My name is Andrea Del Prete. I am an associate professor focusing on the use of optimization algorithms for control, planning and estimation of legged robots. In this website you can find all my publications with links to open-access PDF files, videos and code. Each publication has a dedicated page where visitors can leave comments.
Since 2022 I have been an associate professor in the Industrial Engineering Department of the University of Trento (Italy), where I am teaching robotics and computer programming in C++. From 2019 to 2021 I had been a tenure-track assistant professor (RTD-B) in the same department.
In 2018 I had been a research scientist in the Movement Generation and Control group at the Max-Planck Institute for Intelligent Systems (Tübingen, Germany), under the lead of Ludovic Righetti.
From 2014 to 2017 I had been an associated researcher in the Gepetto team (LAAS-CNRS, Toulouse), where I have been working with the humanoid robot HRP-2. My main collaborations were with Nicolas Mansard, Olivier Stasse, Steve Tonneau and Justin Carpentier.
Before going to LAAS I had spent four years (3 of PhD + 1 of post-doc) at the Italian Institute of Technology (IIT, Genova, Italy), where I had been working with Lorenzo Natale and Francesco Nori on the iCub humanoid robot.
Since 2020 I am serving as an Associate Editor for IEEE Robotics and Automation Letters.
Ahmad Gazar (co-supervision with Ludovic Righetti @ MPI, Tuebingen, Germany): Stochastic and Robust MPC for legged robots. webpage
Gianluigi Grandesso (co-supervision with Patrick Wensing @ University of Notre Dame, USA): Co-Design with Reinforcement Learning. webpage
Francesco Roscia (co-supervision with Michele Focchi @ IIT, Genova, Italy)
Gianni Lunardi: Model predictive control for legged locomotion.
PhD from the Cognitive Humanoids laboratory of the 'Robotics Brain and Cognitive Sciences' department, 2013
IIT and University of Genoa
MEng in Computer Engineering, 2009
2nd faculty of the University of Bologna (Italy)
The project Memmo aims to solve the problem of generating complex movements for arbitrary robots with arms and legs interacting in a dynamic environment by 1) relying on massive off-line caching of pre-computed optimal motions that are 2) recovered and adapted online to new situations with real-time tractable model predictive control and where 3) all available sensor modalities are exploited for feedback control going beyond the mere state of the robot for more robust behaviors.
The goal of this project is to achieve reliable locomotion behaviors in semi-structured environments with the humanoid robots HRP-2 and Pyrene. This project relies on model-based feedback control techniques, such as inverse-dynamics and model predictive control.
Nowadays legged robots are capable of performing locomotion and manipulation in semi-structured environments, but with a low level of reliability, which makes their application in disaster-recovery scenarios difficult, if not impossible. However, if we look at the results that researchers in robotics and animation have achieved in simulation, we can see that simulated robots/avatars can easily and reliably perform dynamic movements such as walking, running, jumping, kicking. What is preventing real robots from showing similar performance?
A 60-hour course for master students. This course focuses on control of robotic systems, with special attention to methods based on optimization techniques. After reviewing the basic principles of robot modeling and numerical optimization, students will learn different control techniques, from the simplest and most well-known, to the most recent and advanced. Methods will be first studied in theory, and then implemented in simulation (with the Python language) to gain practical experience. The lab sessions will focus on industrial manipulators and legged robots, but most of the studied methods could also be applied to flying and wheeled robots. After completing the course, students will be able to: - understand the working principles of several control algorithms for robotic systems - choose the appropriate approach(es) to control a specific system for a given target application - implement, tune, and test control algorithms with the Python language
Students should have consolidated knowledge in
Here you can find videos and slides of all the lectures of two academic years (19 /20 and 20 /21, even though in 19 /20 the first lectures have not been recorded). You can also find a virtual machine (password: iamarobot) containing the software needed for the lab sessions.
A 12-hour course for PhD students about reactive control (TSID) and trajectory optimization for humanoid robots. Here you can find the videos of all the lectures (except the first one, which wasn’t recorded). Here you can find the slides:
A 3-hour class about TSID, covering both theory and implementation. Here you can find videos and slides:
I have also given a class on robust TSID:
Full-day workshop at ICRA 2017, Singapore.
Full-day workshop at ICRA 2016, Stockholm (Sweden).
Full-day workshop at Humanoids 2013, Atlanta (Georgia, USA).