An Efficient Joint Frame and Physical Layer Signaling Code Detection Method for DVB-S2

The Digital Video Broadcasting–Satellite Second Generation (DVB-S2) standard enhances coding and modulation techniques beyond those of DVB-S. It provides a spectrum of coding and modulation options (MODCOD) to suit diverse communication environments. Adaptive coding and modulation (ACM), along with additional signaling in the Physical Layer (PL) Header, are introduced, leading to potential ambiguities in frame synchronization and PL signaling identification for receivers. The straightforward approach is correlating the incoming DVB-S2 signal against all potential PL signaling codes. This process is computationally heavy and becomes impractical with all number of PL signaling codes. Addressing this challenge, this article presents an efficient approach utilizing the fast Walsh-Hadamard transform, significantly simplifying the frame detector’s complexity. Furthermore, the detector facilitates MODCOD detection. The implementation of proposed architecture confirms the method in complexity reduction.


I. INTRODUCTION
The Digital Video Broadcasting -Satellite Second Generation (DVB-S2) is a standard in the realm of satellite broadband communication [1].It was developed by the Digital Video Broadcasting (DVB) Project and officially released by the European Telecommunications Standards Institute (ETSI) in 2003.The DVB-S2 standard incorporates advanced LDPC (Low-Density Parity-Check) code and BCH (Bose-Chaudhuri-Hocquenghem) code, supplanting the Reed-Solomon (RS) code and convolutional code utilized in the DVB-S standard.These enhancements in coding techniques enable DVB-S2 to achieve lower bit error rates (BER) at reduced signal-to-noise ratios (SNR), thereby enhancing overall system performance.Moreover, DVB-S2 extends its modulation capabilities, including up to 32APSK modulation modes.This implies that, with appropriate code rates and SNR conditions, DVB-S2 can more efficiently utilize channel capacity, offering improved data transmission efficiency.
The associate editor coordinating the review of this manuscript and approving it for publication was Tianhua Xu .
The enhanced modulation schemes and coding rates in DVB-S2 afford a suite of flexible options tailored to various transmission scenarios.It brings out these options through adaptive coding and modulation (ACM) and variable coding and modulation (VCM) schemes, which are specified within the physical layer header.This adaptability ensures optimal performance across a spectrum of signal conditions.
The physical header of DVB-S2 dynamically adjusts its content in different transmission conditions, which poses a challenge for the receiver to synchronize the physical layer frame position and identify the signaling transmitted.Unlike the conventional DVB-S system, which relies on fixed frame structure, DVB-S2 with ACM/VCM makes frame synchronization and signaling identification less predictable.
The physical header of DVB-S2 consists of two key components: the Start of Frame (SOF) and the Physical Layer Signaling Code (PLSC).The SOF is a clear marker that indicates the start of each frame and segments the data stream into discrete units, while the PLSC conveys essential information for the receiver to demodulate the frame structure correctly.This strategic integration enables the receivers to identify and process the incoming data frames efficiently, even with varying frame header lengths.
A review of state-of-the-art work indicates that frame synchronization can be divided into two approaches: coherent and noncoherent.Noncoherent approaches refer especially to frequency uncertainty due to Doppler from satellite motion.Differential correlation forms the maximumlikelihood approach, e.g., [2], [3], and [4], and post-detection integral base, e.g., [5], [6], and [7].Another approach uses the modulation property of PL Header to synchronize frames, as in [8].According to the specification [1], DVB-S2 should work even at as low as SNR -2.35dB.Adopting noncoherent approaches gains performance when frequency uncertainty exists but degrades performance in normal schemes.However, coherent approaches are in contrast.Therefore, there is a tradeoff between SNR performance and frequency uncertainty.
Additionally, to improve the signal-to-noise ratio, certain satellite communication systems operate with low-frequency offsets, such as Zhongxing 1E, Galaxy-34, SES-22, Measat 3D, etc.These satellites work in geostationary orbits but face challenges associated with low signal-to-noise ratios.
An intuitive approach for addressing frame synchronization is to create all modes of ACM/VCM correlation, which enables the determination of both the frame's location and the modulation and coding scheme employed.However, this method entails a significant consumption cost of resources in calculation.
This paper presents an efficient frame synchronization method for DVB-S2 system based on the fast Walsh-Hadamard transform.The proposed method reduces the computational complexity and improves the performance under extremely low signal-to-noise ratio.Also a maximum likelihood-based approach, as demonstrated by Massey gain improvement by simply adding a correction term.It is a simple way to implemnetation, therefore the Massey structure is adopted as the DVB-S2 synchronizer in this paper.The peak search algorithm is also optimized by exploiting the properties of the PLSC codes, which further enhances the frame synchronization and reduces the false alarm rate.The proposed method is implemented using FPGA and the resource usage and the synchronization time are evaluated, which demonstrate the feasibility and the efficiency of the method.
The rest of this paper is organized as follows.Section I introduces the Physical Layer frame structure and also gives model definitions to these utilities.Section II derives the PL Header and PLSC detection model and presents the proposed method.Section III focuses on implementing the method using FPGA and evaluating the resource usage and the synchronization time.Section IV compares the proposed method with other existing methods and analyzes the advantages and disadvantages.Section V concludes the paper with comprehensive reports.

II. TRANSMITTER FRAME STRUCTURE
The PL frame of DVB-S2 consists of PL Header and payload [1].The DVB-S2 frame structure is shown in Figure 1. The ).In the signaling information, the first 6 bits are computed with the following generator matrix G,where and each computation generate 32bits ⃗ y = (y 1 y 2 . . .y 32 ).The generated sequence ⃗ y replicates itself and performs exclusive or with the remaining signaling information bit b 7 after each output bit ⃗ p = (y 1 y 1 ⊕ b 7 y 2 y 2 ⊕ b 7 . . .y 32 y 32 ⊕ b 7 ) denoted as (p 1 p 2 . . .p 64 ).The 64-bit serial sequence is scrambled with the code ⃗ c = (c 1 c 2 . . .c 64 ) to obtain the PLSC ⃗ m = (m 1 m 2 . . .m 64 ).Subsequently, the PLSC is integrated with the SOF to form a 90-bit PL Header, which is then modulated using pi/2-BPSK.The physical layer header transmitter system flow chart is depicted in Figure 2.
In their research, Feng-Wen Sun et al. argued that the Start of Frame (SOF) sequence is insufficiently long to provide the receiver with a high degree of confidence for successful frame detection [4].The figure below shows the false detection rate depending on the pattern of SOF and the whole PL Header.
Thus, including the Physical Layer Signaling Code enhanced the probability of detection and emerged as a promising approach.Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.symbol is reasonable as ( 1) and ( 2).The SOF symbol signal can be represented as (1) with time index n.
The m th PLSC signaling symbol signal can be represented as where θ n is the phase of pi/2 BPSK, which is (1 + j)/ The matrix is Hadamard matrix in dyadic order and the rows of W is a Walsh function set [9].Representation the element of y th row x th column is Hence,the unscramble sequence of PLSC s p m n can express as following equation where ⌊ * ⌋ represents the floor function, and m denotes the signaling mode, ranging from 0 to 127.The scalar a m is equal to 1 when the signaling mode is even and −1 when the signaling mode is odd.Finally, the PL frame can express as

III. RECEIVED SIGNALING DETECTION
Assuming perfect synchronization with symbol timing and phase when receiving incoming frame s PLF m n with m th signaling transmitting over an AWGN channel.
where g n is complex AWGN noise with power spectrum density N 0 First, assuming precise frame synchronization at index k, the subsequent task involves the identification of the Physical Layer signaling code index denoted as m from the header.This process is characterized by the conditional probability density function of vector ⃗ r = (r 1 r 2 . . .r PLF ), represented in the equation below:

A. SIGNALING DETECTION
Taking natural logarithm to likelihood function in (11b), a joint optimum decision of frame index m and PLSC k is: Assume PSK payload have been sufficiently scrambled to have equal probability property so the expectation nearly (1/M ) N PLF −N +1 ⃗ d a k e jθ k to zero [3].The optimum decision for PLSC m and frame index k of (12b) should be like: (13) Separating the start of frame signal from the s PLF m n+k , the likelihood function in (13) of physical layer signaling detection can be rewritten as follows: Taking into account the fixed nature of the SOF pattern, it is feasible to execute the PLSC decision apart from ( 14), denoted as m before the completion of frame synchronization.This concept can be expressed mathematically as follows: The block diagram of the frame and PLSC detection is shown in Figure 6 under the conditional PLSC known assumption.In order to make the detection system more efficient, trimming the term of the equation by using the biorthogonal property of the PLSC sequence.Recall the former description, all possible signaling signals s PLSC m n form a biorthogonal signal set with the following correlation property: where N represents the length of the scramble sequence, which is fixed at 64.It is important to observe that the cross-correlation between adjacent PLSC indices exhibits a distinctive characteristic compared to autocorrelation, primarily in terms of its sign.The approach of trimming Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.
the adjacent PLSC correlation for the decision rule can be represented as follows: Furthermore, the process of (17) determining the sign decision is achieved through the representation of the index remainder i as following where m = ĵ + i computation in two parts above.

1) EFFICIENCY DESIGN FOR PLSC
As a consequence of the composition of s PLSC m n , as described in Section II-A along with the pi/2 BPSK symbol mapping and the scramble operation, which primarily involves a phase adjustment applied to s PLSC m , we can define the corrected frame synchronization received PLSC as r PLSC n .The correlation component can then be expressed as follows It is noteworthy that the correlation term can be expressed as a linear combination utilizing Walsh functions, which can be equivalently interpreted as a form of the Fast Walsh-Hadamard Transform [10].Consequently, the decision rule (18) can be redefined as follows.
After k is discovery, the remainder index i decision rule (18) becomes (21) with property form (19 PLSC MODCOD and Type field can be reconigze after behind.The figure below provides a schematic diagram of the continuous transmission of multiple frames, with PLSC indices [32 32 33 33 32 32 44 44 32 32] in sequence.Each received signal is unfolded in three-dimensional space after FWHT operation.From the figure8, it can be observed that the correct PLSC index and Time index of the maximum value occur.

2) CORRECTION TERM OF DETECTION
Massey derived the Maximum Likelihood method by adding a correction term in the correlator [11], which is given in the following equation: And its corresponding block diagram is shown in Figure 8.The magnitude computation is frequently implemented in approximate models or approached using the CORDIC algorithm.This paper adopts an approximation model by [12] and [13], which is given in the following equation In this approximation model, the root mean square (rms) deviation is -20.7 dB, while the peak deviation is −18.6 dB.

3) PERFORMANCE
The figure shows the simulation comparing executing frame synchronization by only the frame start indicator, the entire physical layer header, and the entire physical layer header with the correction term.From Figure 9, the performance of False Detected Rate does improve after adding correction terms to the detector in (20).

IV. IMPLEMENTATION
From a hardware implementation perspective, the utilization of the Fast Walsh-Hadamard Transform within the PLSC detector is pivotal.It offers an efficient approach to resource sharing by facilitating the demapping and descrambling of received symbols.The initial detection design necessitated the deployment of 128 sets of detectors to accommodate all physical layer signaling modes.However, through the implementation of the Fast Walsh-Hadamard Transform as discussed in Section III.B, a substantial reduction in register usage can be achieved.To illustrate the proposed scheme, a block diagram combining the frame and physical layer signaling detector is presented in Figure 10.

1) FPGA DESIGN UTILIZATION
The proposed method has been implemented on the Xilinx FPGA device xc7z020clg484-1 successfully.Concurrently, this paper also presents the implementation of the physical layer signaling detector without the efficiency-oriented design with PSLC partial Mode while only obtaining 32 groups due to being limited by the gate count number of FPGA.In the implementation of the Xilinx device xczu9egffvb1156-2 which carries much more gate count than xc7z020clg484-1, a detailed comparison of resource utilization between the two designs is provided in the table below.As the result, generalizing the results of 32 groups to 128 groups, the LUT utilization should require at least 80k, and the register utilization should require at least 40k.
Compared to the efficient design, the inefficient design requires approximately 80 times more LUT usage and about 4 times more register usage.

V. CONCLUSION
The DVB-S2 standard, an advancement over DVB-S, offers enhanced coding and modulation techniques, providing a range of suitable options for various scenarios.However, it also introduces dynamic changes to the Physical Layer Header, complicating the design of receivers.This article proposes reordering the MODCOD and Type Field bits to present the PLSC sequence in Dyadic order, as generated by the Walsh function.Hence, receivers can significantly reduce the resources needed for frame synchronization and jointly detect the PLSC using the Fast Walsh-Hadamard Transform.
In terms of hardware implementation, this paper demonstrates the intuitive approach with the xczu9egffvb1156-2 platform and implements the proposed architecture on the xc7z020clg484-1 platform.The proposed architecture of joint PLSC and frame deterctor greatly lowing the LUT around 80 times and register around 4 times via Fast Walsh-Hadamard Transform.
Start of Frame (SOF) is a 26-bit fixed unique word with 18D2E82(HEX) ⃗ u = (u 1 . . .u 26 ) which is attached in front of every frame.The PLSC costs 7 bits to represent 128 kinds of VCM/ACM signaling information which consist of MODCOD Field (b 1 b 2 b 3 b 4 b 5 ) and Type Field (b 6 b 7 ) and remapping (b 1 b 2 b 3 b 4 b 5 b 6
n and (1 − j)/ √ 2 for even n.(3) Separating the unscrambled PLSC sequence from the scrambled sequence results in two distinct signal terms, which can be defined as follows s PLSC m n = s p m n • s c n = e j(p m π)• e j(c n π+θ n ) .( * ) T denote transport computation.Corresponding 1 to black and 0 to white in the array and displaying as an image shown in Figure4.Observation reveals that the difference between even and odd rows is determined by the signum function.To reconstruct the array S P , even and odd rows are separated, denoted as W and −W respectively, and shown as such in Figure5(a) for even rows and Figure5(b) for odd rows.

6 b=1 2
x b and y b in (4) are binary bit of time index x and othogonal serial stamp y which shown in x = x b , x = 0, . . ., 63

FIGURE 4 .
FIGURE 4. Array of encoded matrix in gray scale.
which decided by the PLSC symbol s PLSC m n when m th signaling is transmitting.And the PL Header can express as s PLH m n = s SOF n , n = 1, . . ., 26 s PLSC m n−26 n = 27, . . ., 90 is N PLH where can divide into PL Header part with length N PLH and Payload part N PLF − N PLH .

FIGURE 9 .FIGURE 10 .
FIGURE 9. False acquisition rate of different frame synchronization methods.