Search

We define a mathematical notion of complex adaptive system by following the original intuition of G.K. Zipf about the principle of least effort, an intuitive idea which is nowadays informally widespread in complex systems modeling. We call generalized evolution principle this mathematical notion of interaction spaces theory. Formalizing and generalizing Mandelbrot's ideas, we also prove that a large class of these systems satisfy a power law. We finally illustrate the notion of complex adaptive system with theorems describing a Von Th\"unen-like model. The latter can be easily generalized to other complex systems and describes the appearance of emergent patterns. Every notion is introduced both using an intuitive description with lots of examples, and using a modern mathematical language.

We present the first steps of interaction spaces theory, a universal mathematical theory of complex systems which is able to embed cellular automata, agent based models, master equation based models, stochastic or deterministic, continuous or discrete dynamical systems, networked dynamical models, artificial neural networks and genetic algorithms in a single notion. Therefore, interaction spaces represent a common mathematical language that can be used to describe several complex systems modeling frameworks. This is the first step to start a mathematical theory of complex systems. Every notion is introduced both using an intuitive description by listing lots of examples, and using a modern mathematical language.

This paper introduces an end-to-end trajectory planning algorithm tailored for multi-UAV systems that generates collision-free trajectories in environments populated with both static and dynamic obstacles, leveraging point cloud data. Our approach consists of a 2-fork neural network fed with sensing and localization data, able to communicate intermediate learned features among the agents. One network branch crafts an initial collision-free trajectory estimate, while the other devises a neural collision constraint for subsequent optimization, ensuring trajectory continuity and adherence to physicalactuation limits. Extensive simulations in challenging cluttered environments, involving up to 25 robots and 25% obstacle density, show a collision avoidance success rate in the range of 100 -- 85%. Finally, we introduce a saliency map computation method acting on the point cloud data, offering qualitative insights into our methodology.

By means of several examples, we motivate that universal properties are the simplest way to solve a given mathematical problem, explaining in this way why they appear everywhere in mathematics. In particular, we present the co-universal property of Schwartz distributions, as the simplest way to have derivatives of continuous functions, Colombeau algebra as the simplest quotient algebra where representatives of zero are infinitesimal, and generalized smooth functions as the universal way to associate set-theoretical maps of non-Archimedean numbers defined by nets of smooth functions (e.g. regularizations of distributions) and having arbitrary derivatives. Each one of these properties yields a characterization up to isomorphisms of the corresponding space. The paper requires only the notions of category, functor, natural transformation and Schwartz's distributions, and introduces the notion of universal solution using a simple and non-abstract language.

In this paper, we propose the Liquid-Graph Time-constant (LGTC) network, a continuous graph neural network(GNN) model for control of multi-agent systems based on therecent Liquid Time Constant (LTC) network. We analyse itsstability leveraging contraction analysis and propose a closed-form model that preserves the model contraction rate and doesnot require solving an ODE at each iteration. Compared todiscrete models like Graph Gated Neural Networks (GGNNs),the higher expressivity of the proposed model guaranteesremarkable performance while reducing the large amountof communicated variables normally required by GNNs. Weevaluate our model on a distributed multi-agent control casestudy (flocking) taking into account variable communicationrange and scalability under non-instantaneous communication

Neural Style Transfer (NST) refers to a class of algorithms able to manipulate an element, most often images, to adopt the appearance or style of another one. Each element is defined as a combination of Content and Style: the Content can be conceptually defined as the what and the Style as the how of said element. In this context, we propose a custom NST framework for transferring a set of styles to the motion of a robotic manipulator, e.g., the same robotic task can be carried out in an angry, happy, calm, or sad way. An autoencoder architecture extracts and defines the Content and the Style of the target robot motions. A Twin Delayed Deep Deterministic Policy Gradient (TD3) network generates the robot control policy using the loss defined by the autoencoder. The proposed Neural Policy Style Transfer TD3 (NPST3) alters the robot motion by introducing the trained style. Such an approach can be implemented either offline, for carrying out autonomous robot motions in dynamic environments, or online, for adapting at runtime the style of a teleoperated robot. The considered styles can be learned online from human demonstrations. We carried out an evaluation with human subjects enrolling 73 volunteers, asking them to recognize the style behind some representative robotic motions. Results show a good recognition rate, proving that it is possible to convey different styles to a robot using this approach.

In this paper, we aim to find the conditions for input-state stability (ISS) and incremental input-state stability ($\delta$ISS) of Gated Graph Neural Networks (GGNNs). We show that this recurrent version of Graph Neural Networks (GNNs) can be expressed as a dynamical distributed system and, as a consequence, can be analysed using model-based techniques to assess its stability and robustness properties. Then, the stability criteria found can be exploited as constraints during the training process to enforce the internal stability of the neural network. Two distributed control examples, flocking and multi-robot motion control, show that using these conditions increases the performance and robustness of the gated GNNs.

This article is a natural continuation of the paper Tiwari, D., Giordano, P., Hyperseries in the non-Archimedean ring of Colombeau generalized numbers in this journal. We study one variable hyper-power series by analyzing the notion of radius of convergence and proving classical results such as algebraic operations, composition and reciprocal of hyper-power series. We then define and study one variable generalized real analytic functions, considering their derivation, integration, a suitable formulation of the identity theorem and the characterization by uniform upper bounds of derivatives on functionally compact sets. On the contrary with respect to the classical use of series in the theory of Colombeau real analytic functions, we can recover several classical examples in a non-infinitesimal set of convergence. The notion of generalized real analytic function reveals to be less rigid both with respect to the classical one and to Colombeau theory, e.g. including classical non-analytic smooth functions with flat points and several distributions, such as the Dirac delta. On the other hand, each Colombeau real analytic function is also a generalized real analytic function.

We prove that Picard-Lindel\"of iterations for an arbitrary smooth normal Cauchy problem for PDE converge if we assume a suitable Weissinger-like sufficient condition. This condition includes both a large class of non-analytic PDE or initial conditions, and more classical real analytic functions. The proof is based on a Banach fixed point theorem for contractions with loss of derivatives. From the latter, we also prove an inverse function theorem for locally Lipschitz maps with loss of derivatives in arbitrary graded Fr\'echet spaces.

In this research, we aim to answer the question: How to combine Closed-Loop State and Input Sensitivity-based with Observability-aware trajectory planning? These possibly opposite optimization objectives can be used to improve trajectory control tracking and, at the same time, estimation performance. Our proposed novel Control & Observability-aware Planning (COP) framework is the first that uses these possibly opposing objectives in a Single-Objective Optimization Problem (SOOP) based on the Augmented Weighted Tchebycheff method to perform the balancing of them and generation of B\'ezier curve-based trajectories. Statistically relevant simulations for a 3D quadrotor unmanned aerial vehicle (UAV) case study produce results that support our claims and show the negative correlation between both objectives. We were able to reduce the positional mean integral error norm as well as the estimation uncertainty with the same trajectory to comparable levels of the trajectories optimized with individual objectives., Comment: To be published in IEEE International Conference on Robotics and Automation (ICRA), 2022

Using the existence of infinite numbers $k$ in the non-Archimedean ring of Robinson-Colombeau, we define the hyperfinite Fourier transform (HFT) by considering integration extended to $[-k,k]^{n}$ instead of $(-\infty,\infty)^{n}$. In order to realize this idea, the space of generalized functions we consider is that of generalized smooth functions (GSF), an extension of classical distribution theory sharing many nonlinear properties with ordinary smooth functions, like the closure with respect to composition, a good integration theory, and several classical theorems of calculus. Even if the final transform depends on $k$, we obtain a new notion that applies to all GSF, in particular to all Schwartz's distributions and to all Colombeau generalized functions, without growth restrictions. We prove that this FT generalizes several classical properties of the ordinary FT, and in this way we also overcome the difficulties of FT in Colombeau's settings. Differences in some formulas, such as in the transform of derivatives, reveal to be meaningful since allow to obtain also non-tempered global unique solutions of differential equations., Comment: arXiv admin note: text overlap with arXiv:2011.09660

The main aim of the present work is to arrive at a mathematical theory close to the historically original conception of generalized functions, i.e. set theoretical functions defined on, and with values in, a suitable ring of scalars and sharing a number of fundamental properties with smooth functions, in particular with respect to composition and nonlinear operations. This is how they are still used in informal calculations in Physics. We introduce a category of generalized functions as smooth set-theoretical maps on (multidimensional) points of a ring of scalars containing infinitesimals and infinities. This category extends Schwartz distributions. The calculus of these generalized functions is closely related to classical analysis, with point values, composition, non-linear operations and the generalization of several classical theorems of calculus. Finally, we extend this category of generalized functions into a Grothendieck topos of sheaves over a concrete site. This topos hence provides a suitable framework for the study of spaces and functions with singularities. In this first paper, we present the basic theory; subsequent ones will be devoted to the resulting theory of ODE and PDE., Comment: 70 pages, 4 figures

We present an extension of some results of higher order calculus of variations and optimal control to generalized functions. The framework is the category of generalized smooth functions, which includes Schwartz distributions, while sharing many nonlinear properties with ordinary smooth functions. We prove the higher order Euler-Lagrange equations, the D'Alembert principle in differential form, the du Bois-Reymond optimality condition and the Noether's theorem. We start the theory of optimal control proving a weak form of the Pontryagin maximum principle and the Noether's theorem for optimal control. We close with a study of a singularly variable length pendulum, oscillations damped by two media and the Pais-Uhlenbeck oscillator with singular frequencies., Comment: arXiv admin note: text overlap with arXiv:1608.03080

This article is the natural continuation of the paper: Mukhammadiev A.~et al Supremum, infimum and hyperlimits of Colombeau generalized numbers in this journal. Since the ring $\tilde{R}$ of Robinson-Colombeau is non-Archimedean, a classical series $\sum_{n=0}^{+\infty}a_{n}$ of generalized numbers $a_{n}\in\tilde{R}$ is convergent if and only if $a_{n}\to0$ in the sharp topology. Therefore, this property does not permit us to generalize several classical results, mainly in the study of analytic generalized functions (as well as, e.g., in the study of sigma-additivity in integration of generalized functions). Introducing the notion of hyperseries, we solve this problem recovering classical examples of analytic functions as well as several classical results.

This paper addresses the problem of mixed initiative, shared control for master-slave grasping and manipulation. We propose a novel system, in which an autonomous agent assists a human in teleoperating a remote slave arm/gripper, using a haptic master device. Our system is designed to exploit the human operator's expertise in selecting stable grasps (still an open research topic in autonomous robotics). Meanwhile, a-priori knowledge of: i) the slave robot kinematics, and ii) the desired post-grasp manipulative trajectory, are fed to an autonomous agent which transmits force cues to the human, to encourage maximally manipulable grasp pose selections. Specifically, the autonomous agent provides force cues to the human, during the reach-to-grasp phase, which encourage the human to select grasp poses which maximise manipulation capability during the post-grasp object manipulation phase. We introduce a task-relevant velocity manipulability cost function (TOV), which is used to identify the maximum kinematic capability of a manipulator during post-grasp motions, and feed this back as force cues to the human during the pre-grasp phase. We show that grasps which minimise TOV result in significantly reduced control effort of the manipulator, compared to other feasible grasps. We demonstrate the effectiveness of our approach by experiments with both real and simulated robots., Comment: To be appeared in IEEE/RAS IROS 2017

We present an extension of the classical theory of calculus of variations to generalized functions. The framework is the category of generalized smooth functions, which includes Schwartz distributions while sharing many nonlinear properties with ordinary smooth functions. We prove full connections between extremals and Euler-Lagrange equations, classical necessary and sufficient conditions to have a minimizer, the necessary Legendre condition, Jacobi's theorem on conjugate points and Noether's theorem. We close with an application to low regularity Riemannian geometry., Comment: 34 pages, 1 figure

Generalized smooth functions are a possible formalization of the original historical approach followed by Cauchy, Poisson, Kirchhoff, Helmholtz, Kelvin, Heaviside, and Dirac to deal with generalized functions. They are set-theoretical functions defined on a natural non-Archimedean ring, and include Colombeau generalized functions (and hence also Schwartz distributions) as a particular case. One of their key property is the closure with respect to composition. We review the theory of generalized smooth functions and prove both the local and some global inverse function theorems., Comment: 20 pages, minor corrections

In [11], we introduced the notion of asymptotic gauge (AG), and we used it to construct Colombeau AG-algebras. This construction concurrently generalizes that of many different algebras used in Colombeau's theory, e.g. the special one $\mathcal{G}^{\srm}$, the full one $\gse$, the NSA based algebra of asymptotic functions $\hat{\mathcal{G}}$, and the diffeomorphism invariant algebras $\gsd$, $\mathcal{G}^{2}$ and $\hat{\mathcal{G}}$. In this paper we study the categorical properties of the construction of Colombeau AG-algebras with respect to the choice of the AG, and we show their consequences regarding the solvability of generalized ODE.

This paper presents a novel decentralized control strategy for a multi-robot system that enables parallel multi-target exploration while ensuring a time-varying connected topology in cluttered 3D environments. Flexible continuous connectivity is guaranteed by building upon a recent connectivity maintenance method, in which limited range, line-of-sight visibility, and collision avoidance are taken into account at the same time. Completeness of the decentralized multi-target exploration algorithm is guaranteed by dynamically assigning the robots with different motion behaviors during the exploration task. One major group is subject to a suitable downscaling of the main traveling force based on the traveling efficiency of the current leader and the direction alignment between traveling and connectivity force. This supports the leader in always reaching its current target and, on a larger time horizon, that the whole team realizes the overall task in finite time. Extensive Monte~Carlo simulations with a group of several quadrotor UAVs show the scalability and effectiveness of the proposed method and experiments validate its practicability.

We develop the integral calculus for quasi-standard smooth functions defined on the ring of Fermat reals. The approach is by proving the existence and uniqueness of primitives. Besides the classical integral formulas, we show the flexibility of the Cartesian closed framework of Fermat spaces to deal with infinite dimensional integral operators. The total order relation between scalars permits to prove several classical order properties of these integrals and to study multiple integrals on Peano-Jordan-like integration domains., Comment: v2: simplify the thickening for the proof of the Fermat-Reyes theorem in the previous version

We introduce the notion of functionally compact sets into the theory of nonlinear generalized functions in the sense of Colombeau. The motivation behind our construction is to transfer, as far as possible, properties enjoyed by standard smooth functions on compact sets into the framework of generalized functions. Based on this concept, we introduce spaces of compactly supported generalized smooth functions that are close analogues to the test function spaces of distribution theory. We then develop the topological and functional analytic foundations of these spaces., Comment: Some typos corrected, final version

We use the general notion of set of indices to construct algebras of nonlinear generalized functions of Colombeau type. They are formally defined in the same way as the special Colombeau algebra, but based on more general growth condition formalized by the notion of asymptotic gauge. This generalization includes the special, full and nonstandard analysis based Colombeau type algebras in a unique framework. We compare Colombeau algebras generated by asymptotic gauges with other analogous construction, and we study systematically their properties, with particular attention to the existence and definition of embeddings of distributions. We finally prove that, in our framework, for every linear homogeneous ODE with generalized coefficients there exists a minimal Colombeau algebra generated by asymptotic gauges in which the ODE can be uniquely solved. This marks a main difference with the Colombeau special algebra, where only linear homogeneous ODEs satisfying some restriction on the coefficients can be solved., Comment: 30 pages

We define a general notion of set of indices which, using concepts from pre-ordered sets theory, permits to unify the presentation of several Colombeau-type algebras of nonlinear generalized functions. In every set of indices it is possible to generalize Landau's notion of big-O such that its usual properties continue to hold. Using this generalized notion of big-O, these algebras can be formally defined the same way as the special Colombeau algebra. Finally, we examine the scope of this formalism and show its effectiveness by applying it to the proof of the pointwise characterization in Colombeau algebras.

We tackle the problem of finding a suitable categorical framework for generalized functions used in mathematical physics for linear and non-linear PDEs. We are looking for a Cartesian closed category which contains both Schwartz distributions and Colombeau generalized functions as natural objects. We study Fr\"olicher spaces, diffeological spaces and functionally generated spaces as frameworks for generalized functions. The latter are similar to Fr\"olicher spaces, but starting from locally defined functionals. Functionally generated spaces strictly lie between Fr\"olicher spaces and diffeological spaces, and they form a complete and cocomplete Cartesian closed category. We deeply study functionally generated spaces (and Fr\"olicher spaces) as a framework for Schwartz distributions, and prove that in the category of diffeological spaces, both the special and the full Colombeau algebras are smooth differential algebras, with a smooth embedding of Schwartz distributions and smooth pointwise evaluations of Colombeau generalized functions., Comment: 38 pages

Based on a refinement of the notion of internal sets in Colombeau's theory, so-called strongly internal sets, we introduce the space of generalized smooth functions, a maximal extension of Colombeau generalized functions. Generalized smooth functions as morphisms between sets of generalized points form a sub-category of the category of topological spaces. In particular, they can be composed unrestrictedly., Comment: 17 pages, some minor corrections

This work considers the problem of estimating the unscaled relative positions of a multi-robot team in a common reference frame from bearing-only measurements. Each robot has access to a relative bearing measurement taken from the local body frame of the robot, and the robots have no knowledge of a common or inertial reference frame. A corresponding extension of rigidity theory is made for frameworks embedded in the \emph{special Euclidean group} $SE(2) = \mathbb{R}^2 \times \mathcal{S}^1$. We introduce definitions describing rigidity for $SE(2)$ frameworks and provide necessary and sufficient conditions for when such a framework is \emph{infinitesimally rigid} in $SE(2)$. Analogous to the rigidity matrix for point formations, we introduce the \emph{directed bearing rigidity matrix} and show that an $SE(2)$ framework is infinitesimally rigid if and only if the rank of this matrix is equal to $2|\mathcal{V}|-4$, where $|\mathcal{V}|$ is the number of agents in the ensemble. The directed bearing rigidity matrix and its properties are then used in the implementation and convergence proof of a distributed estimator to determine the {unscaled}{} relative positions in a common frame. Some simulation results are also given to support the analysis., Comment: submitted to 2014 European Control Conference