N.B. This page is likely to be outdated, take a look at my publication list for more recent research activities.



Research interests:


  1. Set-Membership approach to autonomous navigation
  2. EKF-based SLAM with linear features
  3. Collective motion for multi-agent systems
  4. Mobile haptic interfaces
  5. Satellite localization for deep space missions
  6. Consensus protocols in presence of bounded errors
















Set-Membership approach to autonomous navigation

Localization and map building are crucial issues when a mobile robot must autonomously navigate in possibly unknown environments. Both problems can be cast as a state estimation problem for an uncertain dynamic system. Depending on the assumptions on the uncertainty affecting the dynamic model and the measurement equations, probabilistic or deterministic techniques can be adopted. When the process disturbance and the measurement noise are supposed to be unknown-but-bounded (UBB), localization and map-building can be tackled via set-membership state estimation techniques [6].

Localization

Given a map of the environment, the objective is to estimate the vehicle pose. In a set-membership framework, this amounts to compute the feasible pose set, i.e. the set containing all vehicle poses that are compatible with the robot motion mode, the collected measurements and the bounds on the errors [1]. Ad-hoc nonlinear estimators computing outer approximations of the actual feasible pose set shave been developed. The resulting estimates are sets with a fixed structure (e.g., boxes or parallelotopes) guaranteed to contain the unknown robot pose.

Simultaneous localization and map building (SLAM)

When a map environment is not available (e.g., in exploration tasks), the localization problem becomes harder. In this case, the robot must build a map while at the same time localizing within it. This is the so-called Simultaneous Localization And Map building problem (SLAM). By include  the landmark positions in the state vector to be estimated, the problem reduces to a state estimation problem. Under the assumption of bounded errors, set-membership estimators have been developed to compute the feasible state set, i.e. the set containing all vehicle poses and landmark locations that are compatible with the available information [2,4]. The resulting techniques feature increased efficiency through state decomposition and set approximation, provide guaranteed regions containing the unknown robot pose and landmark positions, and are suitable for dealing with the data association problem. The set-membership SLAM algorithm has been successfully extended to the multi-robot scenario [3].

Path planning

In the set-membership framework. a measure of the localization uncertainty is given by the size of the estimated sets. This information can be used in order to plan maximally informative paths, i.e. paths minimizing the overall position uncertainty along the path. This problem can be cast as an optimal control problem, by defining a suitable cost function accounting for the expected localization uncertainty [5,7]. The proposed solution is able to deal with sensor limitations and obstacle avoidance through artificial potentials.


References

  1. M. Di Marco, A. Garulli, A. Giannitrapani, A. Vicino. Set membership pose estimation of mobile robots based on angle measurements. In Proceedings of the 40th IEEE Conference on Decision and Control, Orlando (USA),  December 4-7, 2001.
  2. M. Di Marco, A. Garulli, A. Giannitrapani, A. Vicino. Dynamic robot localization and mapping using uncertainty sets. In Proceedings of the 15th World IFAC Congress, Barcelona (Spain), July 2002.
  3. M. Di Marco, A. Garulli, A. Giannitrapani, A. Vicino. Simultaneous localization and map building for a team of cooperating robots: a set membership approach. IEEE Transactions on Robotics and Automation, vol. 19, n. 2, pp. 238-249, April 2003 (Regular paper).
  4. M. Di Marco, A. Garulli, A. Giannitrapani, A. Vicino. A set theoretic approach to dynamic robot localization and mapping. Autonomous Robots, vol. 16, no. 1, pp. 23-47, January 2004 (Regular paper).
  5. N. Ceccarelli, M. Di Marco, A. Garulli, A. Giannitrapani. A set theoretic approach to path planning for mobile robots. In Proceedings of the 43rd IEEE Conference on Decision and Control, Atlantis (Bahamas), December 14-17, 2004.
  6. N. Ceccarelli, M. Di Marco, A. Garulli, A. Giannitrapani, A. Vicino. Set membership localization and map building for mobile robots. Current trends in nonlinear systems and control, L. Menini, L. Zaccarian and C.T. Abdallah, Eds., pp. 289-308, Birkäuser, 2006.
  7. N. Ceccarelli, M. Di Marco, A. Garulli, A. Giannitrapani, A. Vicino. Path planning with uncertainty: a set membership approach. International Journal of Adaptive Control and Signal Processing,  vol. 25, no. 3, pp.  273-287,  March 2011 (Special issue: Bounding methods for state and parameter estimation).

[Top]


EKF-based SLAM with linear features

A common representation of indoor environments is given in terms of linear features (lines and segments). In this case the simultaneous localization and map building problem (SLAM) consists in estimating the robot pose and the parameters of the lines composing the map. A SLAM algorithm is proposed which recursively updates the estimates of the unknown quantities  through a single Extended Kalman Filter (EKF) [1,2]. A mixed line/segment description of the environment is adopted in order to improve the data association algorithm. Linear features present in the environment are extracted from range and bearing measurements. A statistical characterization of the observation process is derived directly from the sensor specs.


References

  1. A. Garulli, A. Giannitrapani, A. Rossi, A. Vicino. Simultaneous localization and map building using linear features. In Proceedings of the 2nd European Conference on Mobile Robots, Ancona (Italy), September 7-10, 2005.
  2. A. Garulli, A. Giannitrapani, A. Rossi, A. Vicino. Mobile robot SLAM for line-based environment representation. In Proceedings of the 44th Conference on Decision and Control, pp. 2041-2046, Seville (Spain), December 12-15, 2005.
  3. D. Benedettelli, A. Garulli, A. Giannitrapani. Multi-Robot SLAM using M-Space feature representation. In Proceedings of the 49th IEEE Conference on Decision and Control,  pp. 3826-3831, Atlanta (USA), December 15-17, 2010.
  4. D. Benedettelli, A. Garulli, A. Giannitrapani. Cooperative SLAM using M-Space representation of linear features. Robotics and Autonomous Systems, vol. 60, no. 10, pp. 1267-1278, October 2012 (Regular paper).

[Top]


Collective motion for multi-agent systems

The problem considered is that of a team of nonholonomic vehicles whose objective is to achieve collective circular motion around a virtual reference beacon. A decentralized control law is proposed [1,4], which is shown to guarantee global asymptotic stability of the counterclockwise circular motion around a fixed beacon, in the single-vehicle case. Equilibrium configurations of the multi-vehicle system are discussed and sufficient conditions for local asymptotic stability are derived in terms of the control law design parameters. Sensory limitations are explicitly taken into account. Specifically, i) each agent can perceive only vehicles lying in a limited visibility region; ii) a vehicle cannot measure the orientation of another vehicle; iii) vehicles are indistinguishable. Guidelines for the choice of the control law parameters are provided, taking into account the trade-off between fast convergence to the equilibrium configuration and safe collision-free trajectories. Experimental validation of the developed control law [2,3,5-7] has shown the effectiveness of the proposed approach also in a real-world scenario featuring  a number of disturbance sources.


References

  1. N. Ceccarelli, M. Di Marco, A. Garulli, A. Giannitrapani. Collective circular motion of multi-vehicle systems with sensory limitations. In Proceedings of the 44th Conference on Decision and Control, pp. 740-745, Seville (Spain),  December 12-15, 2005.
  2. N. Ceccarelli, M. Di Marco, A. Garulli, A. Giannitrapani. Experimental analysis of collective circular motion for multi-vehicle systems. In Proceedings of the 8th International IFAC Symposium on Robot Control. Bologna (Italy), September 6-8, 2006.
  3. D. Benedettelli, N. Ceccarelli, A. Garulli, A. Giannitrapani. Experimental validation of a decentralized control law for multi-vehicle collective motion. In Proceedings of the 2007 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 4170-4175 San Diego (USA), October 29 - November 2, 2007.
  4. N. Ceccarelli, M. Di Marco, A. Garulli, A. Giannitrapani. Collective circular motion of multi-vehicle systemsAutomatica, vol. 44, no. 12, pp. 3025-3035, December 2008 (Regular paper).
  5. D. Benedettelli, M. Casini, A. Garulli, A. Giannitrapani, A. Vicino. A LEGO Mindstorms experimental setup for multi-agent systems. In Proceedings of the 3rd IEEE Multi-Conference on Systems and Control, pp. 1230-1235, St. Petersburg (Russia), July 8-10, 2009.
  6. M. Casini, A. Garulli, A. Giannitrapani, A. Vicino. A Matlab-based remote lab for multi-robot experiments. In Proceedings of the 8th IFAC Symposium on Advances in Control Eduction, Kumamoto (Japan), October 21-23, 2009.
  7. D. Benedettelli, N. Ceccarelli, A. Garulli, A. Giannitrapani. Experimental validation of collective circular motion for nonholonomic multi-vehicle systemsRobotics and Autonomous Systems, vol. 58, no. 8, pp. 1028-1036, August 2010 (Regular paper).
  8. F. Morbidi, A. Giannitrapani, D. Prattichizzo. Maintaining connectivity among multiple agents in cyclic pursuit: a geometric approach. In Proceedings of the 49th IEEE Conference on Decision and Control, pp. 7461-7466,  Atlanta (USA), December 15-17, 2010.
  9. M. Casini, A. Garulli, A. Giannitrapani, A. Vicino. A LEGO Mindstorms multi-robot setup in the Automatic Control Telelab. In Proceedings of the 18th IFAC World Congress, pp. 9812-9817, Milan (Italy), August 28 - September 2, 2011.
  10. M. Casini, A. Garulli, A. Giannitrapani, A. Vicino. A remote lab for multi-robot experiments with virtual obstacles. In Proceedings of the 9th IFAC Symposium on Advances in Control Education, pp. 354-359, Nizhny Novgorod (Russia), June 19-21, 2012.
[Top]


Mobile haptic interfaces


The increasing demand of realism for virtual reality applications calls for the design of advanced force-feedback devices allowing not only kinesthetic interaction with virtual objects but also locomotion and navigation inside virtual worlds. A possible solution is to resort to mobile haptic interfaces (MHI), built by combining standard force-feedback devices with mobile platforms. In [1-4] we investigated both theoretically and experimentally which factors may affect the transparency of this kind of devices, identifying in mobile robot dynamics a possible cause of loss of transparency. A method to analyze dynamic performance of an MHI and some basic guidelines to design controller in order to meet desired specifications are also provided.


References

  1. A. Formaglio, A. Giannitrapani, M. Franzini, D. Prattichizzo, F. Barbagli. Current issues in haptic rendering using mobile haptic interfaces. In Proceedings of the 1st joint EuroHaptics Conference and Symposium on Haptic Interfaces, Pisa (Italy), March 18-20, 2005.
  2. A. Formaglio, A. Giannitrapani, M. Franzini, D. Prattichizzo, F. Barbagli. Performance of Mobile Haptic Interfaces. In Proceedings of the 44th Conference on Decision and Control,  pp. 8343-8348, Seville (Spain),  December 12-15, 2005.
  3. F. Barbagli, A. Formaglio, M. Franzini, A. Giannitrapani, D. Prattichizzo. An experimental study of the limitations of Mobile Haptic Interfaces. Springer Tracts in Advanced Robotics - Experimental Robotics IX, Springer Verlag - Berlin, 2006.
  4. A. Formaglio, D. Prattichizzo, F. Barbagli, A. Giannitrapani. Dynamic performance of Mobile Haptic Interfaces. IEEE Transactions on Robotics, vol. 24, no. 3, pp. 559-575, June 2008 (Regular paper).

[Top]


Satellite localization for deep space missions

When GPS signals are not available (e.g., high-Earth orbits, Earth-to-Moon transfers) estimation of satellite position is a fundamental requirement for autonomous navigation. This is especially true when dealing with spacecraft propelled by electric propulsion systems featuring a continuous thrusting strategy which requires an accurate knowledge of the spacecraft position and attitude. In [1,2] the problem of spacecraft self-localization is addressed using angular measurements. A dynamic model of the spacecraft is derived, which takes into account several perturbing effects such as Earth and Moon gravitational field asymmetry and errors associated with the Moon ephemerides . It is assumed that the spacecraft is equipped with line of sight sensors providing measurements of elevation and azimuth of Moon and Earth with respect to the spacecraft reference system Range measurements, which are often difficult to obtain or not sufficiently reliable, are not required. Position and velocity of the spacecraft are estimated by employing both the classical Extended Kalman Filter (EKF) and the recently developed Unscented Kalman Filter (UKF). The performance of the filters has been tested on simulated data concerning two different missions (Earth-to-Moon transfer and GEO orbit raising), showing that the proposed algorithms provide reliable estimates, whose accuracy is sufficient for autonomous navigation in the considered class of missions.

References

  1. N. Ceccarelli, A. Garulli, A. Giannitrapani, M. Leomanni, F. Scortecci. Spacecraft localization via angle measurements for autonomous navigation in deep space. In Proceedings of the 17th IFAC Symposium on Automatic Control in Aerospace. Toulouse (France), June 25-29, 2007.
  2. A. Giannitrapani, N. Ceccarelli, F. Scortecci, A. Garulli. Comparison of EKF and UKF for spacecraft localization via angle measurements. IEEE Transactions on Aerospace and Electronic Systems, vol. 47, no. 1, pp. 75-84, January 2011
  3. A. Garulli, A. Giannitrapani, M. Leomanni, F. Scortecci. Autonomous low Earth orbit station-keeping with electric propulsion. Journal of Guidance, Control, and Dynamics, vol. 34, no. 6, pp. 1683-1693, November 2011 (Regular paper).
  4. A. Garulli, A. Giannitrapani, M. Leomanni, F. Scortecci. Autonomous station keeping for LEO missions with a hybrid continuous/impulsive electric propulsion system. In Proceedings of the 32nd International Electric Propulsion Conference, Kurhaus (Germany), September 11-15, 2011.

[Top]



Consensus protocols in presence of bounded errors

In [1,2] we analyze two classes of consensus algorithms in a set-theoretic framework. Under the assumption of unknown but bounded measurement errors, the feasible state set (i.e., the set of all states compatible with the bounds on the noise) is explicitly derived. The evolution of the feasible state set is used to evaluate the agreement of the team. Linear consensus protocols adopting both constant weights and vanishing weights are considered, in the case of undirected and stationary communication graph. It is shown that for both types of protocols, asymptotic consensus cannot be guaranteed with respect to all possible noise realizations, and bounds on the asymptotic difference of the agents' states are explicitly derived, as a function of the bounds on the measurement errors and the weight matrix.


References

  1. A. Garulli, A. Giannitrapani. A set-membership approach to consensus problems with bounded measurement errorsProceedings of the 47th IEEE Conference on Decision and Control, pp. 2276-2281, Cancun (Mexico), December 9-11, 2008.
  2. A. Garulli, A. Giannitrapani. Analysis of consensus protocols with bounded measurement errors. Systems and Control Letters, vol. 60, no. 1, pp. 44-52, January 2011.

[Top]

[Home] - [Research] - [Publications] - [Videos] - [Teaching] - [CV]

Valid HTML 4.01 Strict Valid CSS!