A Fuzzy Classifier for Tactile Sensing

In this paper we propose a fuzzy rule-based algorithm for solvingclassification problems related to tactile sensing. A tactile sensing systemis a robotic device which gives an image of contacting objects. While thefine form recognition problem has been widely discussed and severaltechniques have been proposed for its solution (Bayesian approach or Neuralalgorithms), less attention has been paid to the problem of deciding whetherthe object belongs to a particular class or set of objects that share acommon feature, also known as tactile primitive. As input data we considerthe sum of the normal stresses at the sensing sites. Three levels ofclassification, hierarchically connected, are analyzed and, for each level,different basis variables with their membership functions are proposed andcalibrated using a training procedure. The output is an answer regarding thefeatures of the object at each level and is related to the truth values ofthe fuzzy classes. The numerical experiences show that, at least for dataaffected by low noise level, the algorithm has a very high percentage ofcorrect answers.     

A computer algorithm for the optimization of discrete-time pulse frequency modulated systems

A computer algorithm is described for the optimization of discrete-time pulse frequency modulated systems with state and control constraints. The algorithm, based on a modified maximum principle [1], leads to the solution of a Boolean linear programming problem, for which many computer codes are available commercially. An application to a time-shared sampled-data control system is presented. Numerical examples are given.     

Cooperative neural field for the path planning of a robot arm

This paper deals with the problem of finding a good trajectory, from an initial position to a prescribed target point, for the end effector of a robot arm moving on a two-dimensional work field and avoiding obstacles lying on the work field. Two algorithms based on cooperative neural fields are proposed: the former is suited for the case where the location of obstacles is known, the latter doesn't require any a priori knowledge and is based on a very crude collision detector.     

The Auction Technique for the Sensor Based Navigation Planning of an Autonomous Mobile Robot

The problem of finding a path for the motion of a small mobile robot from a starting point to a fixed target in a two dimensional domain is considered in the presence of arbitrary shaped obstacles. No a priori information is known in advance about the geometry and the dimensions of the workspace nor about the number, extension and location of obstacles. The robot has a sensing device that detects all obstacles or pieces of walls lying beyond a fixed view range. A discrete version of the problem is solved by an iterative algorithm that at any iteration step finds the smallest path length from the actual point to the target with respect to the actual knowledge about the obstacles, then the robot is steered along the path until a new obstacle point interfering with the path is found, at this point a new iteration is started. Such an algorithm stops in a number of steps depending on the geometry, finding a solution for the problem or detecting that the problem is unfeasible. Since the algorithm must be applied on line, the effectiveness of the method depends strongly on the efficiency of the optimization step. The use of the Auction method speeds up this step greatly both for the intrinsic properties of this method and because we fully exploit a property relating two successive optimizations, proved on paper, that in practical instances enables the mean computational cost requested by the optimization step to be greatly reduced. It is proved that the algorithm converges in a finite number of steps finding a solution when the problem is feasible or detecting the infeasibility condition otherwise. Moreover the worst case computational complexity of the whole algorithm is shown to be polynomial in the number of nodes of the discretization grid. Finally numerical examples are reported in order to show the effectiveness of this technique.     

On the approximation of the transport equation by the method of weighted residuals

A general class of approximate models for the transport equation is derived by the Weighted Residual Method using polynomial trial and test spaces. For two, particularly relevant, of such models a closed form of the approximate solution is given in the Laplace-transform domain. A strict connection is established between the approximate transfer function and the Padé approximants to exp(-sx). Furthermore it is shown that Galerkin model minimizes in the considered class a measure of the initial state approximation error. Finally an application to the solution of systems of first order hyperbolic equations is reported.