Diseño de demostrador back-end para nuevo sistema radar de navegación

Abstract

El presente trabajo consiste en el diseño y construcción del primer demonstrador back-end para el nuevo sistema radar que está siendo actualmente desarrollado por el instituto Fraunhofer FHR. Este nuevo radar coherente de bajo costo para aplicaciones de navegación marítima, que usa arreglos de antenas activos de barrido electrónico (AESA) busca mejorar el desempeño que presentan los radares de navegación convencionales utilizando técnicas de integración de pulsos y de compresión de pulsos. Este trabajo inicia con una definición general de los sistemas radar y una introducción global del nuevo sistema radar en desarrollo mostrando sus principales características y los parámetros que deben ser considerados en el diseño del demostrador back-end. Consecutivamente el diseño final del back-end es presentado a el cual es aplicado el plan de nivel de potencia (power level plan). Este análisis permite estimar el nivel de potencia de señal y de ruido a lo largo de las cadenas de transmisión y de recepción. A continuación y para empezar el proceso de fabricación del demostrador se llevó a cabo selección de todos los componentes electrónicos commercial off-the-shelf (COTS) requeridos en el diseño. La técnica de compresión de impulsos usada implica la transmisión de secuencias de fase binarias. A fin de reducir las interferencias involucradas debido al uso de estas secuencias, un análisis detallado de los lóbulos laterales de la función de autocorrelación se llevó a cabo. Por lo tanto se indican los algoritmos matemáticos de búsqueda de códigos desarrollados y los resultados obtenidos con estos. Una nueva selección en función del máximo nivel de cross-correlación fue necesaria para minimizar la interferencia debida a los objetos situados fuera del rango de detección inequívoca del radar. Finalmente se muestran los resultados de las mediciones realizadas con el demostrador back-end construido con los dispositivos electrónicos previamente seleccionados. SUMMARY The first radar system was created in 1904 by the German Christian Hülsmeyer, who was the first to use radio waves to detect the presence of distant metallic objects but not its distance. His invention was developed for commercial application as it was used to detect approaching ships on the river. However, either the naval authorities or industry showed interest at that time. It was not until war world II when practical radar were secretly developed to meet the needs of the military applications, and until today this continues to be one of their main applications. Due to the introduction of new low-cost and high performance radar systems, the design and manufacturing of commercial civilian radar system for a wide range of purposes have become exceptionally relevant, especially ship-borne and coast surveillance applications. In this field radar systems are widely used for target detection and collision avoidance. Conventional ship- borne navigation radar systems employ mechanically rotated antennas and non-coherent signals with high peak power usually generated by a magnetron. These characteristics require a constant maintenance and limit the general performance of the radar, without mentioning high operational costs. Currently, the department of antenna technology and electromagnetic modeling (AEM), of the Fraunhofer institute for high frequency physics and radar techniques (FHR) is working on the design and manufacturing of a novel low-cost coherent radar for maritime navigation with active electronically scanned array (AESA) antenna. This new system uses coherent RF pulses enabling pulse integration and pulse compression techniques allowing to cover the same distance ranges with a lower peak signal power levels. Thus, there is no need for the magnetron. This system, in order to provide a large distance range of coverage, defines different operational modes, each covering a different radial range and using different pulse shapes. The structure of the radar is based on four active antenna front-ends arranged in a rectangular configuration, each with ±45° coverage. The antenna front-end is basically composed by the RF feeding network, a large number of T/R modules and a linear array antenna of 109 identical monopoles inserted into a parallel-plate waveguide (PPW) and a horn shaped section. Due to the modularity of the proposed system, also single front-ends may be used separately at critical position on a shore for coast or harbor surveillance. This new radar system is expected to achieve all tasks of the conventional ship-borne navigation and coast surveillance radar system reducing as much as possible the manufacturing costs. The low manufacturing costs and also the innovation of this system is accomplished by using, in the system front-end, custom-made mixed-signal integrated circuits, mixed-signal printed circuit boards (PCBs) and a serial feeding network. The system back- end, on the other hand, reduces costs by utilizing low-cost commercial-of-the- shelf (COTS) components which are commonly used for wireless communication systems and wireless LAN. The system back-end refers to the part of the radar system where the amplification, shaping, modulation and filtering processes are implemented, not only for transmission but also for reception mode. The objectives of this thesis involve the design and manufacturing of the first demonstrator of this new system back-end. The design process included a complete analysis of the signal and noise power levels along the transmitter and the receiver. On the other hand, the selection of all the suitable components, keeping always in mind to reduce as much as possible the manufacturing costs, was made as part of the manufacturing process. The system back-end is the RF unit that is constituted by the transmitter and the receiver. The processes of amplification, shaping, modulation and filtering that accustom or refine the signal are performed in this unit. The transmitter must deliver a signal with the correct power level and shape to be transmitted depending on the selected operational mode. Therefore, the modulation of the signal to generate the pulses and sub-pulses necessary for the pulse integration and pulse compression techniques is made by this subsystem. The receiver, on the other hand, must amplify and refine the received low signal power level from the targets to detect and provide a suitable signal-to-noise ratio so the signal processing process that follows will be able to detect the target and determine correctly the distance at which it is placed. The design process of the system back-end was developed to satisfy the power requirements of the system front-end and the signal processing stage. For the construction of the system back-end demonstrator an important design tool known as the power level plan was used. This analysis allows to determine the signal and noise power level along the transmitter and receiver chains due to a specific generated or received signal power level, respectively. In order to make the power level plan analysis as practical as possible, a constant update of the components' parameters was necessary, substituting the estimated values for those given by the selected component datasheets. The final power level plan was done using the measured values for the actual electrical specification of the selected components. This analysis improved the design process as it allowed to adapt the system back-end theoretical design to the commercial available components. The pulse compression technique chosen to be used by this system implicated the generation of binary phased sequences. In order to reduce the interferences involved, a detailed analysis of the autocorrelation sidelobe level was performed, selecting for this system the sequences that provided the lowest sidelobe levels possible. The computational time required by a certain algorithm to calculate the maximum autocorrelation sidelobe level of all the codes of a certain length and select those who presented the smallest one, depends directly on their length. Due to the different sequence lengths required by the system's operational modes, the search for the most suitable codes could not be done with a single algorithm. In this thesis is presented an exhaustive search algorithm which provides as a result all the optimum sequences (minimum autocorrelation side lobe level) of a specific length. This algorithm is restricted by the practical computational time it should use, which for practical values permits results until a sequence length of about 40. Also in this work is presented the development of an evolutionary algorithm which allows to find near optimum codes within an acceptable computational time for sequence lengths up to 100 binary digits. To avoid interference due to the objects placed outside the unambiguous range of the operational mode in used, different sequences must be used from one pulse to the next one based on the maximum cross- correlation level. For the sequences obtained by the developed algorithm a subsequent selection based on the maximum cross-correlation level was done. The built demonstrator shown in this thesis, satisfied all power requirements and met the expected behavior predicted during the design process and the power level plan analysis. However, future work and improvements must be done for this radar system back-end