Controlling a Class of Nonlinear Systems on Rectangles

Department of Mathematics and Computer Science, Technische Universiteit Eindhoven, Eindhoven, North Brabant, Netherlands
IEEE Transactions on Automatic Control (Impact Factor: 2.78). 12/2006; 51(11):1749 - 1759. DOI: 10.1109/TAC.2006.884957
Source: IEEE Xplore


In this paper, we focus on a particular class of nonlinear affine control systems of the form xdot=f(x)+Bu, where the drift f is a multi-affine vector field (i.e., affine in each state component), the control distribution B is constant, and the control u is constrained to a convex set. For such a system, we first derive necessary and sufficient conditions for the existence of a multiaffine feedback control law keeping the system in a rectangular invariant. We then derive sufficient conditions for driving all initial states in a rectangle through a desired facet in finite time. If the control constraints are polyhedral, we show that all these conditions translate to checking the feasibility of systems of linear inequalities to be satisfied by the control at the vertices of the state rectangle. This work is motivated by the need to construct discrete abstractions for continuous and hybrid systems, in which analysis and control tasks specified in terms of reachability of sets of states can be reduced to searches on finite graphs. We show the application of our results to the problem of controlling the angular velocity of an aircraft with gas jet actuators

Full-text preview

Available from:
  • Source
    • "Our work has some relations with theoretical qualitative control techniques used for piecewise linear systems in the field of genetic regulatory networks (Chaves and Gouzé (2011)). The approach is also similar to the domain approaches used in hybrid systems theory, where there are some (controlled) transitions between regions, forming a transition graph (Belta and Habets, 2006; Habets and van Schuppen, 2004). "
    [Show description] [Hide description]
    DESCRIPTION: Abstract The oscillator made of a negative loop of two genes is one of the most classical motifs of genetic networks. In this paper we give two solutions to control such an oscillator by modifying the synthesis rates. Our models are given by Piecewise Affine systems, and the control is very qualitative, taking only two values. The necessary measurements for implementing this control only depend on the fact that some gene is expressed or not. Our goal is to obtain sustained oscillations.
    Full-text · Research · Oct 2015
  • Source
    • "An infinite input word (sequence with elements from ) is accepted by if the word produces at least one run of that visits set infinitely often. 2) Transition Systems: For a partitioned environment cluttered with possibly overlapping and static regions of interest from set , the movement capabilities of a mobile robot can be abstracted into a finite state transition system as in Def. 2. Environment partitions can be obtained with cell decomposition algorithms [35], [36], while details on abstractions suitable for different robot dynamics and cell shapes can be found in [37], [38]. The idea is that every cell from the partition corresponds to a single state in the finite state description, and transitions between states correspond to movement capabilities. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This research proposes a centralized method for planning and monitoring the motion of one or a few mobile robots in an environment where regions of interest appear and disappear based on exponential probability density functions. The motion task is given as a linear temporal logic formula over the set of regions of interest. The solution determines robotic trajectories and updates them whenever necessary, such that the task is most likely to be satisfied with respect to probabilistic information on regions. The robots' movement capabilities are abstracted to finite state descriptions, and operations as product automata and graph searches are used in the provided solution. The approach builds up on temporal logic control strategies for static environments by incorporating probabilistic information and by designing an execution monitoring strategy that reacts to actual region observations yielded by robots. Several simulations are included, and a software implementation of the solution is available. The computational complexity of our approach increases exponentially when more robots are considered, and we mention a possible solution to reduce the computational complexity by fusing regions with identical observations.
    Full-text · Article · Oct 2015 · IEEE Transactions on Automation Science and Engineering
  • Source
    • "This approach is similar to those deducing the global dynamics from a " transition graph " of possible transitions between regions [5]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We consider the problem of global stabilization of an unstable bioreactor model (e.g. for anaerobic digestion), when the measurements are discrete and in finite number ("quantized"), with control of the dilution rate. The model is a differential system with two variables, and the output is the biomass growth. The measurements define regions in the state space, and they can be perfect or uncertain (i.e. without or with overlaps). We show that, under appropriate assumptions, a quantized control may lead to global stabilization: trajectories have to follow some transitions between the regions, until the final region where they converge toward the reference equilibrium. On the boundary between regions, the solutions are defined as a Filippov differential inclusion. If the assumptions are not fulfilled, sliding modes may appear, and the transition graphs are not deterministic.
    Preview · Article · Aug 2015
Show more