VINS Auto Launcher

This ROS package automates the process of:

1. Launching VINS-Mono (vins_estimator/android.launch)
2. Starting an odometry recorder node or activity predictor node
3. Playing a specified ROS bag file
4. Automatically launch vins-mono and start prediction of the activity in real-time with pre built ML model.

Odometry or activity prediction data is saved or published as appropriate.

---

Prerequisites

• ROS Melodic (tested on Ubuntu 18.04, Python 2.7)
• VINS-Mono installed in your catkin_ws/src
• rosbag installed (comes with ROS by default)
• joblib, numpy (for predictor node, install via pip if needed)

---

Installation

cd ~/catkin_ws/src
git clone https://github.com/sajjad-hm/vins-auto-launcher.git
cd ..
catkin_make
source devel/setup.bash

---

Usage

Odometry Recording Launch Command

Run the full process to record the raw_data feature (VINS-Mono → Bag playback → Feature Recorder):

roslaunch vins_auto_launcher full_pipeline.launch bag_file:=/absolute/path/to/your_file.bag

Example:

roslaunch vins_auto_launcher full_pipeline.launch bag_file:=/home/user/data/sample.bag

Activity Prediction Launch Command

Run the full process with activity prediction (VINS-Mono → Bag playback → Activity Prediction):

roslaunch vins_auto_launcher full_pipeline_predict.launch bag_file:=/absolute/path/to/your_file.bag

Example:

roslaunch vins_auto_launcher full_pipeline_predict.launch bag_file:=/home/user/data/sample.bag

---

---

What Happens

full_pipeline.launch

1. Launches VINS-Mono (vins_estimator/android.launch)
2. Waits 3 seconds
3. Starts vins_odom_recorder.py:
   • Saves odometry from /vins_estimator/odometry
   • Stores CSV in the same folder as the .bag file
   • CSV filename: <bag_name>_<timestamp>.csv
4. Waits 3 seconds
5. Plays the provided .bag file

full_pipeline_predict.launch

1. Launches VINS-Mono (vins_estimator/android.launch)
2. Waits 3 seconds
3. Starts predictor_node.py:
   • Loads ML model and scaler from models/
   • Listens to /vins_estimator/odometry
   • Publishes activity prediction to /activity_prediction
4. Waits 6 seconds
5. Plays the provided .bag file

---

Monitor Activity Prediction in Real-Time

rostopic echo /activity_prediction

CSV Format (Odometry Recorder)

The CSV file contains:

Position_x	Position_y	Position_z	Orientation_x	Orientation_y	Orientation_z	Orientation_w

Example:

0.123, -0.456, 1.789, 0.0, 0.0, 0.707, 0.707

---

File Structure

vins_auto_launcher/
├── CMakeLists.txt
├── package.xml
├── launch/
│   ├── full_pipeline.launch
│   └── full_pipeline_predict.launch
├── scripts/
│   ├── vins_odom_recorder.py
│   └── predictor_node.py

---

Tips

• You must provide an absolute path for bag_file in the launch command.
• The package will not overwrite CSV files from previous runs — each has a unique timestamp.
• Ensure vins_estimator is built and link the location of vins_eastimator in the launch file and can run android.launch independently before using this package.
• For activity prediction, ensure the required model and scaler files are present in the models/ directory.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.