Bit error rate performance analysis of fast frequency hopping receivers
Dyah Rakhma Aryanti
Date of Issue2016-08-16
School of Electrical and Electronic Engineering
Spread-spectrum system is initially developed for military guidance and communication systems. Currently, this system also attracts significant interest in commercial applications such as license-free industrial, scientific, medical bands, and wireless local area network (IEEE 802.11). Spread-spectrum system also has advantages in interference suppression and energy density reduction hence it reduces the probability of being eavesdropped or jammed by unwanted parties. There are two types of spread-spectrum system, namely, direct-sequence and frequency-hopping (FH). In a direct sequence spread-spectrum communication system a data modulated signal modulates a spreading signal or spreading code. While in a FH system, the data modulated signal modulates a carrier signal whose centre frequency is determined in a pseudo-random manner. FH system can be further categorized into two types, namely, slow frequency hopping (SFH) and fast frequency hopping (FFH). In this dissertation, a FFH spread-spectrum system, where the hopping frequency is larger than the symbol frequency or there are several hops in a symbol, is examined. In a FFH SS system, the same signal is duplicated and then transmitted with different frequencies, hence it increases the robustness of the transmitted signal in fighting against intentional or unintentional jamming and fading. In order to get the advantages of diversity in a FH system, an optimum combining mechanism has to be used. There are many types of combining mechanisms that can be used at a receiver such as self-normalize, product, maximum-likelihood (ML) combining, etc. These receivers show the ability of FFH technique in combating interferences such as partial band noise jamming (PBNJ) and multi-tone jamming (MTJ) as well as fading channels. However, not every combining mechanism yields a better performance with the implementation of diversity. Linear combining receiver, for example, has worse bit-error rate (BER) performance as the diversity level is increased. Therefore, it is important to study effective combining mechanisms. In this dissertation, several combining mechanisms used in receivers are examined. Monte Carlo simulations of the receivers’ performance are also conducted to validate the understanding of the knowledge and analyse the receivers’ performance. A ML receiver has a significantly better performance compared to the linear combining receiver under the composite effect of PBNJ and MTJ. The ML receiver uses the ML combining method in combining the outputs of a square-law detector while the linear combining receiver uses a linear combining method. The ML combining method depends on the ratio of probability density function (pdf) expressions of a square-law detector output while these pdf expressions depend on the communication channel conditions. A different channel condition leads to a different receiver structure. Hence, in designing a receiver structure with ML combining mechanism, deriving the pdf of the square law detector output is a milestone to be accomplished. There are many researches studying ML combining receivers under fading channel. However, to the best of our knowledge, there is no research examining this combining mechanism under Nakagami-m fading channel, AWGN and composite effect of interferences (PBNJ and MTJ). Nakagami-m fading has several advantages. It possesses greater flexibility as it can model a wide range of channel condition even for a channel that is less or more severe than Rayleigh fading. It also provides greater accuracy in matching experimental data than Rayleigh, Rician and Hoyt distribution. Furthermore, Nakagami-m distribution gives the best fit to multipath fading channels. Therefore, a ML receiver structure for Nakagami-m fading channel and interferences including PBNJ, MTJ and AWGN was studied in this dissertation by examining methods used to derive the pdf expressions of a square-law detector output. There are three methods to be examined. The implementation of these three methods to the suggested channel condition is also elaborated.