José Luis Guerrero

Robust sensor fusion in real maritime surveillance scenarios

This paper presents the design and evaluation of a sensor fusion system for maritime surveillance. The system must exploit the complementary AIS-radar sensing technologies to synthesize a reliable surveillance picture using a highly efficient implementation to operate in dense scenarios. The paper highlights the realistic effects taken into account for robust data combination and system scalability.

Introducing a robust and efficient stopping criterion for MOEAs

Soft computing methods, and Multi-Objective Evolutionary Algorithms (MOEAs) in particular, lack a general convergence criterion which prevents these algorithms from detecting the generation where further evolution will provide little improvements (or none at all) over the current solution, making them waste computational resources. This paper presents the Least Squares Stopping Criterion (LSSC), an easily configurable and implementable, robust and efficient stopping criterion, based on simple statistical parameters and residue analysis, which tries to introduce as few setup parameters as possible, being them always related to the MOEAs research field rather than the techniques applied by the criterion.

A multi-objective approach for the segmentation issue

This work presents and formalizes an explicit multi-objective evolutionary approach for the segmentation issue according to Piecewise Linear Representation, which consists in the approximation of a given digital curve by a set of linear models minimizing the representation error and the number of such models required. Available techniques are focused on the minimization of the quality of the obtained approximation, being the cost of that approximation considered, in general, only for certain comparison purposes. The multi-objective nature of the problem is analysed and its treatment in available works reviewed, presenting an a posteriori approach based on an evolutionary algorithm. Three representative curves are included in the data set, comparing the proposed technique to nine different techniques. The performance of the presented approach is tested according to single and multiobjective perspectives. The statistical tests carried out show that the experimental results are, in general, significantly better than available approaches from both perspectives.

Piecewise Linear Representation Segmentation as a Multiobjective Optimization Problem

Actual time series exhibit huge amounts of data which require an unaffordable computational load to be processed, leading to approximate representations to aid these processes. Segmentation processes deal with this issue dividing time series into a certain number of segments and approximating those segments with a basic function. Among the most extended segmentation approaches, piecewise linear representation is highlighted due to its simplicity. This work presents an approach based on the formalization of the segmentation process as a multiobjetive optimization problem and the resolution of that problem with an evolutionary algorithm.

Domain Transformation for Uniform Motion Identification in Air Traffic Trajectories

In this paper, we will discuss the viability of a proposed algorithm to segment trajectories based on a study case of recorded opportunity traffic. This segmentation is the first step of the reconstruction process of the trajectory. Our algorithm will try to apply specific models for the three movement possibilities in our trajectories: uniform, turn and acceleration. We will cover specifically the parameters and viability (as a part of the general algorithm) of the uniform movement segmentation, centring our study in the appropriate descriptor attribute extracted from available samples expressed in its original domain. In particular, we detail a comparison between a general statistic such as a correlation coefficient against the residue of best linear unbiased estimator.

Initialization procedures for multiobjective evolutionary approaches to the segmentation issue

Evolutionary algorithms have been applied to a wide variety of domains with successful results, supported by the increase of computational resources. One of such domains is segmentation, the representation of a given curve by means of a series of linear models minimizing the representation error. This work analyzes the impact of the initialization method on the performance of a multiobjective evolutionary algorithm for this segmentation domain, comparing a random initialization with two different approaches introducing domain knowledge: a hybrid approach based on the application of a local search method and a novel method based on the analysis of the Pareto Front structure.

PIECEWISE LINEAR REPRESENTATION SEGMENTATION IN NOISY DOMAINS WITH A LARGE NUMBER OF MEASUREMENTS: THE AIR TRAFFIC CONTROL DOMAIN

The importance of time series segmentation techniques is rapidly expanding, due to the growth in collection and storage technologies. Among them, one of the most used ones is Piecewise Linear Representation, probably due to its ease of use. This work tries to determine the difficulties faced by this technique when the analyzed time series shows noisy data and a large number of measurements and how to introduce the information about the present noise in the segmentation process. Both difficulties are met in the Air Traffic Control domain, which exhibits position measurements of aircrafts trajectories coming from sensor devices (basically surveillance radar and aircraft-derived data), being used as the motivating domain. Results from the three main traditional techniques are presented (sliding window, top down and bottom up approaches) and compared with a new introduced approach, the Hybrid Local Residue Analysis technique.

Air traffic control: a local approach to the trajectory segmentation issue

This paper presents a new approach for trajectory segmentation in the area of Air Traffic Control, as a basic tool for offline validation with recorded opportunity traffic data. Our approach uses local information to classify each measurement individually, constructing the final segments over these classified samples as the final solution of the process. This local classification is based on a domain transformation using motion models to identify the deviations at a local scale, as an alternative to other global approaches based on combinatorial analysis over the trajectory segmentation domain.

A robust memetic algorithm with self-stopping capabilities

Evolutionary algorithms exhibit some traditional handicaps: lack of a stopping criterion, slow convergence towards the minimum, etc. Memetic algorithms try to combine the best exploration qualities of population based approaches with the exploitation qualities of local search ones. The proposed solution in this work, Robust Evolutionary Strategy Learned with Automated Termination Criteria (R-ESLAT) uses a memetic approach, combining an evolutionary strategy with derivative-free local search methods, adding as well a termination criteria based on the population diversity, according to the principles of the original ESLAT algorithm. The original algorithm is analyzed and its features improved towards an increased robustness, comparing the results obtained with the Covariance Matrix Adaptation Evolutionary Strategy (CMAES).

Multi-agent Data Fusion Architecture Proposal for Obtaining an Integrated Navigated Solution on UAV's

MAS have already more than proved their effectiveness while dealing with high level distributed problems, but some domains (usually low level ones) are still reluctant to their use, usually on a performance basis. UAVs multisensor integration systems take information coming from different sensors and integrate them into one global positioning solution, with a pre-analyzed fixed data fusion architecture topology in a changing environment. In this paper we will propose a novel adaptative MAS data fusion architecture for this problem, able to change its topology according to its conditions, and thus effectively improving the overall quality of the system.