Search

Adaptive arrays for spatiotemporal signal processing are expected to be effective multipath measures in future high-speed mobile communications systems. However, a problem is the tremendous computational load in a hardware implementation because the number of weights that are adaptively controlled becomes extremely large. Therefore, the key to its practical use is to reduce the computational load. In this paper, an adaptive array for processing subband signals that can substantially reduce the computational load was implemented in hardware by using a commercially available digital signal processor (DSP). First, we illustrated the scale of the hardware by the number of DSPs needed in each unit that performed the required signal processing. Next, we built a simple multipath environment of two waves in an anechoic chamber and performed experiments to demonstrate the basic performance and verified that performance nearly equal to the theoretical values was obtained. © 2002 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 85(10): 19–30, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.1122 [ABSTRACT FROM AUTHOR]

It has been known that an adaptive array antenna is effective for reducing frequency-depending selective fading, which is a problem in wide-band land-mobile communication systems. In an environment in which many delayed signals arrive, it is desirable to improve their SN ratio by using co-phase synthesis. There have been two types of adaptive array antenna for this purpose: an antenna based on time-domain signal processing using a TDL (tapped delay line), and an antenna based on frequency-domain signal processing using a fully reconfigurable filter. The latter has variations in its configuration, and is useful for making practical apparatus. However, its convergence process and behavior in a fading environment have not been clearly understood. In this paper, the performance and the advantages and disadvantages of adaptive array antennas based on various configurations with time- and frequency-domain signal processing are compared. It is found that combinations of some configurations improve their performance. The improvement in performance is also described. © 2002 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 85(9): 35–42, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.1117 Copyright © 2002 Scripta Technica [ABSTRACT FROM AUTHOR]

In digital beamforming (DBF) antennas, received propagation signals are first converted down into baseband inphase and quadrature signals, then DBF signal processing is done. We propose a method for designing two-dimensional (2D) digital filters utilized in such DBF networks to handle the signals with wide fractional bandwidths up to 40%. First, we consider the relationship between DBF and its 2D normalized frequency description. Here, we show that 2D complex coefficient digital filters with nonsymmetric characteristics are necessary to form broadband beams. Next, based on this consideration, we describe the proposed method, which combines spectral transformation and a window method. Note that approximation is achieved such that the beam pattern, including its sidelobe characteristics, from the resulting broadband DBF network is virtually frequency-independent. As a result, it has become possible to design DBF networks suitable for a multibeamforming network for use in the broadband beamspace adaptive array antenna proposed in Sekiguchi, Miura, and Karasawa [9]. © 1998 Scripta Technica. Electron Comm Jpn Pt 3, 81(3): 22–34, 1998 [ABSTRACT FROM AUTHOR]