| A novel trellis-coded-modulation scheme, named as orthogonal plane sequence modulation (OPSM), is introduced whose signal space defines over L (L ≥ 2) orthogonal planes. Its geometric uniformity and binary isometric labeling is verified. Both quasi-synchronous trellis-coded CDMA (QS-TC-CDMA) system and asynchronous trellis-coded CDMA (A-TC-CDMA) system, which are based on OPSM, are presented, respectively, and their theoretical upper bounds are derived under an AWGN channel plus multi-user interference. The average degradation due to multi-user interference is calculated based on the Gaussian quadrature rules (GQR). Gold sequences are utilized as a pseudo-random sequence and Ungerboeck's systematic convolutional encoders with feedback are employed for the trellis-code implementation. Due to the OPSM's multi-planar signal space expansion, a QS-TC-CDMA system based on OPSM improves the system performance over conventional QS-TC-CDMA systems based two-dimensional signaling, and even over conventional multicode QS-TC-CDMA systems that have the same number of sequences per user. An A-TC-CDMA system based on OPSM has better cross-correlation characteristics over conventional biorthogonal signaling based A-TC-CDMA systems, especially at high data rates, so that at practical error rates the proposed scheme has a superior gain to biorthogonal signaling schemes. Furthermore, compared to standard convolutionally-coded CDMA systems, the proposed TC-CDMA system is superior under the same delay conditions (quasi-synchronous or asynchronous), same processing gain, and same decoder complexity. We also simulate the proposed system under a Rician fading channel, which shows that the system performance gain of TC-CDMA system based on OPSM is more pronounced over conventional TC-CDMA systems. OPSM provides several constellation options for a certain data rate so that the proposed system is highly adaptable to dynamic channel conditions and allows a flexible system design. Additionally, for biorthogonal A-TC-CDMA systems, by the transfer function approach, we obtain uniformly tight error-probability upper bounds compared to the previously published approaches. |
