Inventaire
Site en français
GARONE Emanuele



Units

Department of control engineering and system analysis

Person in charge of the Unit : Oui

The research activities of the SAAS group are essentially oriented in two directions:
- Supervision and fault diagnosis of technical systems based on mathematical models
- Advanced control of nonlinear dynamic system subject to constraints, with emphasis on reference governor methods
 
The developed methods are applied to a wide range of areas including:
- Electric energy production, distribution and storage
- Mechatronics (UAVs and robotics)
- Industrial processes

Projetcs

LI-ION FAST & SAFE: SAFE AND FAST CHARGE OF LI-ION BATTERY

Fast charging is considered a key aspect for increasing the usage of Li-ion batteries, especially within applications where it is fundamental to reduce the downtown of the equipment due to charge (Electric Vehicles, Drones, Power and Gardening Tools, etc.). Most of the current commercial battery management systems (BMSs) available in the market make use of fast-charging options based on logic-based protocols without awareness of the battery dynamics that can dramatically reduce the lifespan of the battery in the long term.

After more than 10 years of experience in the field, this spin-off project, funded by Innoviris and led by Alejandro, aims at improving the current commercially-available fast charging solutions with scientific findings in the area of model-based constrained control strategies. Thus, we develop, test, and validate low-complexity, optimization-free, and close-loop estimation and charging algorithms that reduce the charging time while diminishing the likelihood of an accelerated electrochemical and thermal degradation of Li-ion batteries. Our offer is built upon 4 pillars:

1. An accurate but simple characterization of the electrochemical and thermal dynamics of the battery (either a cell or a pack);

2. An estimation layer able to track the dynamics of the battery;

3. Charging-control algorithms that charger the battery as fast as possible according to the battery dynamics and considering the environmental conditions;

4. Extensive validations campaigns, using commercial batteries (Sony, Samsung, Panasonic, LG, etc):

Such pillars are operative supported by more than 60 testing channels and covered by the patent WO2022189454 A1 (Fast-charge of a Li-ion Battery based on a state feedback on a reduced order electrochemical model).

Robotic Bricklayer: a multi-robot system for sand-lime blocks masonry

In this project we will design and develop an innovative concept for the automatic bricklaying of sand-lime bricks. The robotization of this activity poses a series of fundamental methodological challenges. The first challenge is that the weight of the blocks and the typical distance at which the blocks must be placed makes the classical robotic assumption of “rigid-manipulator” unreasonable. Accordingly we will study a “soft-robotic” mechanical design, together with a suitable sensing equipment and control law.  The second challenge in this project is the design of safe and effective control solutions to perform the construction task. The correct placement and alignment of each block will be done through a combination of position and force control. A prototype will be developed for laboratory tests.

Knee Advanced Rehabilitation Device

Nowadays, knee surgery is a relatively common procedure used to treat cartilage defects and/or soft tissue lesions, osteoarthritis with total or unicondylar knee arthroplasty, and cruciate ligaments (ACL/PCL) injuries.
After surgery, each patient undergoes a long period of rehabilitation (typically from 6 weeks to 6 months) consisting of long sessions of physiotherapy and medical training therapy carried out by qualified personnel. This procedure is long and expensive, and may cause work-related pathologies to physiotherapists because of the significant workload it implies. These considerations highlight the high potential benefits that robotic solutions could bring in this field. As a matter of fact, if technology could provide an effective tool to assist the physiotherapist, the rehabilitation time and cost could be reduced, with important benefits for both the patients and the operators.
Although some knee rehabilitation devices are already available, they have proved to be not very effective and, in some situations, could even harm the patient. The main reason of this is related to the over simplistic and non-physiological assumptions used for the device design: the first one is to consider the knee as a 1 Degree of Freedom joint enabling only flexion-extension. Second, these devices constrain the knee to “blindly” perform movements with very few adaptations to the patients’ anatomy and muscular reaction. 
This project aims to solve this issue by investigating it from a biomechanical and robotic perspective, considering human physiology patient-specific knee kinematics and kinetics as the starting point of the design.

CONSTRAINED CONTROL AND ITS MARINE APPLICATIONS

Frank is currently exploring the realm of control allocation for marine vessels. He is working on optimization-free approaches and also new convex formulations to solve the allocation problem in an efficient way.

BRICKIEBOTS 2.0

This project aims to continue the development of BrickieBots, a multi-agent robotic system that will aid in sand-lime masonry. This robotic system, comprised of an industrial manipulator and a crane, is designed to be able to finely place the blocks autonomously. Doing so increases safety in construction sites by reducing the people that need to be up on scaffoldings and around large, free-hanging blocks.

Navigation and trajectory control of a robotic hummingbird

This project is a part of the COLIBRI project, which aims at building a robot mimicking the flight of a natural hummingbird. The objective of the research is to integrate appropriate sensors (such as vision-based sensors) to the COLIBRI control board in order to, in one hand, increase the accuracy of the hovering flight and, on the other hand, utilize it as a navigation system to perform desired trajectories. Then a control algorithm will be developed to process the real-time signal obtained by the integrated sensor in order to steer the robot through the desired trajectory. The performance of the proposed control algorithm including sensor integration will be evaluated by flight experiments.

AMATS – ADVANCED MULTILAYER ADAPTIVE THIN SHELLS

The project is a collaboration between ULB, MateriaNova and the European Space Agency to manufacture and characterize a thermally stable piezoelectrically actuated flexible polymer reflector. The active material is envisaged to be used in deployable space telescopes for nanosatellites. The built-in active control will allow the reflector to correct disturbances from manufacturing, thermal loads, and stowage-induced creep. The result is the creation of small satellites carrying telescopes with apertures larger than the size of the satellite. Future applications include Earth Observation (EO), Laser Communication (FSOC) and Astronomy.

SELF-ORGANIZED NERVOUS SYSTEMS FOR ROBOT SWARMS

The project aimed to enhance the coordination and performance of heterogeneous robot swarms through a self-organized hierarchical control architecture. Key accomplishments include:
1. Developed the S-Drone, a novel, versatile quadrotor platform for swarm robotics research with onboard processing and sensing capabilities.
2. Designed a hierarchical control architecture enabling dynamic switching between distributed and centralized control, based on task requirements.
3. Implemented decentralized control algorithms, consensus-based control, and leader-follower approaches for efficient swarm coordination.
4. Demonstrated improved swarm performance through simulations and real-world experiments, showcasing enhanced task completion time, robustness, and adaptability.
The project’s findings advance swarm robotics, providing insights into self-organized hierarchical control for multi-robot systems. Practical applications include military and civilian domains

DESIGN OF ROBOT SWARMS WITH GUARANTEED BEHAVIOURAL PROPERTIES

My research focuses on the design of robot swarms with guaranteed behavioural properties. The goal is to conceive a reliable method to design robust robot swarms that could be used in real-world environments.

Testing and preemptive quarantine for the control of epidemics

This project aims at defining smart policies for the selection of the individuals that must be tested during an epidemics with the goal of maximizing the effectiveness of selective quarantine measures. The starting point of this project is to model epidemics using stochastic cellular automata and to devise ways to estimate the probability that each individual is infected. This probability estimation can be used as a guideline to prescribe the tests. This project aims at tackling the most relevant practical (tuning and validation of the model) and theoretical (estimation of the probability and design of the policies) aspects to make this idea a viable tool to support the decisions during an epidemics.

PESTFINDER: MODEL-BASED ESTIMATION AND CONTROL OF AGRICULTURAL INFESTATIONS THROUGH ABIOTIC CHANGES

PestFinder is a project funded by the European Commission in the framework of the Marie Skłodowska Curie Postdoctoral Fellowship actions 2022 (grant n. 101102281). The goal of PestFinder is to contribute to the ongoing “precision agriculture” revolution by developing model-based control strategies for pest control. We firstly aim to develop models and methodologies that can guide in an optimal way the agronomical decisions that impact pest control. To pursue this aim, we will develop and validate models able to capture both the spatial diffusion and the temporal evolution of insects in a plantation. The focus will be then shift on how to use these models for pest monitoring, and on the interplay between quality of the estimation and data collection, to develop optimal data collection policies. In the last part of the project, we will use the developed results to develop model-based optimal pest control policies.