디지털 CATV 망에서 최우도비를 이용한 고차 변조 신호의 검출

Abstract

고화질, 양방향 및 멀티미디어에 대한 서비스의 욕구를 만족시키기 위해서는 광대역, 고속의 정보 전송이 필요하다. 그러나 CATV 망은 제한된 대역폭, 잡음 및 망의 정보량(traffic)에 따라 수시로 변화되는 간섭 신호들로 인하여 전송할 수 있는 단위 시간당 전송 정보량이 제한된다. 따라서 무선 전송에서 효율적이고 안정적으로 정보를 전송하기 위하여 사용되고 있는 적응 변조 기법이 CATV 정보 전송에서도 적용되어야 할 것이다. 특히 제한된 대역폭을 이용하여 단위 시간당 전송하기 위해서는 M-QAM 또는 M-PSK 등의 고차(high order) 변조 방식이 필수적이다. 이러한 고차 변조 방식은 잡음과 간섭에 민감하기 때문에 안정적인 성능을 확보하기 위해서는 높은 신호대 잡음비(SNR)가 요구된다. 그러나 높은 신호대 잡음비를 얻기 위해서 사용해야 하는 전력이 제한되므로, 이를 극복하기 위해서는 채널 부호화의 기법을 이용하여 개선하여야 한다. 채널 부호화 기법 중 LDPC (low density parity check) 부호나 Turbo 부호는 수신기에서 반복 복호를 사용하여 상당량의 부호화 이득(coding gain)을 얻을 수 있으나 이진 반복 복호기를 사용하기 때문에 수신된 고차 변조 신호는 이 반복 복호기를 이용하기 위하여 심벌 신호를 비트 정보로 변환해야 한다. 본 논문에서는 디지털 CATV 망을 이용하여 광대역, 고속의 정보 전송 시 필요한 QAM 및 PSK의 고차변조 신호의 검출 알고리듬과 구현 방법을 제안한다. QAM 신호의 검출(detection)을 위한 구현 알고리듬을 제안하고 이를 PSK 신호에 적용하여 동일한 PAM 신호의 검출기를 이용함으로써 유사하게 신호검출을 수행할 수 있는 알고리듬(algorithm)을 제안한다. 이 알고리듬은 수식이 간단하고, 실제 응용이 가능한 형태로 되어있으며, 기존 이진 데이터 전송에서 사용되던 이진 반복 복호기를 변형 없이 그대로 사용할 수 있다. 또한 QAM과 PSK 신호에 모두 공통적으로 사용할 수 있기 때문에 적응형(adaptive) 변/복조 시스템에도 쉽게 응용이 가능한 특징을 가지고 있다.; Before 1990, the main concern in designing the CATV network had been the downstream transmission from the head-end to the subscribers. At that time, not only was there no study about what kind of additional services were possible in the upstream transmission that transmits from the subscriber to the head-end, but also there was not the necessity for data transmission using upstream channel of the CATV network. Therefore, the CATV network was designed for the downstream transmission, and the equipment and the standard of transmission were for the downstream transmission. After 1990, the various additional communication services such as Internet, VOD (Video-on Demand), and telephony over network became possible with the development of the digital communication technology. The study to offer these additional communication services using the upstream channel of the CATV network has been in its active progress. The upstream channel of the CATV network is 5∼42 MHz and it has poor digital signal transmission environment because of the various source of noise like short-wave broadcasting, citizen band interferences, thermal noise, ingress noise, laser of clipping of the optical equipment, common path distortion, and so on. To provide the bidirectional digital services using CATV network, it is necessary to achieve high spectral efficiency in the return path. However, it is inherently difficult to transmit upstream signal, because of narrow bandwidth, and various kinds of noise and interference. Robust network infrastructure and modulation scheme with high spectral efficiency are required to overcome this problem An M-ary high order modulation/demodulation schemes are a prospective technique to achieve an improved the spectral efficiency without increasing the bandwidth, but it requires high SNR to obtain the sufficient BER performance to sustain stable communication links. Currently, to overcome this disadvantage of the M-ary modulation scheme, which requires high SNR, an error correction coding scheme has been discussed to improve the receiver performance. Turbo and LDPC coding schemes which provide a near Shannon capacity are actively utilized with M-ary modulation because of its noticeable coding gain with iterative decoding at the receiver. In this dissertation, thus, the signal detection algorithms are proposed for high-order modulated signals such as arbitrary rectangular Gray coded QAM and PSK signals. We provide general and simple expressions of bitwise signal detection based on Max-Log-MAP for high order QAM signals. For high order PSK signals, we expand the results of QAM signals. The characteristics of the Gray code mapping such as symmetries and repeated formats of the bit assignment in a symbol among bit groups are applied effectively for the simplification of the LLR expression. In order to reduce the complexity of the Max-Log-MAP algorithm for LLR calculation, we replace the mathematical max or min function of the conventional LLR expression with simple arithmetic functions. A channel gain for bitwise LLR calculation is considered when the perfect channel knowledge is available. In order to effectively derive the M-PSK LLR expression, a two-dimensional PSK signal space is transformed to a one-dimensional PAM signal space. Since the transformed PAM signal space does not preserve perfectly identical characteristics of the original PSK signal space, a virtual constellation concept is introduced to derive a simple LLR expression for M-PSK signals. As a result, it is shown that the PSK LLR expression can be directly derived from the PAM signal space with the virtual constellation, and present a new simple and general LLR expression for an M-PSK signal. In addition, with the provided expressions, implementing methods are proposed for practical application of high order QAM and PSK signals. The proposed expressions and implementing algorithms can be utilized effectively for practical applications such as adaptive modulation and coding schemes and software radio with iterative decoders to convert the demodulated output of Gray coded high order modulated signal into bitwise information or for other related applications.