The Lab focuses on the design of optimization-based algorithms for robot control and coordination in human-inhabited environments. Our goal is to ensure a safe, reliable, secure, and private coordination of the robots among humans.
The main methods involved in our research are nonlinear model predictive control (MPC) and decomposition methods (e.g., ADMM and AMA). We also investigate control strategies using deep reinforcement learning and game theory. Finally, we are exploring how to integrate privacy-preserving algorithms in the context of robot control and planning.
The main application domains of the Lab’s research are automotive, maritime transportation, manipulators, and UAVs.
Optimization
Model Predictive Control
Autonomous Vehicles
Cybersecurity
Reinforcement Learning
Game Theory
Projects
HARMONIA
Seamless Mobile-Robot Coordination In the Real World
Mobile robots, such as autonomous cars, vessels, and drones, can significantly improve our quality of life. These robots are equipped with communication units that allow information sharing and coordination near humans, abilities that are fundamental to improve, for example, traffic efficiency and even save human lives. However, the occurrence of faults and attacks can severely disrupt the coordination with negative consequences on humans’ safety and security. Furthermore, to coordinate, the robots share private users’ information, such as habits, routes, and destinations, which may be inadvertently exposed to prying eyes. Hence, there is an urgent need to address safety, security, and privacy concerns for our society to fully benefit from multi-robot systems.
This project will devise a unified coordination framework for mobile robots to seamlessly perform tasks near humans, providing strong safety, but also security and privacy guarantees.
More Information
This project received fundings from the Dutch Science Foundation NWO-TTW within the Veni project HARMONIA (nr. 18165).
Contact: Dr. L. Ferranti
REMARO
Reliable AI for Marine Robotics
REMARO attacks one of the most pressing problems of modern computing, the safety of AI, in the well defined context of submarine robotics. The REMARO research fellows will develop the first ever submarine robotics AI methods with quantified reliability, correctness specifications, models, tests, and analysis & verification methods. REMARO rests on two founding principles: (i) The submarine robot autonomy requires a comprehensive hybrid deliberative architecture, a robotic brain. (ii) Safety and reliability must be co-designed simultaneously with cognition, not separately, as an afterthought. These principles are used to construct the training program (to train ESRs to deliver required scientific breakthroughs) and the expert consortium (to supervise the ESRs, run secondments and courses).
More Information
REMARO has received funding from the European Union’s EU Framework Programme for Research and Innovation Horizon 2020 under Grant Agreement No 956200.
OpenDR
Open Deep Learning Toolkit for Robotics
The aim of OpenDR is to develop a modular, open and non-proprietary deep learning toolkit for robotics. We will provide a set of software functions, packages and utilities to help roboticists develop and test robotic applications that incorporate deep learning. OpenDR will enable linking robotics applications to software libraries such as tensorflow and the ROS operating environment. We focus on the AI and cognition core technology in order to give robotic systems the ability to interact with people and environments by means of deep-learning methods for active perception, cognition and decisions making. OpenDR will enlarge the range of robotics applications making use of deep learning, which will be demonstrated in the applications areas of healthcare, agri-food and agile production.
More Information
The project is funded by the EU Horizon 2020 program, call H2020-ICT-2018-2020 (Information and Communication Technologies), 2019 – 2023.
SCoop
Safe Cooperation of Autonomous Vehicles in Mixed Traffic
Autonomous vehicles (such as cars and vessels) will be widespread in our daily lives, aiming at reducing pollution while improving traffic efficiency and safety. The ability of these vehicles to cooperate in planning trajectories is one of the main strengths of this technology. The presence of human-operated vehicles and the occurrence of sensor/actuator faults, however, complicate the vehicle cooperation. Failing to handle these mixed-traffic uncertainties and faults in the motion planning strategy can inevitably compromise the cooperation. The goal of this project (SCoop) is to design a cooperation framework to allow autonomous vehicles to safely navigate in the presence of human-operated vehicles and faults. To design a novel safe cooperation framework, the project will rely on tools for uncertainty estimation/fault diagnosis and distributed motion planning. Experiments on real autonomous surface vessels (ASVs) will demonstrate the effectiveness of the proposed design. SCoop is a Cohesion project between the Cognitive Robotics Department and the Maritime and Transport Technology Department.
More Information
Coordinator: TU Delft.
Contact: Dr. L. Ferranti and Dr. V. Reppa
SafeVRU
Safe Interaction of Automated Vehicles with Vulnerable Road Users
SafeVRU addresses the interaction of highly automated vehicles with vulnerable road users (VRUs), such as pedestrians and cyclists, in the context of future urban mobility. Pursue an integrated approach, covering the spectrum of VRU sensing, cooperative localization, behavior modeling & intent recognition, and vehicle control.
More Information
Coordinator: TU Delft.
User Group: Province of Gelderland, TNO, NXP, 2GetThere, SWOV, RDW.
Contact: Prof. Dr. D. M. Gavrila
Formation control for Waterborne Structures
The project investigates the use of distributed optimization techniques for a multi-robot coordination problem, that is, the navigation of autonomous vessels at a canal intersection.
More Information
The project was a collaboration within the faculty of Mechanical, Maritime, and Material Engineering at TU Delft.
Contact: Dr. J. Alonso-Mora, Dr. T. Keviczky, and Prof. R. R. Negenborn.
RECONFIGURE
Reconfiguration of Control in Flight for Integral Global Upset Recovery
The main goal of RECONFIGURE is to investigate and develop aircraft guidance and control (G&C) technologies that facilitate the automated handling of off-nominal/abnormal events and optimize the aircraft status and flight. The automatism of the G&C will help alleviate the pilots’ task and optimize performance by automatically reconfiguring the aircraft to its optimal flight condition. This automatism and optimization must be performed while maintaining the current aircraft safety levels.