Milestone 1

Mission Statement & Scope

Mission Statement

The mission of this project is to develop an autonomous mobile manipulation system capable of locating, retrieving, and dynamically throwing a ball in a simulated environment.

The system integrates:

• Global beacon-based localization
• Vision-based object detection
• Autonomous navigation
• Robotic manipulation

All components operate within a ROS 2 and Gazebo simulation framework.

The robot must autonomously:

1. Detect the ball
2. Navigate toward it
3. Pick it up using a vacuum-based gripper
4. Execute a dynamic throwing action toward a predefined target location

• A TurtleBot mobile base
• A robotic manipulator arm
• A Gazebo-based simulation environment

Scope

This project focuses on the integration of perception, estimation, planning, and actuation modules required for mobile manipulation.

The robotic platform consists of:

• A TurtleBot mobile base
• A Robotic manipulator arm
• A Gazebo-based simulation environment

Perception System

The perception stack combines:

• Beacon-based localization to estimate the global position of the ball
• Camera-based vision module to refine the ball’s local pose for manipulation

System Capabilities

The system implements:

• Autonomous Navigation to approach the ball
• Precise base alignment to ensure the object lies within the manipulator workspace
• A custom dynamic throwing mechanism to return the ball to a target location

Custom Algorithmic Modules

Two task-specific control algorithms are introduced:

• Center-of-Gravity (CG) Stability Controller
  Dynamically adjusts manipulator posture during base motion to maintain system stability.

• Vacuum Throw Release Controller
  Synchronizes suction release with peak tangential velocity during the arm swing to maximize throwing performance.

Environment Assumptions

The current system operates in an open indoor simulated environment without obstacles.

This allows the project to focus on subsystem integration and coordination. Future extensions may introduce:

• Obstacle-rich environments
• More advanced manipulation tasks

The platform consists of a TurtleBot mobile base integrated with a robotic manipulator arm, enabling autonomous object retrieval and dynamic throwing.

Technical Specifications

Robot Platform

The robotic system is implemented using:

• ROS 2 as the middleware framework
• Gazebo for physics-based simulation

The platform consists of a TurtleBot mobile base integrated with a robotic manipulator arm, enabling autonomous object retrieval and dynamic throwing.

Software Frameworks

The system integrates the following ROS 2 packages and tools:

• Gazebo – Physics-based simulation environment
• ROS 2 – Distributed robotic communication framework
• Nav2 – Autonomous navigation stack
• MoveIt2 – Manipulation planning and trajectory execution
• TF2 – Coordinate frame transformation system
• RViz2 – Visualization and debugging interface
• OpenCV – Camera-based object detection

The robot operates in an open environment to simplify navigation and emphasize perception–manipulation integration.

High-Level System Architecture

The system follows a Perception → Estimation → Planning → Actuation pipeline.

1. Beacon localization provides a coarse estimate of the ball position.
2. A goal generation module computes the navigation target.
3. The Nav2 navigation stack drives the robot toward the ball.
4. The vision module detects and refines the ball pose.
5. The base alignment module positions the robot for grasping.
6. The manipulator grasps the ball using a vacuum gripper.
7. The throwing planner computes a slinging trajectory.
8. The release controller disengages suction at the optimal moment.

---

System Architecture

System Modules

Custom Modules

CG Stability Controller

Dynamically adjusts the manipulator posture to maintain stable center-of-gravity positioning during base motion.

Problem: When the TurtleBot accelerates or decelerates, inertial forces shift the combined center of mass of the base-arm system, reducing traction and destabilizing the robot.

Solution: The controller estimates the projected center of gravity and adjusts the arm configuration to keep the CoM within a stable support region — reducing load transfer and improving traction consistency at the drive wheels.

Center of Gravity Estimation:

\[x_{cg} = \frac{\sum m_i x_i}{\sum m_i}\]

---

Vacuum Throw Release Controller

Implements a custom dynamic manipulation algorithm for throwing the ball using a vacuum suction gripper.

How it works:

1. The manipulator performs a forward slinging motion, accelerating the end effector
2. The algorithm monitors the arm trajectory
3. At the moment of maximum tangential velocity aligned with the desired throw direction, vacuum suction is disengaged
4. The ball detaches and travels along a ballistic trajectory

Key equations:

Tangential velocity: \(v_t = r\omega\)

Projectile range approximation: \(R = \frac{v^2 \sin(2\theta)}{g}\) By synchronizing release timing with arm motion, the system maximizes throwing range and repeatability.

---