Repository: https://github.com/Shass27/TurtlePursuit
TurtlePursuit is a robotic simulation framework developed within the Robot Operating System 2 (ROS 2) environment. The project simulates a dynamic pursuit-evasion scenario wherein an autonomous agent (the "Hunter") identifies, tracks, and intercepts dynamically generated targets (the "Prey") within a bounded 2D continuous space. This system demonstrates fundamental concepts in autonomous robotics, including node orchestration, coordinate frame transformation, feedback control loops, and service-based interaction for entity management.
The system is composed of a modular node architecture leveraging the ROS 2 communication graph. The interaction between components is achieved through a combination of asynchronous topics and synchronous service calls.
/turtle_spawner: A stochastic generator that spawns target entities at random coordinates(x, y, θ)within the simulation bounds. It maintains a state registry of active targets and broadcasts this manifest to the wider system via custom interfaces./turtle_controller: The central processing unit for the autonomous agent. It subscribes to the global target state and the agent's own odometry to compute kinematic control commands at a frequency of 100 Hz.turtlesim: The simulation backend responsible for rendering the environment and integrating the kinematic equations of motion.
/turtles
Type: interfaces/msg/TurtleArray
Carries the real-time manifest of all active target turtles.
/turtle1/pose
Type: turtlesim/msg/Pose
Provides telemetry (position and orientation) of the Hunter agent.
/turtle1/cmd_vel
Type: geometry_msgs/msg/Twist
Transmits computed velocity commands to the Hunter's actuator interface.
/spawn: Invoked by the spawner node to instantiate new targets in the simulator./kill: Invoked by the controller node to remove intercepted targets from the simulation.
The autonomous behavior is governed by a Nearest-Neighbor selection algorithm coupled with a Proportional (P) Controller for kinematic regulation.
The agent operates in a continuous control loop where it evaluates the Euclidean distance d_i between its current position (x_h, y_h) and every i-th active target (x_i, y_i):
d_i = sqrt((x_i - x_h)^2 + (y_i - y_h)^2)
The target T_target is selected such that it minimizes the distance metric:
T_target = argmin(d_i)
Once a target is locked, the controller generates velocity commands based on the error state.
The linear velocity is proportional to the distance error, ensuring the agent decelerates as it approaches the interception point to minimize overshoot.
v(t) = K_p_linear · d(t)
Parameter: K_p_linear = 2.0
The angular velocity is driven by the heading error α, defined as the difference between the desired heading angle (towards the target) and the current agent orientation θ_h:
α = atan2(y_target - y_h, x_target - x_h) - θ_h
The error α is normalized to the range [-π, π] to ensure optimal turning direction. The control law is:
ω(t) = K_p_angular · α
Parameter: K_p_angular = 6.0
An interception is successfully registered when the distance d(t) falls below a critical threshold ε:
d(t) < 0.1 ⇒ State: Intercepted
Upon interception, the agent halts (v = 0, ω = 0), issues a service request to remove the entity from the simulation, and updates the internal state registry to prevent re-engagement.
Operating System: Ubuntu 24.04.3 LTS (Noble Numbat)
Middleware: ROS 2 Jazzy Jalisco
Language: Python 3.12+
IDE: Visual Studio Code (VS Code)
Build System: Colcon (Collective Construction)
Clone the repository into your ROS 2 workspace src directory and build using colcon:
mkdir -p ~/ros2_ws/src
git clone https://github.com/Shass27/TurtlePursuit
mv TurtlePursuit/src/* .
rm -rf TurtlePursuit
cd ..
colcon build
source install/setup.bash
The project utilizes a centralized XML launch file to orchestrate the simulation nodes and load configuration parameters:
ros2 launch bringup catch_all.launch.xml
Simulation parameters can be adjusted via catch_all_config.yaml or passed as launch arguments:
spawn_frequency: Rate at which new targets appear (Hz)turtle_prefix: Naming convention for generated targets