Michele Furci

Global trajectory tracking for underactuated VTOL aerial vehicles using a cascade control paradigm

This work proposes a feedback control strategy capable of controlling the dynamics of an under-actuated Vertical Take-Off and Landing (VTOL) aerial vehicle to track a desired trajectory globally with respect to the initial conditions. The novelty of the proposed design is the idea of considering a cascade control paradigm in which the attitude dynamics, which are governed by means of a hybrid controller capable of overcoming the well-known topological constraints, and the position dynamics of the vehicle play respectively the role of the inner and of the outer loop. The stability properties of the proposed controller are then derived by analyzing the interconnection between a hybrid system, namely the closed-loop attitude dynamics, and a continuous time system, given by the closed-loop position dynamics. The proposed algorithms are then demonstrated by means of simulations obtained considering a miniature quadrotor prototype.

Robust Global Trajectory Tracking for Underactuated VTOL Aerial Vehicles Using Inner-Outer Loop Control Paradigms

This paper proposes a feedback control strategy to let the dynamics of a class of under-actuated vertical take-off and landing (VTOL) aerial vehicle tracking a desired position and attitude trajectory globally with respect to the initial conditions. The proposed feedback controller is derived following an inner-outer loop control paradigm, namely by considering the attitude, which is governed by means of a hybrid controller so as to overcome the well-known topological constraints, serving as a virtual input to stabilize the aircraft position. Two different approaches, the first one obtained by assuming perfect knowledge of the vehicle dynamics and the second one obtained by considering uncertainties and exogenous disturbances, are proposed and compared by analyzing the interconnection between the hybrid attitude and the continuous-time position closed-loop subsystems. The effectiveness of the obtained results is demonstrated by means of simulations and experiments using a miniature quadrotor prototype.

A control architecture for multiple drones operated via multimodal interaction in search & rescue mission

An architecture suitable for the control of multiple unmanned aerial vehicles deployed in Search & Rescue missions is presented in this paper. In the proposed system, a single colocated human operator is able to coordinate the actions of a set of robots in order to retrieve relevant information of the environment. This work is framed in the context of the SHERPA project whose goal is to develop a mixed ground and aerial robotic platform to support search and rescue activities in alpine scenario. Differently from typical human-drone interaction settings, here the operator is not fully dedicated to the drones, but involved in search and rescue tasks, hence only able to provide sparse and incomplete instructions to the robots. In this work, the domain, the interaction framework and the executive system for the autonomous action execution are discussed. The overall system has been tested in a real world mission with two drones equipped with on-board cameras.

An energy-efficient parallel algorithm for real-time near-optimal UAV path planning

We propose a shortest trajectory planning algorithm implementation for Unmanned Aerial Vehicles (UAVs) on an embedded GPU. Our goal is the development of a fast, energy-efficient global planner for multi-rotor UAVs supporting human operator during rescue missions. The work is based on OpenCL parallel non-deterministic version of the Dijkstra algorithm to solve the Single Source Shortest Path (SSSP). Our planner is suitable for real-time path re-computation in dynamically varying environments of up to 200 m2. Results demonstrate the efficacy of the approach, showing speedups of up to 74x, saving up to ~ 98% of energy versus the sequential benchmark, while reaching near-optimal path selection, keeping the average path cost error smaller than 1.2%.

An open-source architecture for control and coordination of a swarm of micro-quadrotors

This work presents the design of an open-source control architecture specifically tailored to the rapid development and testing of control and coordination algorithms on micro quadrotors. The proposed design extends an existing open-source and open-hardware quadrotor project, i.e., the Crazyflie nano quadrotor, by adding the guidance, navigation and control layers required to accomplish autonomous flight. The control layer, in particular, is based on a cascade control algorithm able to globally stabilize the position and the attitude of the vehicle. The global property guaranteed by the algorithm allows the quadrotors to perform aggressive maneuvers. The guidance layer is designed to coordinate simultaneously multiple vehicles by generating suitable reference trajectories. Experiments demonstrate the effectiveness of the proposed design.

A supervisory control strategy for robot-assisted search and rescue in hostile environments

This work proposes a reactive supervisory control for a team of heterogeneous robots capable of autonomous navigation and cooperation in unstructured and hostile environment. The supervisor is built using classical theories of discrete event systems (DES), but it differs for its reactive capabilities. This allows to have an online evaluation and optimization for dynamic events and environments. The supervisor (high level control) interacts directly with low level control, taking into account dynamic and kinematic of the agents. The specific robotic platform includes both aerial (UAV) and ground robots but the results and applications are absolutely general.

A smartphone based quadrotor: Attitude and position estimation

This paper describes the results obtained in the implementation of a solution implementing real-time precise and accurate attitude and position estimation algorithms designed for control applications based on an Android smartphone. The topic of the paper is in particular the application of this solution to dynamics sensing in a quadrotor Unmanned Aerial Vehicle (UAV). The attitude estimation is based on an Explicit Complementary Filter (ECF). The position estimation is implemented through a complementary filter using acceleration readings provided by the smartphone sensors and the position derived by a vision based motion capture system. The parameters of the algorithms are tuned through the minimisation of the estimation error for sets of sample acquisitions. Moreover, these settings are chosen in order to minimise the issues related with the fact that Android is not a Real Time Operating System (RTOS). Finally, it was verified that the chosen solution provides performances adequate to the quadrotor stabilisation. The results of a set of experimental tests are used to show the performances of the proposed solution. For the attitude estimation algorithm, a comparison is moreover provided in order to highlight the difference with standard Android orientation estimation algorithm.

A combined planning and control strategy for mobile robots navigation in populated environments

In this work we present a combined planning and control strategy to let agile autonomous robots navigating in real-world populated environments by means of range limited sensors. In order to reduce the computational cost of obtaining a dynamically-feasible path, the proposed solution relies on a simple way-points generation subject to geometrical constraints. Conditions ensuring feasibility of the obtained trajectories also in the presence of possible exogenous disturbances such as wind are then derived by taking into account the interconnection between the vehicle dynamics, the low-level stabilizing controller and the trajectory planner. As an application, the navigation of a quadrotor and of a differential wheel robot in an unknown forest environment is proposed. In particular, the proposed framework is employed to let the vehicle moving at the highest possible speed that guarantees no collision with obstacles.

A robust control strategy for mobile robots navigation in dynamic environments

This work introduces a novel control strategy to allow a class of mobile robots to robustly navigate in dynamics and potentially cluttered environments. The proposed approach combines a high-level motion planner and a low-level stabilizing feedback control law designed considering the nonlinear dynamic model of the vehicle. Taking advantage of a symbolic description of the vehicle dynamics and of the environment, the reference trajectories are sequences of elementary primitives which are obtained with a reduced computational cost. However, the resulting references may fail to be functionally controllable for the actual dynamical model of the vehicle. Accordingly, to obtain a desired tracking error, sufficient conditions are then derived by investigating the interconnection between the discrete time planner and the continuous time closed-loop nonlinear system. The effectiveness of the obtained results is demonstrated by considering, as application, a ground robot navigating in a cluttered environment.

Modeling and Control of a Class of Multi-Propeller Aerial Vehicles

Abstract This work focuses on the modeling and the control design for a class of multi-propeller aerial vehicles. The idea is to consider the aircraft as a modular system obtained by interconnecting a number of actuator modules, each one given by a fixed-pitch propeller driven by a motor, and of payload modules, representing the equipment and all the other passive components installed onboard. A control strategy based on control allocation is then proposed so as to govern a generic multi-propeller aerial vehicle by means of a unique globally stabilizing controller. The effectiveness of the proposed results are then demonstrated by considering the problem of controlling the dynamical model of quad-rotor aerial vehicle.