Alessandro Renzaglia

Short Bio:

I received my BSc and MSc in Physics at the University of Rome, Italy, "La Sapienza" in 2005 and 2007 respectively. In 2012 I obtained my PhD in "Applied Mathematics and Computer Science" from the University of Grenoble, France. During this period I have been part of the e-Motion group at INRIA Rhone-Alpes. The title of my thesis is: "Adaptive stochastic optimization for cooperative coverage with a swarm of Micro Aerial Vehicles".
From September 2012 to July 2014 I have been Postdoc Associate in the Robotic Sensor Networks Lab, at the Computer Science and Engineering Department, University of Minnesota. After this experience, I have been working as Postdoc Researcher in the Robotics and InterectionS (RIS) group at LAAS-CNRS, Toulouse, France.

Main Research

  • Cooperative exploration in currents fields: This work, developed in the framework of the SkyScanner project (, focuses on the optimal trajectory planning for cooperative exploration and data gathering by using a team of fixed-wings UAVs. The goal of this project is to exploit a swarm of UAVs to collect data (winds, pressure, humidity, temperature, etc.) within low-altitude convective clouds. The resulting study of their behavior and evolution is aimed to improve meteorological models and, consequently, small and large scale weather forecasts. The central research topic I am focusing on is the achievement of complex cooperative tasks in highly dynamic environments and in presence of a current field, where currents simultaneously represent the objective of study and a strong constraint for the motion.

  • Limited energy optimal search: This research addresses a fundamental search problem in which a searcher subject to time and energy constraints tries to find a mobile target. The target's motion is modeled as a random walk on a discrete set of points. We initially consider 1D bounded environments and study the process taking into account two different detection models. The objective is to optimize the motion, seen as a sequence of actions, such that the probability of capturing the random-walking target is maximized. We assume that move and stay actions correspond to different energy costs. The problem has been approached analytically as well as modeled as a Partially Observable Markov Decision Process (POMDP) and solved using a reduced state-space representation of the belief.
    This work is part of a collaborative project funded by the National Science Foundation:
    A Robotic Network for Locating and Removing Invasive Carp from Inland Lakes

  • Stochastic optimization for optimal coverage: This research has been carried out in the framework of the European project sFly ( The objective of this project was to develop several small and safe helicopters which can fly autonomously in city-like environments and which can be used to assist humans in rescue and monitoring tasks. The main motivations were not only to achieve tasks impossible for a human team, but also to be able to substitute the human intervention in very dangerous scenarios. This means that the helicopters must be able to operate in complex environments where GPS signals are often shadowed, cooperatively and in a complete autonomous way. This involves a number of challenges on all levels of helicopter design, perception, actuation, control, navigation and power supply that have yet to be solved.
    My research focused on the problem of deploying a swarm of Micro Air Vehicles (MAV) to perform surveillance coverage mission over an unknown terrain of arbitrary morphology. Since the considered terrain's morphology is unknown and it can be complex and non-convex, standard optimization algorithms are not applicable to this particular problem. For this reason, a new approach based on the Cognitive-based Adaptive Optimization (CAO) algorithm has been proposed. A fundamental property of this approach is that it shares the same convergence characteristics as those of constrained gradient-descent algorithms, which require perfect knowledge of the terrain's morphology. In addition, we also proposed a different formulation of the problem in order to obtain a distributed solution, which allows us to overcome the drawbacks of a centralized approach and to consider also limited communication capabilities. The proposed method was finally implemented in a real swarm of MAVs to carry out surveillance coverage in an outdoor complex area near Zürich.

Professional Activities

Member of Program Committee:

  • DARS 2014 - 2016 - International Symposium on Distributed Autonomous Robotic System
  • RSS 2015 -  Robotics: Science and Systems Conference 
  • WiSARN 2016 - International Workshop on Wireless Sensor, Actuator and Robot Networks
  • RoboSense 2014 - The International Workshop on Cooperative Robots and Sensor Networks