Oscar Y. Takeshita

A rank criterion for QAM space-time codes

Space-time coding has been studied extensively as a powerful error correction coding for systems with multiple transmit antennas. An important design goal is to maximize the level of space diversity that a code can achieve. Toward this goal, the only systematic algebraic coding theory so far is binary rank theory by Hammons and El Gamal (see ibid. vol. 46, p.524-42, 2000) for binary phase-shift keying (BPSK) modulated codes defined over binary field and quaternary phase-shift keying (QPSK) modulated codes defined over modulo four finite ring. To design codes with higher bandwidth efficiency, we develop an algebraic rank theory to ensure full space diversity for 2/sup 2k/ quadrature and amplitude modulated (QAM) codes for any positive integer k. The theory provides the most general sufficient condition of full space diversity so far. It includes the BPSK binary rank theory as a special case. Since the condition is over the same domain that a code is defined, the full space diversity code design is greatly simplified. The usefulness of the theory is illustrated in examples, such as analyses of existing codes, constructions of new space-time codes with better performance, including the full diversity space-time turbo codes.

Coded modulation for satellite broadcasting

In this paper, three-level block coded 8-PSK modulations, suitable for satellite broadcasting of digital TV signals, are presented. A design principle to achieve unequal error protection is introduced. The coding scheme is designed in such a way that the information bits carrying the basic definition TV signal have a lower error rate than the high definition information bits. The large error coefficients, formally associated with standard mapping by set partitioning, are reduced by considering a nonstandard partition of an 8-PSK signal set. The bits-to-signal mapping induced by this partition allows the use of suboptimal low-complexity soft-decision decoding of binary block codes. Parallel operation of the first and second stage decoders is possible, for high data rate transmission. Furthermore, there is no error propagation from the first-stage decoder to the second-stage decoder.

Joint Analog and Digital Interference Cancellation

The mitigation of in-band interference from co-located emitters is an increasingly important problem in commercial and military wireless communications systems design. In this work, we present a co-located interference mitigation approach that is targeted for scenarios in which the interfering signal is received at an extremely high power and the delay between the received (over-the-air) and reference (wired) copies of the interferer is not negligible. Such scenarios are common in military applications (e.g., simultaneous jamming and communications). The proposed Joint Analog and Digital Interference Cancellation (JADIC) system is self-configuring and can provide over 100 dB of suppression of wideband and narrowband jammers.

A low complexity packet detection algorithm for a CPM modem

This paper discusses the design of a packet detection algorithm for a burst mode continuous phase modulation (CPM) modem over flat Rayleigh fading channels in the 220 MHz frequency band. The modem design focuses on bandwidth efficiency while maintaining good synchronization performance and low complexity. The packet structure meets the Federal Communications Commission frequency emission mask for the 220 MHz frequency band. The packet detection algorithm uses a degraded maximum likelihood (ML) amplitude estimate to provide the initial timing needed for synchronization and demodulation. The ML amplitude estimate is shown to be a function of a one dimensional parameter search. The parameter search is degraded to reduce complexity while maintaining an accurate amplitude estimate.