Home

Awesome

Delft AI Toolkit

Visual Authoring for AI Prototyping

UPDATE as of March 15, 2021

NOTE: We've just released version 3.1.0. The new version includes the following changes:

Issues:

Overview

The Delft AI Toolkit is a system for designing and researching smart things. It provides a visual authoring environment that simplifies the use of machine learning, cognitive APIs, and other AI approaches. The node-and-connection visual graphs combine the advantages of behavior trees and dataflow models to create smart behavior in autonomous devices.

The Delft AI Toolkit enables quick prototyping, experimentation and iteration of AI interactions with powerful nodes that support behavioral sequences, sensing, decision making, state management, and the simple invocation of AI techniques such as speech to text, text to speech, and visual object recognition. In addition, it encourages design strategies that start with simulation and smoothly progress toward a working prototypes of machine learning and hardware.

The goal of this project is to develop an approach to authoring AI that enables designers to easily and iteratively sketch working smart things. This approach includes the ability to Wizard-of-Oz AI behaviors and simulate physical hardware in 3D, and then migrate these simulations to working prototypes that use machine learning and real hardware.

<!-- TOC START min:2 max:3 link:true asterisk:false update:true --> <!-- TOC END -->

Video Introduction

<img src="docs/images/toolkitintro.jpg" width="700">

System Diagram

system diagram

Documentation

System Components

Visual Authoring & Control System running on a computer

Robot/Device

Unity and the robot have a codebase here, and these include a range of open source libraries. This system also uses xNode, which is being enhanced by Siccity as part of this project.

Current Features

<img src="docs/images/speak-recognize-graph.png" width="700">

Future Roadmap as of April 2021

System Architecture

system architecture

Hardware

The physical robot is currently based on a simple robot platform from Adafruit, and uses a Raspberry Pi 4 to perform the local edge AI, local text-to-speech, and make use of cloud APIs. The RPi uses the Adafruit CRIKIT Raspberry Pi Hat to support audio, NeoPixels, DC motors, and Servos. The RPi communicates over WiFi to Unity on the computer with the OSC network protocol.

More details on the hardware.

<img src="docs/images/robot2020-1.jpg" width="512"><br><br> <img src="docs/images/robot2020-5.jpg" width="512"><br><br> <img src="docs/images/robot2020-3.jpg" width="512"><br><br>

Raspberry Pi Disk Image

Getting Started

Starting the System Up

  1. Power the Robot: Power on the Raspberry Pi (RPi):

    • RPi: Connect a 5V 2A AC adapter or a USB battery to the barrel jack on the CRICKIT -- this will also power the RPi

  2. Get the WiFi IP address of the Robot

    • Your RPI must already connect to WiFi - If you haven't already, set up your RPI to connect to your local WiFi using an ethernet cable connecting a computer to the RPi.
    • Connect by Ethernet if necessary - Some institutional networks don't allow connection by the .local (Bonjour) name, and you may need to use the IP address of the RPi. If so, hook up an ethernet cable between your computer and the Raspberry Pi (RPi) on the robot (you may need a USB-C adapter to get the IP.
    • Login to the RPi - Open a terminal window on your computer, and log into the RPi by typing in the below. Change the number at the end of delftbt0 to match the number of your robot if you changed the name.
    ssh pi@delftbt0.local
    • Password - The default password for the standard Delft AI Toolkit disk image is "adventures"
    • Get the RPi WiFi IP Address - Once logged in, copy and save the IP address of the RPi by typing "ifconfig" as below. You'll see an entry for "wlan0" which is the WiFi connection - from there copy the IP address (e.g. 10.4.27.47)
    ifconfig
    • Logout - of the RPI by typing
    exit
    • Disconnect - Unplug the ethernet cable

  3. Start the Delft Toolkit software on the RPi

    • Login to RPi from your computer - In a terminal app log in to the RPi over WiFi by typing:
    ssh pi@10.4.27.47 # replace this IP address with the one you got from ifconfig above
    • Start the robot software - After logging in, type the below command, putting the IP address of your Unity computer at the end (get the IP address of your computer by opening the Network Preferences panel)
    ./delft-ai-toolkit/start-delft-toolkit.sh 192.168.1.21 hello

    • Startup Sequence - The software will take a little time to start up. When it finishes, the robot will say "hello" (or whatever text you put at the end of the line)

    • IMPORTANT: Powering off the RPi Procedure: Before you disconnect the power from the RPi, to prevent harm to the SD storage, you must properly shut it down with the following command. Wait for 10 seconds after the poweroff command, then it is safe to unplug the power from the RPi.

    sudo poweroff

  4. Start the software running in Unity

    • Open Project - Open the "delft-toolkit" project in Unity3D (minimum version is 2019.4.22f)
    • Open Scene - In the Project, open the the scene that matches the toolkit graph you are using (e.g. Assets>Scenes>MainExamples)
    • Open a Visual Graph - In the Unity Project Assets>DelftToolkitGraphs directory, double click on the toolkit visual graph you are currently using (e.g. Assets>DelftToolkitGraphs>MainExamples)
    • Physical Robot - If you are using the physical robot (the Unity toolkit will work fine without the robot):
      • From the Unity menubar, select Delft AI Toolkit>Show Settings. This will select the central settings document (Assets>Resources>DelftAIToolkitSettings) and you can make changes in the Inspector panel.<br><img src="docs/images/settings.png" width="310">
      • In the inspector for "ding1", paste in the IP address of the robot where it says "Robot IP"
      • If you are not using the physical robot, leave the Ding IP address set to 127.0.0.1
    • IBM Watson - If you are using any of the Watson services (e.g. TextToSpeech or SpeechToText), these are usable on both the computer and on the robot (selectable). To use these services, you'll need an IBM Cloud account (free for limited use).
      • Sign up for an account - http://cloud.ibm.com
      • Set up the service(s) you want to use in the toolkit
      • Create credentials for using the service in the API Key format (also called IAM Key).<br><img src="docs/images/watson-credentials.png" width="370">
      • Enter the API Key/IAM Key in the corresponding field in the toolkit settings asset. Delft AI Toolkit>Show Settings, then open the inspector panel in Unity: Window>General>Inspector
    • Start Robot Software - If you are using the robot, you need to start the software running on the robot before starting up the graph in Unity with the PLAY button. This allows Unity to initialize the Watson software on the robot.
    • Play - Click on the Unity Play button
    • Start Graph - In the xNode Toolkit graph pane, click on the "Start" node Trigger button to run the whole graph, or use Trigger on any individual node
      • For keyboard or OSC - Click on the Game pane (this is to ensure Unity is receiving all commands -- if you find it is not responding to the keyboard or OSC, click this pane)

Essential Robot Linux Commands

# login to the RPi via ethernet
ssh pi@delftbt0.local
# get the IP address of the RPi from the wlan0 section
ifconfig
# login to the RPi via WiFi, change the example IP to that RPi
ssh pi@10.4.27.47
# start the RPi software, change the example IP to that of the computer running Unity
./delft-ai-toolkit/start-delft-toolkit.sh 192.168.1.21 hello
# shutdown before disconnecting the power, then wait for 10 seconds
sudo poweroff

Installing The Software

See installation instructions on the hardware documentation page