Cross-coupled DOA trackers

output and the exact derivative of the bandlimited signal used in this simulation. Similar to the previous examples of Figs. 5 and 8, the interpolation error may be controlled by the tradeoff between computing time and accuracy. In this case, only the first six rows of the matrix Cu were retained for an overall error of 075 dB. IV. CONCLUSIONS The design [3] of digital filters whose output equations may, optionally, produce interpolated values at predetermined times is extended, in this correspondence, to those cases that require real-time computation of the interpolation time instants. In all its variants, the method provides the accurate digital replica of an analog prototype, with the exact mapping of the transfer function poles mentioned above, in Section II-C. Due to this accuracy, the proposed design preserves the advantages of optimal filter design methods. For example , an optimal FIR filter may be approximated with an analog filter, which will generate the interpolating digital filter operators. This is illustrated by the linear-phase digital filters of Tables II and III and Figs. 3–5. The main contribution of this correspondence is the simple tradeoff between the accuracy and number of multiplications per interpolation point resulting from the properties of the operator C T 1; T 2. Moreover, in the frequent case of digital filters followed by sampling rate conversion, a unique interpolating digital filter with linear phase was shown to eliminate the additional " interpolation filters " required by conventional approaches. A second class of applications is provided by the simulation software for performance evaluation of communication systems [2]. Thus, analog filters, together with the interpolated output, are accurately and efficiently modeled. In addition, the flexibility of the proposed design methods allows for simple simulation of multirate processes that are frequently encountered in the practice of simulation. Finally, the accuracy of the proposed digital filters can be used to advantage in the design of equipment for performance measurement. Prolate spheroidal wave functions, Fourier analysis, and uncertainty principle T and II, " Bell Syst. Tech. Abstract— A new robust, low complexity algorithm for multiuser tracking is proposed, modifying the two-stage parallel architecture of the estimate-maximize (EM) algorithm. The algorithm copes with spatially colored noise, large differences in source powers, multipath, and crossing trajectories. Following a discussion on stability, the simulations demonstrate an asymptotic and tracking behavior that neither the EM nor a nonparallelized tracker can emulate.