DeepField Workshop 2022+2023

Exploiting (supervised) learning- based detectors for on-board, real time localization and target tracking

Installation:

1. Clone repository with submodules (git clone --recursive)
2. Install ROS noetic on Ubuntu 20.04 LTS

Tutorial:

Task: Detect and localize a cup with a simulated robot (using the laptop webcam)

• install requirements (ROS noetic)
• run network detection server
• run detection and tracking with launch file packages/simulation/aircap/launch/detect_cup_with_webcam.launch
• have neural network detector (pytorch) detect the cup and localize in ROS
• visualize with rviz deefield_workshop.rviz

Folders:

• ./Slides: lecture slides and system presentation
• ./ssd_pytorch_code - submodule - python code for neural network detector service - used by AirCap simulation tutorial code
• ./AirCap_code - submodule - Simulation Tutorial Source Code - bases for exercise