Geo-synchronous SPACEWAY® Ka-band satellite system offers a wide range of user data rates and services on a "bandwidth-on-demand" (BOD) basis. The 30 GHz user uplink (UL) employs narrow-band FDMA/TDMA, whereas 20 GHz user downlink (DL) employs broadband TDMA. Satellite payload consists of a fixed multi-beam receive antenna, a bank of onboard digital de-multiplexing/demodulation/decoding processors, a dynamic packet routing switch, a bank of multiplexing/modulation/encoding processors, a bank of power amplifiers and a hopping-beam transmit array antenna per polarization. This payload architecture enables single-hop, low-latency, full-meshed, user terminal to user terminal connectivity by avoiding double hops to a central hub. Delivery of services to users is provided via low-cost, small-aperture terminals (e.g., 1W HPA and <;0.7m antenna). SPACEWAY® satellite supports a peak capacity of 10 Gbps of raw user data when operating in a point to point (PTP) transmission mode. Dynamic power control is employed to minimize co-channel interference in the user uplinks, and to combat rain fades in both up- and down- links. Payload bandwidth and time slots are assigned based on demand.

The performance of coherent QPSK, offset QPSK, and MSK modulation systems is investigated via computer simulation for TDMA transmission over satellite channels with cascaded nonlinear elements, up- and downlink fades, intersymbol interference, adjacent channel interference, and thermal noise. Three satellite transponder concepts are considered: TWTA transponders nominally operated at their saturation power under clear sky conditions, TWTA transponders nominally operated at input overdrive, and hard-limiting transponders. The performance of these modulation techniques in terms of bit error rate versus energy per information bit over thermal noise power densityE_{b}/N_{0}has been evaluated for the three system concepts as a function of various system parameters. For transponderBTproducts in the range of 1.5-2.0, modem filter characteristics and waveform shaping have been carefully and nearly "optimally" selected to combat adjacent channel interference, which is the major source of impairments, especially under uplink fades. It has been found that, for certain system environments, MSK and OQPSK outperform QPSK. As a byproduct of this investigation, various means to minimize the effects of adjacent channel are proposed.

The modulation transfer noise caused by the "on-off" bursts of TDMA signals is modeled and analyzed in detail. Expressions of the power spectra of the RF and the baseband modulation transfer noise have been obtained for the most general case ofLTDMA andNFM carriers in a zero-memory nonlinear amplifier such as a helix-type TWTA. From these general expressions, the RF power spectrum and the baseband modulation transfer noise onto the FM carriers can be readily computed. It is shown that there are only line spectra and no continuous spectrum in the modulation transfer noise caused by the TDMA on-off bursting. The separation of the spectral lines is equal to the TDMA frame rate. The maximum of the power spectrum of the modulation transfer noise is inversely proportional to the square of the TDMA frame duration and is independent of the separation between TDMA and FM carriers. The modulation transfer noise is in general a complicated function of TDMA burst levels, burst patterns, burst lengths, FM carrier levels, carrier sizes, device nonlinearities, and operating points.

When TDMA and FM carriers are commonly amplified by a memoryless device with nonlinear amplitude and phase characteristics, modulation transfer noise induced by the on-off bursting of the TDMA signals will be present at the FM carriers. In Part II of this paper, closed form expressions of the power spectra of the modulation transfer noise are derived for many special carrier configurations of interest. Numerical examples are also provided in terms of the signal-to-modulation transfer noise ratio, SNR, in the baseband of the FM carriers. It is shown how SNR varies as a function of TDMA burst levels, burst patterns, burst lengths, FM carrier levels, carrier sizes, and device nonlinearities and operating points. With these detailed analyses and numerical examples, insight can be gained into the mechanism of modulation transfer from TDMA on-off bursting to the FM signals. Results presented in this part of the paper should be useful for transmission systems planning and evaluation, and for TDMA burst scheduling.

A unified approach is presented for evaluation of the error probabilities of a class of digital communications systems in additive noise and interference. This class of systems includes coherent systems such asM-ary amplitude-shift keying (ASK),M-ary phase-shift keying (PSK), andM-ary amplitude-and-phase keying (APK); it also includes differential coherent systems such as binary differential PSK (DPSK). The noise is not necessarily Gaussian. The interference can be intersymbol interference, co-channel interference, adjacent-channel interference, any of their linear combinations, or intermodulation prodducts at the output of some nonlinear device. This approach essentially expands the characteristic function of the interferences into a power series so that the desired error probability can be evaluated as the sum of terms representing perturbations around the error probability due to additive noise alone. Bounds on three kinds of truncation errors, which are simple and applicable to all aforementioned digital systems, are obtained. As a result, any desired accuracy in the evaluation of the error probabilities can be achieved with this approach. In the special case in which the noise is Gaussian, explicit bounds on truncation errors are also obtained. Examples are given to illustrate how the unified analysis can be applied to evaluate the error probabilities of various digital systems. More specifically, the combined effects of Gaussian noise, intersymbol interference, and co-channel interference on the error performance ofM-ary coherent PSK and APK (MCPSK and MCAPK) systems are computed. The probability of error of a binary DPSK (BDPSK) system in the presence of Gaussian noise and intersymbol interference is analyzed. The intermodulation products at the output of a hardlimiter are also determined.

A useful method is presented for the evaluation of the error probability ofM-ary phase-shift-keyed systems with Gaussian noise and intersymbol interference. By regarding a portion of the intersymbol interference as an equivalent Gaussian noise, this method suggests first calculating the probability of error caused by the combined effects of the Gaussian noise and the equivalent noise. Necessary correction terms are then added to the calculation to take into account the fact that the intersymbol interference is actually non-Gaussian. It is shown that with this method any desired accuracy can be theoretically achieved in the evaluation of the probability of error. This method has been programmed for numerical computations. Some examples are also given for illustrative purposes.

A variable-length, nonsystematic, convolutional encoding, and successive-decoding scheme is devised to establish significant improvements in the reliability functions of memoryless channels with noiseless decision feedback. It is shown that, for any but pathological discrete memoryless channels with noiseless feedback, there exists a variable-length convolutional code such that the reliability function of the channel can be bounded below by the channel capacityCfor all transmission rates less thanC. By employing a modified version of this scheme, it is also constructively shown that, for an additive-white-Gaussian-noise (AWGN) channel with noiseless feedback it is possible to find a variable-length convolutional code such that the channel reliability function can be bounded below by\alpha_0 c_{infty}for all rates less than the channel capacityC_{\infty}, where\alpha_0 = max (1, \gamma/2)and\gammais the maximum allowable expected-peak-to-expected-average-power ratio at the transmitter.

For the transmission of a Gaussian information source over an additive white Gaussian noise channel, several noiseless-feedback linear estimation schemes are shown to be the same in the sense that they not only achieve the rate-distortion bound on the minimum attainable mean-square error, but also possess identical system parameters. Moreover, Butman's noiseless-linear-feedback scheme for the transmission of a digital information source withMpossible messages over additive white Gaussian noise channels can, in fact, easily be obtained from any of these equivalent schemes by simply imposing a decision structure on it and by using the error probability criterion.

Linear preemphasis and deemphasis filters are found for matching a class of colored Gaussian sources and wide-band additive white Gaussian-noise channels by the utilization of noiseless feedbacks. These filters are optimum in the sense that they give an overall mean-square error that corresponds to the rate-distortion bound.

The performance of a band-limited signaling scheme for an additive white Gaussian noise channel with noiseless feedback is presented.