Andrea F. Daniele

Chief Technology Officer  •  Robotics Engineer  •  Computer Scientist




I am a Computer Engineer and currently the Chief Technology Officer at Duckietown, Inc.

I received my Ph.D. in Computer Science from the Toyota Technological Institute at Chicago (TTIC), a philanthropically endowed academic computer science institute located on the University of Chicago campus. I was a member of the Robot Intelligence through Perception Laboratory (RIPL) where I worked under the supervision of Prof. Matthew Walter.

I received a Master's Degree in Computer Science from the Toyota Technological Institute at Chicago (TTIC), as part of the Ph.D. program.

I received a Master's Degree in Artificial Intelligence and Robotics from the University of Rome - ”La Sapienza”, (Italy), where I worked under the advisement of Prof. Daniele Nardi.

I received a Bachelor’s Degree in Computer Engineering from the University of Calabria - UNICAL (Italy).

I am interested in developing robots that are able to move autonomously in unstructured environments and work alongside people. My research focused mainly on human-robot cooperation and coordination based on natural language interaction as well as fleet-level robot-robot communication and coordination in the development of smart cities and self-driving vehicles.




Work Experience


Chief Technology Officer

Nov. 2022 - present

I took the position of CTO at Duckietown, where I lead the design and development of all new Duckietown robots and software products.

Duckietown, Inc.

Boston, MA, USA

Lead Software Engineer

Jan. 2021 - Nov. 2022

Officially joined the Duckietown Team as Lead UX Software Engineer. I led the effort of improving and developing new software modules aimed at improving the user experience (UX) on the Duckietown robots.

Duckietown - Duckieworks AG

Zürich, Switzerland

Software Engineer Intern

Jan. 2019 - May 2019

During a 4-months internship as a Software Engineer in the Simulation Team at drive.ai, I worked on the problem of generating behaviors for dynamic simulated agents using semantically rich graphs computed from annotated HD maps.

drive.ai

Mountain View, CA, USA

Software Engineer

Sep. 2017 - Jan. 2021

I contributed to the Duckietown project as a volunteer at the Duckietown Foundation. I led the effort of improving accessibility in robot programming, debugging, and monitoring with the development of better user interfaces.

The Duckietown Foundation

Chicago, IL, USA

Software Engineer

Aug. 2013 - Jan. 2016

Developed and implemented Web-based applications, websites, web-APIs, and interactive applications for desktop environments and mobile devices.

Cloud4Service.net

Petilia Policastro, KR, Italy





Education


Ph.D. in Computer Science

Sep. 2016 - Sep. 2023

Toyota Technological Institute at Chicago

Chicago - IL - USA

M.S. in Computer Science

Sep. 2016 - Sep. 2019

Toyota Technological Institute at Chicago

Chicago - IL - USA

M.S. in Artificial Intelligence and Robotics

Oct. 2013 - Dec. 2016

University of Rome - La Sapienza

Rome - RM - Italy

B.S. in Computer Engineering

Oct. 2009 - Jul. 2013

University of Calabria - UNICAL

Rende - CS - Italy





Portfolio


Duckiematrix

A high-performance robot simulator

Roles: Lead Architect Main Developer

Current Owner / License: Duckietown, Inc.

The Duckiematrix is a high-performance robot simulator for the Duckietown platform that allows for the simulation of large fleets of robots in real-time. It is based on the Unity game engine.

The Duckiematrix supports the concurrent simulation of multiple robots, such as ground vehicles, drones, smart city infrastructure, and all their sensors and actuators.

It is also designed to support the use of multiple renderers to distribute the rendering load across multiple GPUs or nodes across the network. Learn more from the official documentation.


DD24 Autonomous Drone

A state-of-the-art educational autonomous drone

Roles: Project Leader

Current Owner / License: Duckietown, Inc.

The DD24 autonomous drone is a state-of-the-art educational platform by Duckietown, Inc. With the DD24, we wanted to create a drone that was easy to use, program, debug, and understand. At the same time, we wanted a drone that was powerful enough to be used in advanced robotics courses and research projects.

Compared to its predecessor, the DD21, the DD24 is 36% more powerful, 30% smaller, has 3x the number of sensors, and new obstacle detection capabilities with 270deg coverage.

Learn more from the official documentation.


Interactive Robot Assembly Tool

An interactive tool for assembling robots in 3D

Roles: Lead Architect Main Developer

Current Owner / License: Duckietown, Inc.

The Interactive Robot Assembly Tool is a web-based tool that allows users to assemble physical robots while following the step-by-step assembly process in 3D. The view can be rotated, zoomed, and panned to provide a better understanding of all the steps that make up the assembly process.

The Interactive Robot Assembly Tool is based on the Unity game engine. This provides a photo-realistic rendering engine in which high-fidelity 3D models of the robot components can be used.

Learn more from the official video and live example.


SHARC

Shared autonomy for robotic underwater exploration

Roles: Co-Inventor

Current Owner / License: Patented

Conventional underwater intervention operations using robotic vehicles require expert teleoperators on a support vessel deployed nearby the intervention site.

SHARC (SHared Autonomy for Remote Collaboration) is a framework that enables operators to cooperatively conduct robotic underwater sampling and manipulation tasks from thousands of miles away. With SHARC, operators can execute manipulation tasks using natural language or hand gestures through a virtual reality interface. SHARC is readily extensible to other tasks and domains such as planets exploration.

Learn more from our scientific publication or the project page.


The Duckietown Dashboard

A browser-accessible control center and robot data visualizer

Roles: Lead Architect Main Developer

Current Owner / License: Duckietown, Inc.

The Duckietown Dashboard is a browser-accessible control center and robot data visualizer for the Duckietown platform.

I started the development of the Dashboard in 2017 as a tool to help students and researchers to interact with the Duckietown robots in a more intuitive way. It is now a key component of the Duckietown platform and is used by thousands of users worldwide. The Duckietown Dashboard is created using the \compose\ framework.

Learn more from the official documentation.


aavm

Almost A Virtual Machine

Roles: Lead Architect Main Developer

Current Owner / License: MIT License

While Virtual Machines are a powerful tool for isolated deployments, they are often too heavy to justify their use for simple applications. The Almost A Virtual Machine (aavm) is a lightweight alternative to full-blown VMs that allows you to run applications in sandboxed environments with minimal overhead.

aavm uses Docker containers configured with an isolated instance of systemd to provide a full Linux environment for your applications. An example use case is that of running applications that usually require a full desktop environment on headless systems, e.g., Unity applications. aavm can, in fact, run instances of the X server with GPU access on systems without a display.

Learn more from the official GitHub repository.


\compose\

A lightweight web-based CMS for robot telemetry

Roles: Lead Architect Main Developer

Current Owner / License: MIT License

\compose\ is an open-source web-based CMS platform that provides all the functionalities needed for the fast development and deployment of web applications. In particular, it focuses on the development of web applications that require real-time data visualization and interaction.

This makes it the ideal framework for the development of robot telemetry dashboards and control centers. An example use case is the Duckietown Dashboard. \compose\ is written in PHP and designed to be lightweight, modular, and easy to use.

Learn more from the official GitHub repository.


cpk

the Code Packaging toolKit

Roles: Lead Architect Main Developer

Current Owner / License: MIT License

cpk stands for Code Packaging toolKit and is designed to standardize the way code in a project is structured and packaged for maximum portability, readability and maintainability. cpk is the result of years of experience in cross-user, cross-machine, cross-architecture development and deployment of software modules.

cpk organizes code in projects. A cpk project is a directory containing everything that is needed for the project to be built, packaged, documented and deployed. An easy-to-use command-line interface is provided to manage the project. It leverages the power of Docker to provide consistent and reproducible, yet dynamic development and deployment environments.

Learn more from the official GitHub repository.






Events


    

Arctic Code Vault

The cold storage that will last 1,000 years

Some of my open-source contributions were selected for archival in the Arctic World Archive (AWA), an archival facility designed to preserve data for 1,000 years.
A total of 186 reels of film were stored in a steel-walled container, inside a sealed chamber, within a decommissioned coal mine, buried deep into the permafrost of the Svalbard archipelago, in Norway. Internationally recognized as a demilitarized zone, Svalbard is the world’s northernmost town and one of the most remote and geopolitically stable human habitations on Earth.


2021


    

SHARC: SHared Autonomy for Remote Collaboration at WHOI

August 2021 - October 2021

I worked on the development of "SHARC: SHared Autonomy for Remote Collaboration" at the Woods Hole Oceanographic Institution. SHARC is a multi‑modal interface that enables remote scientists to perform high‑level tasks using an underwater manipulator, while deferring low‑level control to the robot. We successfully deployed and tested SHARC during the OECI Technology Demonstration: Nereid Under Ice (NUI) Vehicle + Mesobot expedition aboard E/V Nautilus.


    

MOOC course on edX called 'Self-Driving Cars with Duckietown'

March 2021 - August 2021

I co-organized the first hardware based massive online open course (MOOC) in AI and robotics, free on edX. Aimed at teaching autonomy hands-on by making robots that can take their own decisions and accomplish broadly defined tasks. The course guides learners step-by-step from the theory, to the implementation, to the deployment in simulation as well as on real robots (Duckiebots).


2019


    

Internship at drive.ai - Software Engineer - Simulation Team

January 2019 - May 2019

Leveraging static annotations and HD maps, we created a semantically rich lane-level graph, that simulated agents can use to navigate the map. This lane graph is comprised of three layers: topological, metrical, and semantic. Topological and metrical layers were extracted from HD maps and static annotations, while the semantic layer was constructed from annotations only.


2018


    

Workshop on Models and Representations for Natural Human-Robot Communication at the RSS18 Conference

June 2018

Together with my adviser Prof. Matthew Walter, Jacob Arkin, Nakul Gopalan, Prof. Thomas Howard, Jesse Thomason, and Lawson Wong, I organized the Workshop on Models and Representations for Natural Human-Robot Communication (MRHRC) at the RSS (Robotics: Science and Systems) 2018 conference.


    

UR5-equipped Robot playing Checkers, 2018 National Robotics Week at MSI

April 2018

Under the supervision of my adviser Prof. Matthew Walter, I and other colleagues showed our UR5-equipped Husky A200 robot safely playing Checkers against human opponents at the Museum of Science and Industry in Chicago for the 2018 National Robotics Week exhibit. This robot is developed in the RIPL lab at TTI-Chicago.


2017


    

Duckietown 2017

September 2017

In Autumn 2017 I joined the international team of Duckietown (a robotics education and outreach effort) working as a Teaching Assistant for the course TTIC 31240 - Self-driving Vehicles: Models and Algorithms for Autonomy taught by Prof. Matthew Walter at TTIC. Click on the blue button to see my contributions to the project.


    

Husky A200 robot for the National Robotics Week at the MSI-Chicago

April 2017

Under the supervision of my adviser Prof. Matthew Walter, I and other colleagues showed our Husky A200 robot at the Museum of Science and Industry in Chicago for the National Robotics Week. This robot is developed in the RIPL lab at TTI-Chicago as part of the Robotics Collaborative Technology Alliance (RCTA) research program. CBS 2’s Vince Gerasole interviewed me on that occasion.


2016


    

NLIGEN - Natural Language Instruction Generation

January 2016 - September 2016

Under the supervision of my adviser Prof. Matthew Walter, I developed a model that enables robots to generate natural language instructions that allow humans to navigate a priori unknown environments. The model first decides which information to share with the user according to their preferences, then “translates” this information into a natural language instruction.





Publications


A Shared Autonomy System for Precise and Efficient Remote Underwater Manipulation.
Amy Phung, Gideon Billings, Andrea F. Daniele, Matthew R Walter, and Richard Camilli.
IEEE, 2024.

[ pdf, bibtex, abstract ]


Enhancing scientific exploration of the deep sea through shared autonomy in remote manipulation.
Amy Phung, Gideon Billings, Andrea F. Daniele, Matthew R Walter, and Richard Camilli.
Science Robotics - Vol 8 / No 81, 2023.

[ pdf, bibtex, abstract ]


Ph.D. Thesis - Accessible Interfaces for the Development and Deployment of Robotic Platforms.
Andrea F. Daniele.
Ph.D. Thesis, 2023.

[ pdf, bibtex, abstract ]


Toward Efficient Under-Ice Exploration of Ocean Worlds Using Distributed Autonomy and 3D Workspace Reconstruction Presented in VR for Intuitive Understanding.
Amy Phung, Gideon Billings, Andrea F. Daniele, Matthew Walter, and Richard Camilli.
The Astrobiology Science Conference (AbSciCon) 2022, 2022.

[ pdf, bibtex, abstract ]


Language Understanding for Field and Service Robots in a Priori Unknown Environments.
Matthew R Walter, Siddharth Patki, Andrea F. Daniele, Ethan Fahnestock, Felix Duvallet, Sachithra Hemachandra, Jean Oh, Anthony Stentz, Nicholas Roy, and Thomas M. Howard.
Journal of Field Robotics (IJFR 2021), 2021.

[ pdf, bibtex, abstract ]


Integrated Benchmarking and Design for Reproducible and Accessible Evaluation of Robotic Agents.
Jacopo Tani, Andrea F. Daniele, Gianmarco Bernasconi, Amaury Camus, Aleksandar Petrov, Anthony Courchesne, Bhairav Mehta, Rohit Suri, Tomasz Zaluska, Matthew R. Walter, Emilio Frazzoli, Liam Paull, and Andrea Censi.
In Proceedings of the IEEE International Conference on Intelligent Robots and Systems (IROS 2020), July 2020.

[ pdf, bibtex, abstract ]


DIODE: A Dense Indoor and Outdoor DEpth Dataset.
Igor Vasiljevic, Nick Kolkin, Shanyi Zhang, Ruotian Luo, Haochen Wang, Falcon Z. Dai, Andrea F. Daniele, Mohammadreza Mostajabi, Steven Basart, Matthew R. Walter, and Gregory Shakhnarovich.
CoRR volume abs/1908.00463, August 2019.

[ pdf, bibtex, abstract ]


Inferring Compact Representations for Efficient Natural Language Understanding of Robot Instructions.
Siddharth Patki, Andrea F. Daniele, Matthew R. Walter, and Thomas M. Howard.
In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), 2019, May 2019.

[ pdf, bibtex, abstract ]


The AI Driving Olympics at NeurIPS 2018.
Julian Zilly, Jacopo Tani, Breandan Considine, Bhairav Mehta, Andrea F. Daniele, Manfred Diaz, Gianmarco Bernasconi, Claudio Ruch, Jan Hakenberg, Florian Golemo, A. Kirsten Bowser, Matthew R. Walter, Ruslan Hristov, Sunil Mallya, Emilio Frazzoli, Andrea Censi, and Liam Paull.
arXiv:1903.02503, March 2019.

[ pdf, bibtex, abstract ]


A Multiview Approach to Learning Articulated Motion Models.
Andrea F. Daniele, Thomas M. Howard, and Matthew R. Walter.
In Proceedings of the International Symposium of Robotics Research (ISRR), 2017, December 2017.

[ pdf, bibtex, abstract ]


Navigational Instruction Generation as Inverse Reinforcement Learning with Neural Machine Translation.
Andrea F. Daniele, Mohit Bansal, and Matthew R. Walter.
In Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI), March 2017.

[ pdf, bibtex, abstract ]


Natural Language Generation in the Context of Providing Indoor Route Instructions.
Andrea F. Daniele, Mohit Bansal, and Matthew R. Walter.
In Proceedings Robotics: Science and Systems Workshop on Model Learning for Human-Robot Communication, May 2016.

[ pdf, bibtex, abstract ]