ALU - Autonomous Lung Ultrasound

Overview

ALU (Autonomous Lung Ultrasound) is a project designed to autonomously perform lung ultrasound scanning using a robotic system.

Installation

Clone the repository:

git clone https://github.com/yourusername/yourproject.git

Navigate to the directory:

cd /ALU-Autonomous-Lung-Ultrasound

Install the required packages

pip install -r requirements.txt

Prerequisites

Ensure you have ROS2 Humble installed. You can follow the installation instructions here.

Usage

Setting the Controllers

To set up the controllers, follow the instructions in this repository.

Note: By default, the controllers do not publish the current robot position information. To enable this feature, you need to modify the following flags in the cartesian_controller_base/src/cartesian_controller_base.cpp file:

//Line 104
auto_declare<bool>("solver.publish_state_feedback", false);
m_initialized = true;

//Line 129
auto_declare<int>("solver.iterations", 1);
auto_declare<bool>("solver.publish_state_feedback", false);

Both to:

auto_declare<bool>("solver.publish_state_feedback", true);

Build Packages

To build the packages navigate to ros2Packages and do:

colcon build --symlink-install

Don't forget to source both the controller workspace and this one.

Launching the controller

Note For the control logic to work correctly if using another end-effector you need to update the URDF in the ur_description ros2 package.

To run the simulation controller using CoppeliaSim:

ros2 launch ur_coppeliasim ur_coppelia_controllers.launch.py

To launch the controller for the real robot:

ros2 launch ur-launch ur_compliance_controller.launch.py

The controller being used is the cartesian_compliance_controller which allows the robot to not excert excesive force onto the phantoms.

Running the Lung Ultrasound Nodes

Two nodes were created, one for the acquisition of the data necessary for the development of the system, and the other one for the implementation of said system.

To start data acquisition node:

ros2 run lung_us talker

The parameters for the movement can be changed when the program is running, but the base and center target position has to be updated in the aks.py file

To start the autonomous acquisition nodes:

1. Move the robot to a desired starting position, assuring contact with the skin
2. Start the matlab ros2 node that constantly acquires the images and sends them. To do this run in matlab the file ros2autoAq.m
3. Start the acquisition node:

ros2 run lung_us listener

The system will perform N number of bayesian optimization cycles and early stop if the last 5 acquisitions have an error change of less than 0.02.

Additional things

Re-run the Robot Program

To resend the robot program if it closes:

ros2 service call /io_and_status_controller/resend_robot_program std_srvs/srv/Trigger {}