Nicla Vision ROS package

:rocket: Check out the ROS2 version: Nicla Vision ROS2 repository :rocket:

Description

This ROS package enables the Arduino Nicla Vision board to be ready-to-use in the ROS world! :boom:

The implemented architecture is described in the above image: the Arduino Nicla Vision board streams the sensors data to a ROS-running machine through TCP/UDP socket or Serial, i.e. UART over the USB cable. This package will be running on the ROS-running machine, allowing to deserialize the received info, and stream it in the corresponding ROS topics

Here a list of the available sensors with their respective ROS topics: - 2MP color camera streams on - /nicla/camera/camera_info - /nicla/camera/image_raw - /nicla/camera/image_raw/compressed - Time-of-Flight (distance) sensor streams on: - /nicla/tof - Microphone streams on: - /nicla/audio - /nicla/audio_info - /nicla/audio_stamped - /nicla/audio_recognized - Imu streams on: - /nicla/imu

The user can easily configure this package, by launch parameters, to receive sensors data via either UDP or TCP socket connections, specifying also the socket IP address. Moreover, the user can decide which sensor to be streamed within the ROS environment. In this repository you can find the Python code optimised for receiving the data by the board, and subsequently publishing it through ROS topics.

Table of Contents

1. Installation
2. Usage
3. Package List
4. Video
5. License
6. Cite

Installation

Step-by-step instructions on how to get the ROS package running

Binaries available for noetic:
sudo apt install ros-$ROS_DISTRO-nicla-vision-ros
For ROS2, check https://github.com/ADVRHumanoids/niclavisionros2.git

Source installation

Usual catkin build : bash $ cd <your_workpace>/src $ git clone https://github.com/ADVRHumanoids/nicla_vision_ros.git $ cd <your_workpace> $ catkin build $ source <your_workpace>/devel/setup.bash

Arduino Nicla Vision setup

After having completed the setup steps in the Nicla Vision Drivers repository, just turn on your Arduino Nicla Vision. When you power on your Arduino Nicla Vision, it will automatically connect to the network and it will start streaming to your ROS-running machine.

Optional Audio Recognition with VOSK

It is possible to run a speech recognition feature directly on this module, that will then publish the recognized words on the /nicla/audio_recognized topic. At the moment, VOSK is utilized. Only Arduino version is supported.

VOSK setup

1. pip install vosk
2. Download a VOSK model https://alphacephei.com/vosk/models
3. Check the recognition arguments in the nicla_receiver.launch file

Usage

Follow the below steps for enjoying your Arduino Nicla Vision board with ROS!

Run the ROS package

• Launch the package: bash $ roslaunch nicla_vision_ros nicla_receiver.launch receiver_ip:="x.x.x.x" connection_type:="tcp/udp/serial" <optional arguments>

  • Set the socket type to be used, either TCP or UDP (connection_type:=tcp, udp, or serial).
  • Set the receiver_ip with the IP address of your ROS-running machine. You can get this IP address by executing the following command: bash $ ifconfig and taking the "inet" address under the "enp" voice. Note argument is ignored if connection_type:="tcp/udp/serial

  Furthermore, using the <optional arguments>, you can decide: - which sensor to be streamed in ROS (e.g. enable_imu:=true enable_range:=true enable_audio:=true enable_audio_stamped:=false enable_camera_compressed:=true enable_camera_raw:=true), with Serial connection, set camera_receive_compressed according to how nicla is sending images (COMPRESSIMAGE) - on which socket port (default `receiverport:=8002). For Serial connection,receiverportargument is used to specify the port used by the board (e.g.receiverport:=/dev/ttyACM0`)

  Once you run it, you will be ready for receiving the sensors data

Simulated board

• For simulating the Arduino Nicla Vision in Gazebo and Rviz: bash $ roslaunch nicla_vision_ros nicla_sim.launch <optional arguments> Using the <optional arguments>, you can decide if to run the simulation in Gazebo or in Rviz, and which sensor to simulate (everything set to true as default).

Optional Micropython Version

Note this version is not supported and maintained anymore, we defitively moved to the Arduino version.

If you want to use the micropython driver of the nicla (after setup it accordingly), simply run the nicla_receiver.launch with driver_version:=micropython as argument.

Video Demonstration

https://github.com/ADVRHumanoids/niclavisionros/assets/26459008/a3eaf921-02ea-4482-80a0-5830a338eb74

Package List

Here some useful links:

• Nicla Vision Drivers repository
• Nicla Vision ROS repository
• Nicla Vision ROS2 repository

License

Distributed under the Apache-2.0 License. See LICENSE for more information.

ROSCon24 Talk

• Video https://vimeo.com/1026030054
• Slides https://roscon.ros.org/2024/talks/UnlockingthePotentialoftheNiclaVisionBoardwith_ROS-ROS2.pdf

Citation

:raised_hands: If you use this work or take inspiration from it, please cite (to be published): bibtex @inproceedings {DelBianco2024, author = {Del Bianco, Edoardo and Torielli, Davide and Rollo, Federico and Gasperini, Damiano and Laurenzi, Arturo and Baccelliere, Lorenzo and Muratore, Luca and Roveri, Marco and Tsagarakis, Nikos G.}, booktitle={2024 IEEE-RAS 23rd International Conference on Humanoid Robots (Humanoids)}, title = {A High-Force Gripper with Embedded Multimodal Sensing for Powerful and Perception Driven Grasping}, year={2024}, volume={}, number={}, pages={149-156}, doi={10.1109/Humanoids58906.2024.10769951} }