Fernando Puente León

Full length article: Information fusion for automotive applications - An overview

This article focusses on the fusion of information from various automotive sensors like radar, video, and lidar for enhanced safety and traffic efficiency. Fusion is not restricted to data from sensors onboard the same vehicle but vehicular communication systems allow to propagate and fuse information with sensor data from other vehicles or from the road infrastructure as well. This enables vehicles to perceive information from regions that are hardly accessible otherwise and represents the basis for cooperative driving maneuvers. While the Bayesian framework builds the basis for information fusion, automobile environments are characterized by their a priori unknown topology, i.e., the number, type, and structure of the perceived objects is highly variable. Multi-object detection and tracking methods are a first step to cope with this challenge. Obviously, the existence or non-existence of an object is of paramount importance for safe driving. Such decisions are highly influenced by the association step that assigns sensor measurements to object tracks. Methods that involve multiple sequences of binary assignments are compared with soft-assignment strategies. Finally, fusion based on finite set statistics that (theoretically) avoid an explicit association are discussed.

Deflectometric Measurement of Specular Surfaces

The presented deflectometric method utilizes the surface under inspection as a mirror in a known surrounding and achieves a highly accurate measurement of gradient errors. The comparison of the measured imaging function with reference data allows a precise assessment of the surface quality. Since this approach mimics the inspection behaviour of humans, the results match the errors perceived by a human. However, it is often difficult to provide adequate reference data. A solution to this problem could be based on using CAD data of the part, which is often available from the product design process

Design and capabilities of the Munich Cognitive Automobile

This paper presents the design of the cognitive automobile in Munich. The focus of the capabilities shown here is the navigation on highways and rural roads. The emphasis on higher speed requires early detection of far field objects, so a multi focal active vision with gaze control is essential. For increased robustness lidar range sensors are combined with vision using an object fusion approach. An elaborate safety concept and a verification stage ensure a safe behavior of the vehicle in all situations. A communication system enables the vehicle to perform cooperative perception and action together with similar intelligent vehicles.

Enhanced imaging by fusion of illumination series

A novel image processing method is presented that allows to obtain images with maximal contrast by means of fusing a series of images which were acquired under different illumination constellations. For this purpose, the notion of illumination space is introduced, and strategies for sampling this space are discussed. It is shown that the signal of interest contained in a physical texture often would be lost if standard image acquisition methods were used. In contrast to this, the presented approach shows a robust and reproducible way to obtain high-contrast images containing the relevant information for subsequent processing steps.© (1997) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Non-linear, shape independent object tracking based on 2D lidar data

The paper presents a new lidar-based approach to object tracking. To this end, range data are recorded by two vehicle-born lidar scanners and registered in a common coordinate system. In contrary to common approaches, particle filters are employed to track the objects. This ensures no linearization of the underlying non-linear process model and, thus, a decreasing estimation error. For the object association, a new method is proposed that considers the knowledge about the object shape as well. Based on a statistical formulation, this ensures a robust object assignment even in ambiguous traffic scenes.

Information fusion to detect and classify pedestrians using invariant features

A novel approach to detect pedestrians and to classify them according to their moving direction and relative speed is presented in this paper. This work focuses on the recognition of pedestrian lateral movements, namely: walking and running in both directions, as well as no movement. The perception of the environment is performed through a lidar sensor and an infrared camera. Both sensor signals are fused to determine regions of interest in the video data. The classification of these regions is based on the extraction of 2D translation invariant features, which are constructed by integrating over the transformation group. Special polynomial kernel functions are defined in order to obtain a good separability between the classes. Support vector machine classifiers are used in different configurations to classify the invariants. The proposed approach was evaluated offline considering fixed sensors. Results obtained based on real traffic scenes demonstrate very good detection and classification rates.

Informationsfusion – Eine Übersicht (Information Fusion – An Overview)

Mit Informationsfusion wird der Prozess bezeichnet, Daten aus unterschiedlichen Sensoren oder Informationsquellen mit dem Ziel zu verknüpfen, neues oder präziseres Wissen über physikalische Größen, Ereignisse und Situationen zu gewinnen. Im Beitrag werden eine Systematisierung der verschiedenen Ansätze und Modelle der Informationsfusion vorgenommen und allgemeine Kriterien bei der Herangehensweise an die Fusionsaufgabe vorgestellt. Ausgehend von den Anforderungen an die Messaufgabe bestimmen die Art und das bekannte oder erlernbare Wissen über die Informationsquellen und die Messobjekte sowie der vertretbare mathematische Aufwand die Wahl der Fusionsalgorithmen. Hierzu wird ein Überblick über verschiedene methodische Ansätze gegeben, welche in vielen Anwendungen eine wichtige Rolle spielen. Information fusion denotes the process of combining data from different sensors or information sources to obtain new or more precise knowledge on physical quantities, events, or situations. This paper undertakes a systematization of the existing models and approaches to information fusion and presents general criteria to accomplish the fusion task. Starting with the requirements of the measurement task, both the type and the knowledge of the information sources and the measurands as well as the mathematical complexity constrain the choice of the fusion algorithms. We give an overview of different approaches which play an important role in many applications.

A Bayesian approach to information fusion for evaluating the measurement uncertainty

The Bayesian approach to uncertainty evaluation is a classical example of the fusion of information from different sources. Basically, it is founded on both the knowledge about the measurement process and the influencing quantities and parameters. The knowledge about the measurement process is primarily represented by the so-called model equation, which forms the basic relationship for the fusion of all involved quantities. The knowledge about the influencing quantities and parameters is expressed by their degree of belief, i.e. appropriate probability density functions that usually are obtained by utilizing the principle of maximum information entropy and the Bayes theorem. Practically, the Bayesian approach to uncertainty evaluation is put into effect by employing numerical integration techniques, preferably Monte-Carlo methods. Compared to the ISO-GUM procedure, the Bayesian approach does not have any restrictions with respect to nonlinearities and calculation of confidence intervals.

Fiber Optic Sensor Network for Lightning Impact Localization and Classification in Wind Turbines

As wind turbines are increasing both in number and in height, they are exposed to a major threat in form of lightning strikes. The protection of these structures from the effects of lightning is an important issue in todays wind turbine development. However, as lightning is random in nature, a complete protection against its damages is not achievable. The presented method for lightning impact localization and classification using a fiber optic current sensor network helps to detect damages caused by lightning and to monitor the blades. The system is connected to the wind turbine control and monitoring system

Model-based analysis of striation patterns in forensic science

We present a new image processing strategy that enables an automated extraction of signatures from striation patterns. To this end, a signal model is proposed that allows a suitable description of the interesting features of forensically relevant striation marks. To provide for a high image quality, several images of the same surface area are recorded under systematically varying conditions. The images obtained are then combined to an improved result by means of appropriate sensor fusion techniques. Based upon the signal model, the signal of interest is concentrated, and a compact representation of the grooves is obtained. To enable an efficient description of the relevant features even in the cases of deformed surfaces or curved striation marks, a straightening of the grooves is performed before. In the following, a meaningful signature describing the information of interest is extracted using the whole length of the grooves. This signature can be used for an objective evaluation of similarity of striation patterns.