

System Interfaces

Vision Interface Service

• Service Type: gbt_stacking_interface/srv/GetAgileGaze
• Description: Gets object coordinates from vision system

# Request
---
# Response
gbt_stacking_interface/AgileGaze AgileGaze

See gbt_stacking_interface/AgileGaze.msg for AgileGaze interface details.

RVIZ Visualization Service

• Service Type: gbt_stacking_interface/srv/PalletVisualizer
• Description: Updates visualization coordinates

# Request
geometry_msgs/Point[] grasp_points
geometry_msgs/Point[] place_points

# Response
bool success     
string message

---

Quick Start Guide (Simulation Mode)

Prerequisites

1. Install ROS2 Humble and dependencies
2. Install Agilebot ROS driver (Installation Guide)
3. Configure robot IP in gbt_driver/config/robot_config.yaml
4. Configure gripper/IO mapping in Agilelink (default DO port 1)
5. Servo on

Launch Command

# Build project
colcon build
source install/setup.bash

# Start simulation
ros2 launch gbt_stacking gbt_stacking.launch.py \
  fake:=True \
  robot_type:=<C5A/C7A/C12A/C16A>
  
# terminal 2 (manul trigger)
source install/setup.bash
ros2 service call /gbt_stacking/external_trigger  gbt_stacking_interface/srv/ExternalStackingTrigger "{trigger: True}"

---

Vision System Integration

Using AgileGaze

1. Camera Calibration:
   • Calculate coordinate transformation using AgileGaze calibration board
2. Workspace Setup:
   • Define grasp workspace origin in base coordinates
   • Update uf_grasp_points in stacking_params.yaml
3. Placement Area Setup:
   • Define placement origin in base coordinates
   • Update uf_place_points in config
4. Initial Pose:
   • Configure safe joint angles init_joints

Custom Vision Integration

# Custom vision node template
import rclpy
from gbt_stacking_interface.srv import GetAgileGaze

class CustomVisionNode(Node):
    def __init__(self):
        super().__init__('custom_vision')
        self.srv = self.create_service(
            GetAgileGaze, 
            '/gbt_vision/service/AgileGaze', 
            self.vision_callback)
    
    def vision_callback(self, request, response):
        # Implement vision logic
        response.agile_gaze = ... 
        return response

---

Important Notes

1. Safety:
   • Test first in simulation mode
   • Confirm emergency stop availability
   • Ensure clear workspace
2. Configuration:
   • Verify all YAML configuration files before startup