Search

Numerous applications require aerial surveillance. Civilian applications include monitoring forest fires, oil fields and pipelines, and tracking wildlife. Applications to homeland security include border patrol and monitoring the perimeter of nuclear power plants. Military applications are numerous. The current approach to these applications is to use a single manned vehicle for surveillance. However, manned vehicles are typically large and expensive. In addition, hazardous environments and operator fatigue can potentially threaten the life of the pilot. Therefore, there is a critical need for automating aerial surveillance using UAVs. This chapter describes a cooperative control strategy for aerial surveillance that has been successfully flight tested on small (48 inch wingspan) UAVs. The cooperative control strategy follows the design methodology described in Section 8.3. [ABSTRACT FROM AUTHOR]

This chapter applies the design scheme in Section 8.4.1 to deep space spacecraft formation flying. We propose a decentralized formation control scheme that is built on the combined strength of consensus algorithms and the virtual structure approach. By following a decentralized coordination architecture via the virtual structure approach, we introduce decentralized formation control strategies, which are appropriate when a large number of spacecraft are involved and/or stringent interspacecraft communication limitations are applied. In our decentralized scheme, each spacecraft in the formation instantiates a local copy of the state of the virtual structure, which is the coordination variable using the virtual structure framework. The local instantiation of the coordination variable in each spacecraft is then driven into consensus by communication with its neighbors following a bidirectional ring topology. The effectiveness of the proposed control strategies is demonstrated through simulation results. [ABSTRACT FROM AUTHOR]

The objective of this chapter is to explore the feasibility of using multiple low-altitude, short endurance (LASE) UAVs to monitor and track cooperatively the propagation of large forest fires. A real-time algorithm is described for tracking the perimeters of fires with an onboard infrared sensor. Using this algorithm, we apply the design scheme in Section 8.4.1 to develop a distributed multi-UAV approach to monitoring the perimeter of a fire. We propose a distributed load balancing algorithm that extends the discrete-time consensus algorithm in Chapter 2. The UAVs are assumed to have limited communication and sensing range. The effectiveness of the approach is demonstrated in simulation using a six degree-of-freedom dynamic model for the UAV and a numerical propagation model for the forest fire. Salient features of the approach include the ability to monitor a changing fire perimeter, the ability to add and remove UAVs systematically from the team, and the ability to supply time-critical information to firefighters. [ABSTRACT FROM AUTHOR]

In this chapter, we apply the design scheme in Section 8.4.2 to formation maneuvers for a group of wheeled mobile robots. Complex formation maneuvers are decomposed into a sequence of maneuvers between formation patterns. We present three decentralized behavior-based formation control strategies. The first strategy uses relative position information configured in a bidirectional ring topology to maintain the formation. The second strategy injects interrobot damping via passivity techniques. The third strategy accounts for actuator saturation. Hardware results on differentially driven wheeled mobile robots demonstrate the effectiveness of the proposed control strategies. [ABSTRACT FROM AUTHOR]

In this chapter, we apply the consensus-based design scheme in Section 8.4.1 to formation control of multiple wheeled mobile robots with a virtual leader. We propose a unified, distributed formation control architecture that accommodates an arbitrary number of subgroup leaders and arbitrary information flow among the robots. The architecture requires only neighbor-to-neighbor information exchange. In particular, the consensus tracking algorithm introduced in Chapter 3 is applied on the group level to guarantee consensus on the time-varying group reference trajectory in a distributed manner. Based on the group-level consensus tracking algorithm, a consensus-based formation control strategy is then applied for vehicle level control. The proposed architecture is experimentally implemented and validated on a multirobot platform under neighbor-to-neighbor information exchange with both single and multiple subgroup leaders involved. [ABSTRACT FROM AUTHOR]

This chapter extends the consensus algorithms for rigid body attitude dynamics in Chapter 6 in two ways. First, we propose algorithms to guarantee that multiple rigid bodies can maintain given relative attitudes and angular velocities during the transition. Both fixed and time-varying relative attitude maintenance are considered. Second, we propose algorithms to guarantee that multiple rigid bodies can track a given, time-varying, reference attitude when the reference attitude is available only to a subgroup of team members. Both cases where the attitudes are represented by Euler parameters and modified Rodriguez parameters are considered. Simulation results on reference attitude tracking are presented to validate our algorithms. [ABSTRACT FROM AUTHOR]

In this chapter, we apply the consensus-based design scheme in Section 8.4.1 to two applications, rendezvous and axial alignment. In the rendezvous application, multiple mobile robots simultaneously arrive at a common a priori unknown location determined through team negotiation. In the axial alignment application, multiple mobile robots collectively align their final positions along a line. The two applications rely on the consensus algorithms introduced in Chapter 2, and are experimentally implemented and validated on a mobile actuator and sensor network platform subject to directed, possibly switching interaction topologies to explore issues and challenges in cooperative control with communication constraints. The experimental results show the effectiveness and robustness of the consensus algorithms, even in the presence of physical platform limitations, packet loss, and information delay. [ABSTRACT FROM AUTHOR]

This chapter introduces a few design methodologies for distributed multivehicle cooperative control based on consensus algorithms. Our approach to distributed multivehicle cooperative control problems can be summarized in four steps: (1) the definition of a cooperation constraint and cooperation objective; (2) the definition of a coordination variable as the minimal amount of information needed to effect cooperation; (3) the design of a centralized cooperation strategy; and (4) the use of consensus algorithms to transform the centralized strategy into a distributed strategy. In addition, we overview research in formation control and UAV cooperation. [ABSTRACT FROM AUTHOR]

This chapter extends the consensus algorithms for double-integrator dynamics in Chapter 4 to a reference model. We first consider consensus algorithms that ensure that the derivatives of the information states follow a reference. Both strategies with and without coupling between neighbors' information state derivatives are addressed. We then consider consensus algorithms that ensure that the information states and their derivatives evolve according to a reference model. Three strategies, full access to the reference model, leader-following strategy, and partial access to the reference model, are addressed. [ABSTRACT FROM AUTHOR]

This chapter considers consensus algorithms for rigid body attitude dynamics. We propose algorithms for attitude consensus among multiple rigid bodies for three different cases. In the first case, multiple rigid bodies align their attitudes with zero final angular velocities. In the second case, multiple rigid bodies converge to a constant reference attitude while aligning their attitudes during the transition with zero final angular velocities without the requirement for absolute and relative angular velocity measurements. In the third case, multiple rigid bodies converge to the same angular velocity while aligning their attitudes during the transition. Simulation results for attitude consensus among six rigid bodies demonstrate the effectiveness of the algorithms. [ABSTRACT FROM AUTHOR]

This chapter studies consensus tracking algorithms for single-integrator dynamics, where the information states of all vehicles approach a constant or time-varying consensus reference state. We first propose and analyze a consensus tracking algorithm with a constant consensus reference state. We then propose consensus tracking algorithms with a time-varying consensus reference state and show conditions under which consensus is reached on the time-varying consensus reference state in four cases, a directed fixed interaction topology, a directed switching interaction topology, bounded control inputs, and vehicle information feedback. The time-varying consensus reference state may be an exogenous signal or evolve according to a nonlinear model. These consensus tracking algorithms are also extended to achieve relative state deviations between vehicles. [ABSTRACT FROM AUTHOR]

This chapter introduces consensus algorithms for double-integrator dynamics that take into account motions of the information states and their derivatives, extending the algorithms for single-integrator dynamics. We first propose a fundamental consensus algorithm and derive conditions under which consensus is reached over directed fixed and switching interaction topologies, respectively. Unlike the single-integrator case, we show that having a directed spanning tree is a necessary rather than a sufficient condition for consensus-seeking with double-integrator dynamics. We focus on a formal analysis of interaction topologies that permit consensus for double-integrator dynamics. Given its importance for the stability of the coordinated system, an analysis of the consensus term control gains is also presented, specifically the strength of the information states relative to their derivatives. We then propose two other consensus algorithms that, respectively, take into account bounded control effort and remove the requirement for measurements of relative information state derivatives. We derive conditions under which consensus is reached for these two algorithms over an undirected fixed interaction topology. [ABSTRACT FROM AUTHOR]

This chapter introduces consensus algorithms for single-integrator dynamics. Both continuous-time and discrete-time consensus algorithms are analyzed under both fixed and dynamically changing interaction topologies. This chapter shows, respectively, that consensus under fixed interaction topologies can be achieved asymptotically if and only if the directed interaction topology has a directed spanning tree and consensus under dynamically changing interaction topologies can be achieved asymptotically if there exist infinitely many uniformly bounded, consecutive time intervals such that the union of the directed interaction topologies across each such interval has a directed spanning tree. [ABSTRACT FROM AUTHOR]

This chapter overviews consensus algorithms in cooperative control. The motivation for information consensus in cooperative control is given. A literature review on consensus algorithms is provided. Theoretical results regarding consensus-seeking under both time-invariant and dynamically changing communication topologies are summarized. A few specific applications of consensus algorithms to multivehicle cooperative control are described. The organization of the monograph is also introduced. [ABSTRACT FROM AUTHOR]

In this chapter we apply the I&I methodology to three examples of electromechanical systems. The first is a simple model of a generator connected to an infinite bus, where the objective is to estimate the rotor angle and the total inductance of the bus. This model has been used extensively in the study of the transient stability of power systems1 and, despite its simplicity, it poses a challenging observer design problem due to the fact that it is nonlinear in the unmeasured states. [ABSTRACT FROM AUTHOR]

In this chapter we revisit some of the examples of mechanical systems introduced in Chapters 2 and 5, and design controllers using the I&I approach. First, we extend the result presented in Proposition 2.4 to a more general class of flexible joints robots. In particular, we show that I&I provides a solution to the problem of global tracking for flexible joints robots with full inertia matrix. [ABSTRACT FROM AUTHOR]

This chapter presents four case studies which illustrate the applicability of the I&I technique to practical control design problems for electrical systems and in particular to the control of power electronic systems. First, we describe an application of the methodology in Section 3.1 to the design of a controller for a thyristor-controlled series capacitor (TCSC), which is a power electronic device that is used to regulate the power flow in an AC transmission line. Then, the results of Chapters 3, 5 and 6 are applied to the design of nonlinear adaptive controllers for single-phase power converters and an experimental study of their properties is carried out. [ABSTRACT FROM AUTHOR]

A novel approach to stabilisation and trajectory tracking for nonlinear systems with unknown parameters and uncertain disturbances is developed in this chapter. We depart from the classical adaptive control approach consisting of a parameterised feedback law and an identifier, which tries to minimise a tracking (or prediction) error. Instead, we propose a simple nonlinear proportional-integral (PI) structure that generates a stable error equation—similar to the off-the-manifold dynamics obtained in the I&I approach—with a perturbation function that exhibits at least one root. Trajectories are forced to converge to this root by suitably adjusting the proportional and integral gains of the nonlinear PI controller. [ABSTRACT FROM AUTHOR]

In this chapter the I&I methodology is applied to the problem of asymptotic stabilisation by means of output feedback. It is known that for general nonlinear systems the combination of an exponentially convergent observer with a globally stabilising, state feedback control law may lead to instability and even finite escape time1. One way to overcome this problem (without modifying the observer design) is to construct state feedback controllers that ensure the closed-loop system is either input-to-state stable or Lp stable with respect to the state estimation error. A less conservative condition can be obtained by rewriting the closed-loop dynamics as a system with a globally asymptotically stable equilibrium perturbed by an integrable term2. [ABSTRACT FROM AUTHOR]

In this chapter a general framework for constructing globally convergent (reduced-order) observers for classes of nonlinear systems is presented. Instrumental to this development is to formulate the observer design problem as a problem of rendering invariant and attractive an appropriately selected manifold in the extended state-space of the plant and the observer. [ABSTRACT FROM AUTHOR]

Chapters 2 and 3 have addressed the problem of (adaptive) stabilisation via state feedback for general nonlinear systems, based on the solution of a set of partial differential equations. In this chapter these ideas are exploited to develop constructive adaptive control algorithms for specific classes of nonlinear systems with parametric uncertainties. In particular, the proposed approach is applied to the class of systems in so-called parametric feedback form, yielding new adaptive control laws with interesting properties. These are compared with the classical adaptive (feedback linearising) controllers by means of a practical example. [ABSTRACT FROM AUTHOR]

In this chapter it is shown how the I&I methodology can be used to develop a novel framework for adaptive stabilisation of nonlinear systems with parametric uncertainties that is not certainty equivalent and does not rely on the assumption of linear parameterisation. A key step is to add to the classical certainty-equivalent control law a new term that, together with the parameter update law, is designed to achieve adaptive stabilisation. The role of this new term is to shape the manifold into which the adaptive system is immersed. [ABSTRACT FROM AUTHOR]

This chapter contains the basic result for I&I stabilisation, namely a set of sufficient conditions for the construction of globally asymptotically stabilising, static, state feedback control laws for general, control affine, nonlinear systems. Note, however, that similar considerations can be given to non-affine systems and to tracking problems, while local versions follow mutatis mutandis. [ABSTRACT FROM AUTHOR]

We present a new method for designing nonlinear and adaptive controllers for (uncertain) nonlinear systems. The method relies upon the notions of system immersion and manifold invariance, which are classical tools from nonlinear regulator theory and geometric nonlinear control, but are used in the present work from a new perspective. We call the new methodology immersion and invariance (I&I). The basic idea of the I&I approach is to achieve the control objective by immersing1 the plant dynamics into a (possibly lower-order) target system that captures the desired behaviour. [ABSTRACT FROM AUTHOR]

Design and control of Power Supplies (PSs) feeding the magnets of a Synchrotron Light Source have to match severe specifications; high accuracy in the range of ppm in output current tracking is required for the correct operation of the magnets, while a Power Factor (PF) close to the unit is demanded at the input section due to the high power involved. In this paper an advanced control strategy is presented for a particular kind of Quadrupole Magnet Power Supply, where variable output current has to be imposed. The case of the "switch-mode" multilevel power converter for booster quadrupole magnets of the DIAMOND synchrotron radiation facility under construction at the Harwell Chilton Science Campus, Didcot, has been considered. High accuracy in the tracking of the desired output current reference is reached by means of a digital internal model-based controller. A multivariable controller is adopted in order to ensure current balancing between the stages of the multilevel converter. Front-end topology selection, proper dimensioning and control design are exploited to guarantee high power factor and low harmonic distortion of the input currents, and to avoid low-frequency components related to the quadrupole magnets' oscillating currents. For this purpose, confined oscillatory behavior imposed to the voltage of the DC-link capacitors plays a key role. Simulations and experimental validations are reported that confirm the expected results. [ABSTRACT FROM AUTHOR]

This paper gives an overview of a framework for analyzing hybrid dynamical systems. The emphasis is on modeling assumptions that guarantee robustness. These conditions lead to a general invariance principle and to results on the existence of smooth Lyapunov functions (converse theorems) for hybrid systems. In turn, the stability analysis tools motivate novel hybrid control algorithms for nonlinear systems. [ABSTRACT FROM AUTHOR]

This chapter presents a structural analysis approach for the design of fault tolerant estimation algorithms. The general fault tolerance problem setting is first given, and structural analysis is presented in the component based modeling frame. An original condition for structural observability is developed, which is constructive, since it allows to identify those Data Flow Diagrams by which unknown variables can be estimated, both in healthy and in faulty conditions. The link with two basic dependability concepts, namely critical faults and reliability is shown. [ABSTRACT FROM AUTHOR]

The Hamiltonian approach has turned out to be an effective tool for modeling, system analysis and controller design in the lumped parameter case. There exist also several extensions to the distributed parameter case. This contribution presents a class of extended distributed parameter Hamiltonian systems, which preserves some useful properties of the well known class of Port Controlled Hamiltonian systems with Dissipation. In addition, special ports are introduced to take the boundary conditions into account. Finally, an introductory example and the example of a piezoelectric structure, a problem with two physical domains, show, how one can use the presented approach for modeling and design. [ABSTRACT FROM AUTHOR]

In this chapter, we present the scenario approach, an innovative technology for solving convex optimization problems with an infinite number of constraints. This technology relies on random sampling of constraints, and provides a powerful means for solving a variety of design problems in systems and control. Specifically, the virtues of this approach are here illustrated by focusing on optimal control design in presence of input saturation constraints. [ABSTRACT FROM AUTHOR]

The goal of this paper is to propose a unique vision able to frame a number of results recently proposed in literature to tackle problems of output regulation for nonlinear systems. This is achieved by introducing the so-called asymptotic internal model property as the crucial property which, if fulfilled, leads to the design of the regulator for a fairly general class of nonlinear systems satisfying a proper minimum-phase condition. It is shown that recent frameworks based upon the use of nonlinear high-gain and adaptive observer techniques for the regulator design can be cast in this setting. A recently proposed technique for output regulation without immersion is also framed in these terms. [ABSTRACT FROM AUTHOR]

This paper surveys recent work of the author with several collaborators, principally Feng Lin, Weilin Wang, and Tae-Sic Yoo; they are kindly acknowledged. Decentralized control of discrete event systems, where local controllers cannot explicitly communicate in real-time, is considered in the first part of the paper. Then the problem of real-time communication among a set of local discrete-event controllers (or diagnosers) is discussed. The writing is descriptive and is meant to inform the reader about important conceptual issues and some recently-completed or on-going research efforts. [ABSTRACT FROM AUTHOR]

We offer an alternative way of evaluating the relative importance of the state coordinates of a nonlinear control system. Our approach is based on making changes of state coordinates to bring the controllability and observability functions into input normal form. These changes of coordinates are done degree by degree and the resulting normal form is unique through terms of degree seven. [ABSTRACT FROM AUTHOR]

We introduce two conceptual models for wireless sensing and control with power-limited sensors and controllers. The limited battery power of the wireless device is captured in the models by imposing hard constraints on either the number of available transmissions the device can make, or on the number of cycles it can stay awake. Such hard constraints can be viewed as a measurement budget, under which estimation or control policies will have to be developed over a given decision horizon. Among the two representative models studied here, the first one is one of optimal scheduling of a finite measurement budget for a Gauss-Markov process over an observation horizon. The second one is an optimal estimation problem where the number of transmissions the wireless sensor can make is limited to a number, M, which is less than the observation horizon, N. It is shown that both problems can be solved by employing dynamic-programming type arguments, and their solutions have a threshold characterization. [ABSTRACT FROM AUTHOR]

Petri nets are a family of powerful discrete event models whose interest has grown, within the automatic control community, in parallel with the development of the theory of discrete event systems. In this tutorial paper our goal is that of giving a flavor, by means of simple examples, of the features that make Petri nets a good model for systems theory and of pointing out at a few open areas for research. We focus on Place/Transitions nets, the simplest Petri net model. In particular we compare Petri nets with automata, and show that the former model has several advantages over the latter, not only because it is more general but also because it offers a better structure that has been used for developing computationally efficient algorithms for analysis and synthesis. [ABSTRACT FROM AUTHOR]

In this paper an internal model based approach to periodic input disturbance suppression for port-Hamiltonian systems is presented; more specifically, an adaptive solution able to deal with unknown periodic signal belonging to a given class is introduced. After an introductive section, the adaptive internal model design procedure is presented in order to solve the input disturbance problem. This theoretical machinery is specialized for the energy-based port-Hamiltonian framework in order to prove the global asymptotical stability of the solution. Finally, in order to clearly point out the effectiveness of the presented design procedure a tracking problem is solved for a robotic manipulator affected by torque ripples. [ABSTRACT FROM AUTHOR]

The minimum phase property is an important notion in systems and control theory. In this paper, a characterization of the minimum phase property of nonlinear control systems in terms of a dissipation inequality is derived. It is shown that this dissipation inequality is equivalent to the classical definition of the minimum phase property in the sense of Byrnes and Isidori, if the control system is affine in the input and the so-called input-output normal form exists. [ABSTRACT FROM AUTHOR]

The extension of application domains of robotics from factories to human environments leads to implementing proper strategies for close interaction between people and robots. In order to avoid dangerous collision, force and vision based control can be used, while tracking human motion during such interaction. [ABSTRACT FROM AUTHOR]

In the average consensus a set of linear systems has to be driven to the same final state which corresponds to the average of their initial states. This mathematical problem can be seen as the simplest example of coordination task and in fact it can be used to model both the control of multiple autonomous vehicles which all have to be driven to the centroid of the initial positions, and to model the decentralized estimation of a quantity from multiple measure coming from distributed sensors. In general we can expect that the performance of a consensus strategy will be strongly related to the amount of information the agents exchange each other. This contribution presents a consensus strategy in which the exchanged data are symbols and not real numbers. This is based on a logarithmic quantizer based state estimator. The stability of this technique is then analyzed. [ABSTRACT FROM AUTHOR]

This chapter presents an overview on the modeling and model-based control of autonomous helicopters. Firstly it introduces some of the platforms and control architectures that has been developed in the last 15 years. Later, the Chapter considers the modeling of the helicopter and the identification techniques. Then, it overviews different linear and non-linear model-based control approaches. This section also includes experiments on the control of the helicopter vertical motion that illustrate the presented techniques and point out the interest of nonlinear analysis methods to study the dynamic behavior of the helicopter. Finally, the Chapter presents open research lines coming from two challenging applications: the autonomous landing in oscillating platforms and the lifting and transporting of a single load with several helicopters. [ABSTRACT FROM AUTHOR]

The antiwindup and override design techniques have been developed in industrial design practice for controlling plants with input and output constraints. At present, various analysis and synthesis techniques are well established. These techniques have shown to be both effective and efficient for plants of dominant low order, which are either asymptotically stable or at least marginally stable. [ABSTRACT FROM AUTHOR]

Often one desires to control a nonlinear dynamical system in an optimal way taking constraints on the states and inputs directly into account. Classically this problem falls into the field of optimal control. Often, however, it is difficult, if not impossible, to find a closed solution of the corresponding Hamilton-Jacobi-Bellman equation. One possible control strategy that overcomes this problem is model predictive control. In model predictive control the solution of the Hamilton-Jacobi-Bellman equation is avoided by repeatedly solving an open-loop optimal control problem for the current state, which is a considerably simpler task, and applying the resulting control open-loop for a short time. [ABSTRACT FROM AUTHOR]

The aim of this chapter is the presentation of a technique for the design of stabilizing dynamic output feedback controllers for discrete-time linear systems with rate and amplitude saturating actuators. Two synthesis objectives are considered: the maximization of the domain of attraction of the closed-loop system under some time-domain performance constraints; and the guarantee that the trajectories of the system are bounded in the presence of L2-bounded disturbances, while ensuring an overbound for the L2 gain from the disturbance to the regulated output. [ABSTRACT FROM AUTHOR]

This chapter deals with two problems on stabilization of linear discrete systems by static feedbacks which are bounded and time-delayed, namely global asymptotic stabilization and finite gain lp-stabilizability, p ∈ [1,∞]. Furthermore, an explicit construction of the corresponding feedback laws is given. The feedback laws constructed also result in a closedloop system that is globally asymptotically stable. [ABSTRACT FROM AUTHOR]

We are concerned here with the stabilization of a linear time invariant system subject to actuator saturation via decentralized control while using linear time invariant dynamic controllers. When there exists no actuator saturation, i.e. when we consider just linear time invariant systems, it is known that global stabilization can be done via decentralized control while using linear time invariant dynamic controllers only if the so-called decentralized fixed modes of it are all in the open left half complex plane. [ABSTRACT FROM AUTHOR]