Giovanni Franzini

Nonlinear H-infinity control of relative motion in space via the state-dependent Riccati equations

Relative guidance algorithms for space applications were identified by several space agencies as an enabling technology for future missions development. Whenever two or more space vehicles must coordinate their motion, or a terminal rendezvous has to be performed, a robust control of the relative motion occurring between the objects is necessary. Control must guarantee operation safety, and minimize fuel consumption, since refueling operations are currently too expensive. In this context, the paper proposes an extended linearization technique to design a nonlinear H-infinity controller for the relative motion. The developed controller is designed to minimize propellant consumption and to attenuate disturbances due to typical perturbations of low Earth orbits, such as atmospheric drag and J2 perturbation. Terminal rendezvous and formation control simulations were performed using data from realistic missions.

Swarm obstacle and collision avoidance using descriptor functions

The descriptor function framework is used as tool for the control management of a swarm of dynamic agents. In this framework, a provision is made for obstacle and collision avoidance, thus improving the potential of the methodology from previous results. Obstacle and collision avoidance terms are added to the overall mission performance index, and the resulting control law moves the agents along obstacle and collision free trajectories. The analytical derivation is validated via numerical simulations.

H-infinity controller design for spacecraft terminal rendezvous on elliptic orbits using differential game theory

The paper presents a H-infinity guidance law for spacecraft low-thrust terminal rendezvous on elliptic orbits. The dynamics of the rendezvous on elliptic orbits are governed by a set of linear time-varying equations, in literature known as linear equations of relative motion. Standard H-infinity controller design technique for linear systems cannot be adopted, since the system is time-varying. Therefore, the problem is formulated as a zero-sum two-person differential game following the minimax H-infinity design technique developed by Başar and Bernhard. The main result is a closed-form solution of the terminal rendezvous on elliptic orbits H-infinity control problem. In addition, we prove that the H-infinity norm of the closed-loop system is bounded.

Human-Machine Interface for Multi-agent Systems Management Using the Descriptor Function Framework

Human-machine interfaces for command and control of teams of autonomous agents is an enabling technology for the development of reliable multi-agent systems. Tools for proper modelling of these systems are sought in order to ease the creation of efficient interface that allow a single operator to control several agents, as well as monitor the execution state of the tasks the team is demanded to accomplish. If humans are present in the environment, the agents must sense their presence and collaborate with them toward the mission accomplishment. In this context, the descriptor function framework is a versatile tool that allows the human integration at two levels: the development of human-machine interfaces and the achievement of human-machine teaming. In this paper, we show how such results can be obtained and we propose a possible architecture for the framework implementation.

A Visual-Haptic Display for Human and Autonomous Systems Integration

This paper introduces a novel concept of visual-haptic display for situational awareness improvement for crowded and low altitude airspace situations. The visual augmentation display that constitutes of Virtual Fences delimiting no-fly zones, and a specific tri-dimensional highlight graphics that enhances visibility of other remotely piloted or autonomous agents, as well as conventional manned aircraft operating in the area is presented first. Then the Shared Control paradigm and the Haptic Force generation mechanism, based on a Proportional-Derivative-like controller applied to repulsive forces generated by the Virtual Fences and other UAVs are introduced and discussed. Simulations with 26 pilots were performed in a photo-realistic synthetic environment showing that the combined use of Visual-haptic feedback outperforms the Visual Display only in helping the pilot keeping a safe distance from no-fly zones and other vehicles.