【正文】 te a variety of mobile munication environments. The performance is analyzed and pared to existing fixed threshold systems through deriving formulas for the detection probability, false alarm rate, and mean acquisition time of the proposed system.This paper is organized as follows. Section 2 describes the acquisition scheme and Section 3 derives the expressions for the detection probability, false alarm rate, and mean acquisition time using the proposed adaptive threshold value method. The numerical results of the proposed system are presented in Section 4 along with parisons to the conventional fixed threshold doubledwell system. The findings and conclusions are then discussed in Section 5.2. System descriptionThe system under consideration in this paper is a doubledwell serial search scheme that consists of two matched filter (MF) correlators connected in a serial manner, as shown in Fig. 1, and two adaptive detectors (ADs). The shorter the correlation tap size, the higher the false alarm rate. Therefore, to pensate for this problem, the second detector is constructed with a long correlation tap size so that the acquisition system can reject those cells that are not in phase quickly. As a result, if the threshold value of the detector is controlled properly, a doubledwell system can both reduce the mean acquisition time and have a high detection probability.The system operates as follows. The transmitted pseudonoise (PN) signal, noise, and any interference arrive at each adaptive detector. If the first AD indicates that the present cell is the correct cell, then the second AD starts working. If the second AD also indicates synchronization, then a decision is made to stop the search. In contrast, if the detector fails to indicate that the present cell is correct, then the relative time delay of the local PN signal is retarded by the chip time △TC for the next search. Usually the value of △ is or 1 [1]. In this paper, △is set at 1. Plus a false alarm is considered to have occurred when both ADs declare synchronization when the received PN signal and locallygenerated PN signal are not actually synchronized. The output distributions of the two Ads are independent because each AD works separately [4].For the adaptive operation of the decision processor, the CACFAR algorithm is selected from among several CFAR algorithms due to its low hardware plexity. The threshold value of the parator in the AD is updated in accordance with the magnitude of the ining signals, as shown in Fig. 2. The correlators in Fig. 2 are IQ noncoherent matched filters. Accordingly, the outputs of the correlator are sent serially to a shift register of length M +1. The first register, denoted as Y, stores the output of the correlator to be tested. The following M registers, denoted by Uj, j=1, 2, … M, called the reference window, store the outputs of the previous M phases. The variable X is the summation of the value in the reference window. The system estimates the background noise power level of the ining signals as follows: (1)The values in the reference window are summed and scaled by T, where T is set according to the desired false alarm rate based on the CACFAR algorithm. Therefore, the adaptive threshold value of an AD is TX.3. Analysis of proposed systemThe detection probability, false alarm rate, and mean acquisition time of the proposed system are derived for additive white Gaussian noise (AWGN) and Rayleigh fading channels. In the derivation of the formulas, the assumptions considered in [2] are also used, namely, (1) there is only one sample corresponding to the correct phase (one H1 cell only), (2) all samples are independent, (3) the correlation tap sizes（≧1），i=1,2, are selected such that the correlation between the received sequence and the local code isabout zero when they are not in phase (H0 cells), (4) the uncertainty region is the full code length L, and (5) for simplicity and to check and pare the trend of the proposed and the conventional systems, Rayleigh fading is slow enough so that the channel is constant over Ni chips as well as fast enough so that the correlator outputs of the adaptive detectors 1 and 2 are essentially independent [6].. AWGN channelIn AWGN channel conditions, the received signal can be written as (2)where b is the received code offset, C(t) is the PN sequence, y is the uniformly distributed random phase （0≦θ≦2π）, is the carrier frequency in radians/second, S is the received signal power, and n(t) is the AWGN with a zero mean and a onesided power spectral density of N0 [Watts/Hertz]. Fig. 3 illustrates the structure of the matched filter used in Fig. 2. Fig. 4 shows the matched filter correlator used in Fig. 3. Referring to Figs. 2–4, the probability density function (pdf) of the H1 sample in AWGN, is a noncentral chisquare distributionwith two degrees of freedom, which can be expressed as [2] (3)Where the noise variance is, the mean signal power is, and is the modified Bessel function of the first kind and zero order. The pdf of the H0 sample in AWGN, is a central chisquare distribution with two degrees of freedom and can be expressed as [2] (4)Since the reference signals in the reference window can be assumed to be noise signals (H0 cells) [4], the pdfs of Uj, j= 1, 2, … M, are the same as (4) and are independent and identically distributed with a distribution of , where is the Gamma distribution. Therefore, the pdf of Uj is (5)Since has a pdf of , the pdf of the random variable X, , can be written as. Therefore, (6)whereis the Gamma function.When the CACFAR and doubledwell systems are bined, the detection probabilities of the ith, i Di, are (7)where represents Marcum’s Qfunction. In (7), Marcum’s Qfunction is difficult to pute accurately. Even when the detection probabil