Advanced Technology in ATSC3.0: Boosting Data Signal Rates Using Polarization Properties and FSS ^†

Abstract

The ATSC 3.0 standard supports ultra-high-definition broadcasting and introduces transmitter identification (TxID) to enable robust operation in single-frequency networks (SFNs). TxID typically relies on long spreading codes to minimize interference, but this approach limits the achievable data transmission rate. To address this limitation, we propose a novel scheme that combines cross-polarization discrimination (XPD) and frequency symbol spreading (FSS) to reduce interference from additional data on standard broadcast signals without increasing the system complexity. In the proposed system, vertically polarized antennas transmit standard broadcast signals, while horizontally polarized antennas transmit additional data. FSS utilizes orthogonal Hadamard codes in the frequency domain to enhance signal robustness in multipath fading environments. The simulation results demonstrate improved bit error rate (BER) and throughput under varying XPD conditions (5–15 dB), with further gains achieved through the use of longer spreading codes and adaptive modulation. The proposed method requires only minor hardware modifications and is fully compatible with existing ATSC 3.0 infrastructure.

1. Introduction

In recent years, the Advanced Television Systems Committee 3.0 (ATSC 3.0) standard has been developed and adopted for terrestrial broadcasting to support higher-quality and more flexible services. ATSC 3.0 is gaining significant attention as a next-generation technology that enables the convergence of broadcasting and telecommunications. Compared to its predecessor, ATSC 1.0, it introduces numerous technical advancements, including ultra-high-definition (UHD) video delivery, support for mobile device reception, and enhanced error correction mechanisms [1,2]. ATSC 3.0 is currently being deployed in North America, South Korea, and other regions around the world.

The physical layer of ATSC 3.0 employs advanced transmission technologies such as bit-interleaved coded modulation (BICM), orthogonal frequency division multiplexing (OFDM), cell multiplexing, and layered division multiplexing (LDM) [3]. These technologies enable more efficient and flexible utilization of frequency resources compared to previous digital television broadcasting standards. Notably, the establishment of a single frequency network (SFN) is a key feature of ATSC 3.0, contributing to improved spectrum efficiency [4]. To support wide-area coverage in SFN configurations, ATSC 3.0 employs transmitter identification (TxID) techniques to distinguish signals from multiple transmitting stations [5,6]. TxID sequences, which function as electronic watermarks, are embedded in each transmission and assign unique identifiers to each on-channel repeater (OCR). Historically, TxID schemes based on Kasami sequences have been widely used; however, their effectiveness is hindered by truncation errors and inter-code interference (ICI), leading to degraded identification accuracy [7]. To address these issues, alternative TxID sequences utilizing efficient matched filter (MF) architectures have been proposed [8]. However, these methods typically rely on extremely long spreading codes (e.g., 64,896 chips) to mitigate interference with broadcast signals, which significantly reduces the data transmission rate [9,10].

In this study, we propose a novel transmission scheme for ATSC 3.0 that leverages the polarization properties of electromagnetic waves to mitigate interference and enhance the data transmission rate. When a dipole antenna is used for signal transmission, the resulting electromagnetic waves exhibit a directional electric field—referred to as polarization—which affects the received power based on the antenna’s orientation relative to the wave’s polarization direction [11]. The maximum received power occurs when the antenna is aligned with the polarization, whereas misalignment leads to attenuation. Specifically, when using vertically and horizontally oriented antennas for transmission and reception, cross-polarization discrimination (XPD) results in a power attenuation of approximately 5–15 dB [12]. XPD, which reflects the variation in received power due to polarization orientation, is effective at reducing interference to broadcast signals. By exploiting XPD, it becomes possible to transmit additional data without significantly interfering with standard broadcast signals, even when using shorter spreading codes. This, in turn, improves the overall transmission rate. Moreover, a system based on XPD can be implemented without increasing the complexity of the existing ATSC 3.0 architecture [13,14]. Its integration requires only the addition of antennas with differing orientations, avoiding increased processing demands at both the transmitter and receiver. Additionally, frequency symbol spreading (FSS)—a technique that spreads OFDM symbols across the frequency domain—has gained attention for its ability to enhance the bit error rate (BER) performance in multipath fading environments [15]. In addition, frequency symbol spreading (FSS) is highly compatible with adaptive modulation techniques [16,17]. In adaptive modulation, the modulation scheme is dynamically adjusted based on the received signal level, while transmit power control (TPC) is used to maintain the desired signal quality. However, in OFDM and OFDMA systems, applying adaptive modulation to all subcarriers requires substantial control information, leading to increased overhead and reduced system efficiency [18,19]. FSS offers an effective solution to this challenge. The FSS process equalizes the received signal-to-noise ratio (SNR), allowing for the implementation of adaptive modulation and TPC using a single feedback value [20]. As a result, the amount of required control information is significantly reduced, while effective modulation adaptation and power control are maintained. This paper demonstrates the effectiveness of utilizing XPD in ATSC 3.0 through simulation results, and some of the content is an extended version of our previous work [21].

This paper is organized as follows. The system model is described in Section 2. In Section 3, we show the proposed system. In Section 4, we show the computer simulation results. Finally, the conclusion is given in Section 5.

2. System Model

2.1. Channel Model

We assume a multipath fading environment, modeled as follows.

$$\begin{array}{c}\hfill h(\tau ,t)=\sum _{j=0}^{J-1}{h}_j\left(t\right)\delta (\tau -{\tau }_j),\end{array}$$

(1)

$$\begin{array}{c}\hfill h_j\left(t\right)={\displaystyle \frac{g_j}{\sqrt{K}}}\sum _{k=1}^{K}exp\left[j(2\pi f_{D}tcos{\alpha }_k+{\phi }_k)\right],\end{array}$$

(2)

where $h_j$ represents the complex channel coefficient, $\delta$ denotes the Dirac’s delta function, and J and ${\tau }_j$ indicate the number of discrete paths and the delay time, respectively. On the receiver side, K waves arrive at ${\tau }_j$, with $g_j, {\alpha }_k$ and ${\phi }_k$ representing the gain of the j-th path, the angle of arrival (AoA) of the k-th wave as well as its initial phase, respectively. $f_D$ denotes the Doppler frequency. We assume normalized path gains, meaning ${\sum }_{j=0}^{J-1}E\left[\right|h_j^2\left|\right]=1$, where $E[·]$ is the expectation operation (ensemble average). Thus, the frequency response $H(f,t)$ can be derived by applying Fourier transform to the impulse response, as

$$\begin{array}{c}\hfill H(f,t)={\int }_0^{\infty }h(\tau ,t)exp(-j2\pi f\tau )d\tau ,\end{array}$$

(3)

$$\begin{array}{c}\hfill =\sum _{j=0}^{J-1}{h}_j\left(t\right)exp(-j2\pi f\tau ),\end{array}$$

(4)

where f represents the carrier frequency. In a mobile wireless communication environment, the frequency response is typically not flat. When $J\ge 1$, it results in a frequency-selective fading channel, where $\left|H\right(f,t\left)\right|$ varies across the transmission bandwidth.

2.2. XPD

Figure 1 illustrates the configuration of the transmitter and receiver equipped with vertical and horizontal polarization antennas. Figure 2 shows the incident electric field vectors arriving at the receiver, which are used to derive the analytical model of XPD in a multipath fading environment. For clarity, we assume that the two receiving antennas lie in the y–z plane. The incident electric field at the location of the receiving antennas is represented by

$$\begin{array}{cc}\mathcal{E}\hfill & ={\mathcal{E}}_v+{\mathcal{E}}_h\hfill \\ {\mathcal{E}}_v\hfill & ={\mathcal{E}}_{v,1}{\hat{d}}_{v,1}+{\mathcal{E}}_{v,2}{\hat{d}}_{v,2}\hfill \\ {\mathcal{E}}_h\hfill & ={\mathcal{E}}_{h,1}{\hat{d}}_{h,1}+{\mathcal{E}}_{h,2}{\hat{d}}_{h,2},\hfill \end{array}$$

(5)

where the unit vectors ${\hat{d}}_{v,1}, {\hat{d}}_{h,1}, {\hat{d}}_{v,2}$ and ${\hat{d}}_{h,2}$ are perpendicular to the direction of propagation. The unit vector ${\hat{d}}_{v,1}$ and ${\hat{d}}_{h,1}$ are oriented in the horizontal plane. In contrast, ${\hat{d}}_{v,2}$ and ${\hat{d}}_{h,2}$ are tilted away from the vertical axis by the elevation angles ${\gamma }_v$ and ${\gamma }_h$, as illustrated in Figure 1 and Figure 2,

$$\begin{array}{cc}\hfill {\hat{d}}_{v,1}=& -sin{\phi }_v·\hat{x}+cos{\phi }_v·\hat{y},\hfill \end{array}$$

(6)

$$\begin{array}{cc}\hfill {\hat{d}}_{h,1}=& -sin{\phi }_h·\hat{x}+cos{\phi }_h·\hat{y},\hfill \end{array}$$

(7)

$$\begin{array}{cc}\hfill {\hat{d}}_{v,2}=& -sin{\gamma }_v·cos{\phi }_v·\hat{x}-sin{\gamma }_v·sin{\phi }_v·\hat{y}\hfill \\ \hfill & +cos{\gamma }_v·\hat{z},\hfill \end{array}$$

(8)

$$\begin{array}{cc}\hfill {\hat{d}}_{h,2}=& -sin{\gamma }_h·cos{\phi }_h·\hat{x}-sin{\gamma }_h·sin{\phi }_h·\hat{y}\hfill \\ \hfill & +cos{\gamma }_h·\hat{z},\hfill \end{array}$$

(9)

The elevation angles are defined as ${\gamma }_v={\displaystyle \frac{\pi }{2}}-{\vartheta }_v$ and ${\gamma }_h={\displaystyle \frac{\pi }{2}}-{\vartheta }_h$. The unit vectors $\hat{x}, \hat{y}$ and $\hat{z}$ indicate directions along the Cartesian axes. The vectors representing the orientation of the receiving antennas are given by

$$\begin{array}{cc}\hfill & \widehat{s_1}=sin\xi ·\hat{y}+cos\xi ·\hat{z},\hfill \end{array}$$

(10)

$$\begin{array}{cc}\hfill & \widehat{s_2}=-sin\xi ·\hat{y}+cos\xi ·\hat{z}.\hfill \end{array}$$

(11)

Given that the incident electric field follows a Rayleigh distribution, let

$$\begin{array}{c}\hfill {\mathcal{E}}_{v,1}={\mathcal{A}}_{v,1}cos(\omega t+{\alpha }_{v,1}),\end{array}$$

(12)

$$\begin{array}{c}\hfill {\mathcal{E}}_{h,1}={\mathcal{A}}_{h,1}cos(\omega t+{\alpha }_{h,1}),\end{array}$$

(13)

$$\begin{array}{c}\hfill {\mathcal{E}}_{v,2}={\mathcal{A}}_{v,2}cos(\omega t+{\alpha }_{v,2}),\end{array}$$

(14)

$$\begin{array}{c}\hfill {\mathcal{E}}_{h,2}={\mathcal{A}}_{h,2}cos(\omega t+{\alpha }_{h,2}),\end{array}$$

(15)

where the random variables ${\mathcal{A}}_{v,1}, {\mathcal{A}}_{v,2}, {\mathcal{A}}_{h,1}$ and ${\mathcal{A}}_{h,2}$ are mutually independent and follow a Rayleigh distribution. The phase variables ${\alpha }_{v,1}, {\alpha }_{h,1}, {\alpha }_{v,2}$ and ${\alpha }_{h,2}$ are independent and identically distributed (i.i.d.) according to a uniform distribution. Based on Equations (5)–(15), the signal received at antenna $R_1$ is proportional to

$$\begin{array}{cc}\hfill R_1=& R_{v,1}+R_{h,1}\hfill \\ \hfill R_{v,1}=& {\mathcal{E}}_v·{\hat{s}}_1={\mathcal{E}}_{v,1}{\hat{d}}_{v,1}·{\hat{s}}_1+{\mathcal{E}}_{v,2}{\hat{d}}_{v,2}·{\hat{s}}_1\hfill \\ \hfill =& {\mathcal{E}}_{v,1}cos{\phi }_{v}sin\xi +{\mathcal{E}}_{v,2}cos{\gamma }_{v}cos\xi \hfill \\ \hfill & -{\mathcal{E}}_{v,2}sin{\gamma }_{v}sin{\phi }_{v}sin\xi \hfill \\ \hfill =& {\mathcal{A}}_{v,1}cos{\phi }_{v}sin\xi cos(\omega t+{\alpha }_{v,1})\hfill \\ \hfill & +{\mathcal{A}}_{v,2}[cos{\gamma }_{v}cos\xi -sin{\gamma }_{v}sin{\phi }_{v}sin\xi ]\hfill \\ \hfill & ·cos(\omega t+{\alpha }_{v,2})\hfill \\ \hfill =& ({\mathcal{A}}_{v,1}{\mu }_{v}cos{\alpha }_{v,1}+{\mathcal{A}}_{v,2}{\nu }_{v}cos{\alpha }_{v,2})cos\omega t\hfill \\ \hfill & -({\mathcal{A}}_{v,1}{\mu }_{v}sin{\alpha }_{v,1}+{\mathcal{A}}_{v,2}{\nu }_{v}sin{\alpha }_{v,2})sin\omega t\hfill \\ \hfill R_{h,1}=& {\mathcal{E}}_h·{\hat{s}}_1={\mathcal{E}}_{h,1}{\hat{d}}_{h,1}·{\hat{s}}_1+{\mathcal{E}}_{h,2}{\hat{d}}_{h,2}·{\hat{s}}_1\hfill \\ \hfill =& {\mathcal{E}}_{h,1}cos{\phi }_{h}sin\xi +{\mathcal{E}}_{h,2}cos{\gamma }_{h}cos\xi \hfill \\ \hfill & -{\mathcal{E}}_{h,2}sin{\gamma }_{h}sin{\phi }_{h}sin\xi \hfill \\ \hfill =& {\mathcal{A}}_{h,1}cos{\phi }_{h}sin\xi cos(\omega t+{\alpha }_{h,1})\hfill \\ \hfill & +{\mathcal{A}}_{h,2}[cos{\gamma }_{h}cos\xi -sin{\gamma }_{h}sin{\phi }_{h}sin\xi ]\hfill \\ \hfill & ·cos(\omega t+{\alpha }_{h,2})\hfill \\ \hfill =& ({\mathcal{A}}_{h,1}{\mu }_{v}cos{\alpha }_{h,1}+{\mathcal{A}}_{h,2}{\nu }_{h}cos{\alpha }_{h,2})cos\omega t\hfill \\ \hfill & -({\mathcal{A}}_{h,1}{\mu }_{v}sin{\alpha }_{h,1}+{\mathcal{A}}_{h,2}{\nu }_{h}sin{\alpha }_{h,2})sin\omega t,\hfill \end{array}$$

(16)

where ${\mu }_v=sin\xi cos{\phi }_v, {\mu }_h=sin\xi cos{\phi }_h, {\nu }_v=cos{\gamma }_{v}cos\xi -sin{\gamma }_{v}sin{\phi }_{v}sin\xi$ and ${\nu }_h=cos{\gamma }_{h}cos\xi -sin{\gamma }_{h}sin{\phi }_{h}sin\xi$. Similarly, the signal strength received by antenna $R_2$ is directly proportional to the corresponding component of the transmitted signal, as

$$\begin{array}{cc}\hfill R_2=& R_{v,2}+R_{h,2}\hfill \\ \hfill R_{v,2}=& {\mathcal{E}}_v·{\hat{s}}_2={\mathcal{E}}_{v,1}{\hat{d}}_{v,1}·{\hat{s}}_2+{\mathcal{E}}_{v,2}{\hat{d}}_{v,2}·{\hat{s}}_2\hfill \\ \hfill =& -{\mathcal{E}}_{v,1}cos{\phi }_{v}sin\xi +{\mathcal{E}}_{v,2}cos{\gamma }_{v}cos\xi \hfill \\ \hfill & +{\mathcal{E}}_{v,2}sin{\gamma }_{v}sin{\phi }_{v}sin\xi \hfill \\ \hfill =& -{\mathcal{A}}_{v,1}cos{\phi }_{v}sin\xi cos(\omega t+{\alpha }_{v,1})\hfill \\ \hfill & +{\mathcal{A}}_{v,2}[cos{\gamma }_{v}cos\xi +sin{\gamma }_{v}sin{\phi }_{v}sin\xi ]\hfill \\ \hfill & ·cos(\omega t+{\alpha }_{v,2})\hfill \\ \hfill =& (-{\mathcal{A}}_{v,1}{\mu }_{v}cos{\alpha }_{v,1}+{\mathcal{A}}_{v,2}{\hat{\nu }}_{v}cos{\alpha }_{v,2})cos\omega t\hfill \\ \hfill & -(-{\mathcal{A}}_{v,1}{\mu }_{v}sin{\alpha }_{v,1}+{\mathcal{A}}_{v,2}{\hat{\nu }}_{v}sin{\alpha }_{v,2})sin\omega t\hfill \\ \hfill R_{h,2}=& {\mathcal{E}}_h·{\hat{s}}_2={\mathcal{E}}_{h,1}{\hat{d}}_{h,1}·{\hat{s}}_2+{\mathcal{E}}_{h,2}{\hat{d}}_{h,1}·{\hat{s}}_2\hfill \\ \hfill =& -{\mathcal{E}}_{h,1}cos{\phi }_{h}sin\xi +{\mathcal{E}}_{h,2}cos{\gamma }_{h}cos\xi \hfill \\ \hfill & +{\mathcal{E}}_{h,2}sin{\gamma }_{h}sin{\phi }_{h}sin\xi \hfill \\ \hfill =& -{\mathcal{A}}_{h,1}cos{\phi }_{h}sin\xi cos(\omega t+{\alpha }_{h,1})\hfill \\ \hfill & +{\mathcal{A}}_{h,2}[cos{\gamma }_{h}cos\xi +sin{\gamma }_{h}sin{\phi }_{h}sin\xi ]\hfill \\ \hfill & ·cos(\omega t+{\alpha }_{h,2})\hfill \\ \hfill =& (-{\mathcal{A}}_{h,1}{\mu }_{h}cos{\alpha }_{h,1}+{\mathcal{A}}_{h,2}{\hat{\nu }}_{h}cos{\alpha }_{h,2})cos\omega t\hfill \\ \hfill & -(-{\mathcal{A}}_{h,1}{\mu }_{h}sin{\alpha }_{h,1}+{\mathcal{A}}_{h,2}{\hat{\nu }}_{h}sin{\alpha }_{h,2})sin\omega t,\hfill \end{array}$$

(17)

where ${\hat{\nu }}_v=cos{\gamma }_{v}cos\xi +sin{\gamma }_{v}sin{\phi }_{v}sin\xi , {\hat{\nu }}_h=cos{\gamma }_{h}cos\xi +sin{\gamma }_{h}sin{\phi }_{h}sin\xi$. The amplitudes of the received vertical polarization component signals by $R_1$ and $R_2$ are, therefore, proportional to

$$\begin{array}{cc}\hfill {\Lambda }_{v,1}=& [{({\mathcal{A}}_{v,1}{\mu }_{v}cos{\alpha }_{v,1}+{\mathcal{A}}_{v,2}{\nu }_{v}cos{\alpha }_{v,2})}^2\hfill \\ \hfill & +{({\mathcal{A}}_{v,1}{\mu }_{v}sin{\alpha }_{v,1}+{\mathcal{A}}_{v,2}{\nu }_{v}sin{\alpha }_{v,2})}^2{]}^{1/2}\hfill \\ \hfill =& [{\mathcal{A}}_{v,1}^2{\mu }_v^2+{\mathcal{A}}_{v,2}^2{\nu }_v^2+2{\mathcal{A}}_{v,1}{\mathcal{A}}_{v,2}{\mu }_v{\nu }_v\hfill \\ \hfill & ·cos({\alpha }_{v,1}-{\alpha }_{v,2}){]}^{1/2}\hfill \\ \hfill {\Lambda }_{v,2}=& [{\mathcal{A}}_{v,1}^2{\mu }_v^2+{\mathcal{A}}_{v,2}^2{\hat{\nu }}_v^2-2{\mathcal{A}}_{v,1}{\mathcal{A}}_{v,2}{\mu }_v{\hat{\nu }}_v\hfill \\ \hfill & ·cos({\alpha }_{v,1}-{\alpha }_{v,2}){]}^{1/2}\hfill \end{array}$$

(18)

According to Equation (18), the mean received signal strength at the receiving antennas is lower than that received by a vertically polarized antenna. The mean signal received by the vertically polarized antenna is expressed as $\langle {\Lambda }_{v,0}^2\rangle =\langle {\mathcal{A}}_{v,2}^2\rangle cos{\gamma }_v$, while the mean signal level received by the polarization diversity antenna $R_1$ is $\langle {\Lambda }_{v,1}^2\rangle$. The ratio between these two signal levels is, therefore, given by:

$$\begin{array}{cc}\hfill L_{v,1}=& {\displaystyle \frac{\langle {\Lambda }_{v,1}^2\rangle }{\langle {\Lambda }_{v,0}^2\rangle }}={\displaystyle \frac{\langle {\mathcal{A}}_{v,1}^2{\mu }_v^2+{\mathcal{A}}_{v,2}^2{\nu }_v^2\rangle }{\langle {\mathcal{A}}_{v,2}^2\rangle cos{\gamma }_v}}\hfill \\ \hfill =& {\displaystyle \frac{\langle {\mathcal{A}}_{v,1}^2\rangle {\mu }_v^2}{\langle {\mathcal{A}}_{v,2}^2\rangle cos{\gamma }_v}}+{\displaystyle \frac{{\nu }_v^2}{cos{\gamma }_v}}\hfill \\ \hfill =& {\displaystyle \frac{{\mu }_v^2}{cos{\gamma }_v{X}_v}}+{\displaystyle \frac{{\nu }_v^2}{cos{\gamma }_v}}.\hfill \end{array}$$

(19)

Likewise, the mean received signal strength at the polarization diversity antenna $R_2$ is $\langle {\Lambda }_{v,2}^2\rangle$. Therefore,

$$\begin{array}{cc}\hfill L_{v,2}=& {\displaystyle \frac{\langle {\Lambda }_{v,2}^2\rangle }{\langle {\Lambda }_{v,0}^2\rangle }}={\displaystyle \frac{\langle {\mathcal{A}}_{v,1}^2{\mu }_v^2+{\mathcal{A}}_{v,2}^2{\hat{\nu }}_v^2\rangle }{\langle {\mathcal{A}}_{v,2}^2\rangle cos{\gamma }_v}}\hfill \\ \hfill =& {\displaystyle \frac{\langle {\mathcal{A}}_{v,1}^2\rangle {\mu }_v^2}{\langle {\mathcal{A}}_{v,2}^2\rangle cos{\gamma }_v}}+{\displaystyle \frac{{\hat{\nu }}_v^2}{cos{\gamma }_v}}\hfill \\ \hfill =& {\displaystyle \frac{{\mu }_v^2}{cos{\gamma }_v{X}_v}}+{\displaystyle \frac{{\hat{\nu }}_v^2}{cos{\gamma }_v}}.\hfill \end{array}$$

(20)

As shown in Figure 3, the simulated XPD values indicate the degree of signal attenuation experienced by the receiver as a function of polarization mismatch, specifically for $\xi$ = 45°. This validates the effectiveness of polarization isolation in suppressing cross-polarized interference. Referring to Figure 3, the received signal in one branch is degraded (between −4 dB and −23 dB for $R_1$) relative to that received by a vertical antenna. In horizontal and vertical polarization antennas, the value of the cross-polarization discrimination (XPD) value vary between 5–15 dB depending on the environment [12].

3. Proposed System

In this section, we present an OFDM-based transmission scheme that utilizes cross-polarization discrimination (XPD).

Figure 4 provides an overview of the proposed system architecture. In particular, Figure 4a shows the architecture of the proposed transmitter. In this design, data streams are first converted from serial to parallel (S/P) and then modulated using QPSK. The broadcasting data are transmitted using the vertical antenna, while additional data are transmitted via the horizontal antenna. To generate the OFDM signal for the additional data, frequency symbol spreading (FSS) is applied before IFFT, as illustrated in Figure 5. This technique uses Hadamard codes to spread symbols across subcarriers, providing frequency diversity and improving robustness in multipath fading environments.

$u_{dd}(m,i)$ represent the i-th symbol at the m-th subcarrier following FSS, and it is given by,

$$\begin{array}{cc}\hfill u_{dd}(m,i)=\sum _{k=0}^{N_{SF}-1}& c_k(mmod{N}_{SF})\hfill \\ \hfill & ·d_{dd}(\lfloor m/N_{SF}\rfloor N_{SF}+k,i),\hfill \end{array}$$

(21)

where $N_{SF}$ represents the length of the spreading code. The term $d_{dd}(m,i)$ refers to the additional data associated with the i-th symbol on the m-th subcarrier, and is normalized, such that $E\left[\right|d_{dd}(m,i)\left|\right]=1$. The notation $\lfloor a\rfloor$ indicates the greatest integer less than or equal to a. The function $c_k\left(n\right)$ denotes the Hadamard code used for FSS, which satisfies the orthogonality condition:

$$\begin{array}{c}\hfill \sum _{k=0}^{N_{SF}-1}{c}_k\left(n\right)c_{\Omega }^{\ast }\left(n\right)=\left\{\begin{array}{cc}{N}_{SF}\hfill & (k=\Omega )\hfill \\ 0\hfill & (k\ne \Omega ),\hfill \end{array}\right.\end{array}$$

(22)

where $|c_k\left(n\right)|=1$ and * denotes the complex conjugate.

After that, the OFDM time signal is generated by IFFT and the guard interval (GI) is inserted. The transmit antennas consist of two antennas (one vertical and one horizontal). $s_v$ and $s_h$ indicate the signals transmitted from the base station vertically and horizontally, respectively. They are expressed as follows.

$$\begin{array}{cc}\hfill s_v\left(t\right)=& \sum _{i=0}^{N_p+N_d-1}g(t-iT)\sqrt{{\displaystyle \frac{2S}{N_c}}}\hfill \\ \hfill & ·[\sum _{i=0}^{N_c-1}{d}_{bc}(m,i)expj2\pi (t-iT)m/T_s],\hfill \end{array}$$

(23)

$$\begin{array}{cc}\hfill s_h\left(t\right)=& \sum _{i=0}^{N_p+N_d-1}g(t-iT)\sqrt{{\displaystyle \frac{2S}{N_c}}}\hfill \\ \hfill & ·[\sum _{i=0}^{N_c-1}{u}_{dd}(m,i)expj2\pi (t-iT)m/T_s],\hfill \end{array}$$

(24)

where $N_p$, $N_d$, and $N_c$ denote the number of pilot symbols, data symbols, and subcarriers, respectively. The term $d_{bc}(m,i)$ represents the broadcast signal of the i-th symbol on the m-th subcarrier. T denotes the total duration of an OFDM symbol, including the GI of duration $T_g$, while $T_s$ refers to the effective symbol duration (excluding the GI) systems, the GI is inserted to mitigate inter-symbol interference (ISI) caused by multipath fading. S represents the average transmit power, and $g\left(t\right)$ is the transmission pulse, which is defined as follows:

$$\begin{array}{c}\hfill g\left(t\right)=\left\{\begin{array}{cc}1,\hfill & -T_g\le t\le T_s\hfill \\ 0,\hfill & \mathrm{otherwise}.\hfill \end{array}\right.\end{array}$$

(25)

Figure 6 illustrates the configuration of cross-polarized antenna pairs used at both the transmitter and receiver. In this setup, vertical and horizontal polarization components are transmitted and received independently, enabling the system to exploit the benefits of cross-polarization discrimination (XPD). The receiver employs two antennas: one aligned for vertical polarization and the other for horizontal polarization. The received signals at these antennas, denoted as $r_v$ and $r_h$, are mathematically expressed, as follows:

$$\begin{array}{cc}\hfill r_v& =h_{vv}{s}_v+h_{hv}{s}_h+n_1\hfill \\ \hfill & =h_{vv}{s}_v+\widetilde{n_1},\hfill \end{array}$$

(26)

$$\begin{array}{cc}\hfill r_h& =h_{vh}{s}_v+h_{hh}{s}_h+n_2\hfill \\ \hfill & =h_{hh}{s}_h+\widetilde{n_2},\hfill \end{array}$$

(27)

where $n_1, n_2$ is the Additive White Gaussian Noise (AWGN), and $h_{vh}$, $h_{hv}$ are the channel response between the different polarization antennas. $\widetilde{n_1}$, $\widetilde{n_2}$ considers the signals from different polarization antennas as residual noise, taking into account the attenuation due to XPD. Therefore, by reducing only the received power from different polarization antenna, it is possible to receive the desired data by switching between vertical and horizontal receiving antennas.

As shown in Figure 4b, the receiver removes the GI, performs S/P conversion, and applies fast Fourier Transform (FFT). Lastly, this frequency domain signal is detected and demodulated.

When receiving additional data, the converted signals are despread by spreading code in the frequency domain. When despreading additional data, spread gain is obtained. Therefore, additional data can be received with the same performance as broadcast. However, multipath fading disrupts the orthogonality among spread symbols. To restore it, we utilize a minimum mean square error combining MMSEC-based frequency equalization for detection. The MMSEC weight is defined as follows:

$$\begin{array}{c}\hfill {\Omega }_{\mathrm{MMSEC}\left(m\right)}={\displaystyle \frac{h^{\ast }\left(m\right)}{{\left|h\left(m\right)\right|}^2+2{\sigma }^2}},\end{array}$$

(28)

where $h\left(m\right)$ denotes the channel estimate of the m-th OFDM subcarrier, and $\sigma$ represents the Additive White Gaussian Noise (AWGN) and residual noise per subcarrier [16].

Figure 5 illustrates the overall mechanism of FSS. Within the FSS block, each subcarrier symbol is spread in the frequency domain using an orthogonal spreading code. The FSS block refers to a group of subcarriers to which the spreading code is applied, and the resulting spread data symbols are combined across these subcarriers. At the receiver side, the original data can be retrieved through a despreading operation. This helps mitigate the effects of frequency-selective fading by effectively distributing its influence across the entire FSS block [15,17]. This approach allows FSS to provide frequency diversity, thereby suppressing the impact of frequency-selective fading.

The conventional TxID spreading codes often use Gold codes with spreading code lengths of $2^n-1(n=1,2,3\dots )$. However, in this case, since spread is performed in the frequency domain by FSS, FFT must be performed after spreading. Considering that the FFT size must be a power of 2. Therefore, Hadamard codes with spreading code lengths of $2^m(m=0,1,2,3\dots )$ are used in this paper.

4. Computer Simulation Result

We conducted computer simulations to evaluate the proposed method described in the previous section. The simulation parameters are shown in

Table 1. Simulation parameters.

	Broadcasting	Data Communication
Modulation Scheme	QPSK	QPSK
Number of Subcarriers	64	16
Guard Interval	16	-
Spreading Code	-	Hadamard code
Spreading Length	-	4, 16
Data Length	20	20
Pilot Signals	2	2
Channel Model	5 path Rayleigh fading	5 path Rayleigh fading
Transmission Rate	20 Mbps	20 Mbps
Doppler frequency	10 Hz	10 Hz

. In these simulations, we assumed a 5-path Rayleigh fading channel, which results in random phase variations and fluctuations in received signal strength for each multipath component. The Doppler shift was set to 10 Hz to reflect a typical mobile environment.

The only broadcasting baseline refers to the conventional ATSC 3.0 system, where the interference from TxID signals is assumed to be negligible, essentially at the noise level. Therefore, in this paper, we model this baseline by considering data transmission exclusively through the vertically polarized antenna, with interference from the horizontally polarized antenna assumed to be zero. This approach reflects the ideal condition where the impact of cross-polarization interference is negligible.

For the proposed system, we evaluated the performance under different cross-polarization discrimination (XPD) values, which varied between 5 and 15 dB depending on the environmental conditions. Simulations were performed with XPD values of 5, 10, and 15 dB, using a spreading code length of 4. Subsequently, we fixed the XPD at 15 dB and compared the FSS (frequency symbol spreading) spreading code lengths of 4 and 16 to evaluate the impact of spreading on system performance.

For forward error correction (FEC), convolutional coding with a rate of 1/2 and constraint length of 7 was employed, along with Viterbi decoding. This combination is well-suited for mitigating bit errors in frequency-selective fading channels. Channel estimation was assumed to be ideal, and perfect synchronization was considered in both the time and frequency domains, ensuring accurate demodulation and decoding throughout the simulation process.

Figure 7 illustrates the bit error rate (BER) performance of the proposed method compared to only broadcasting. Assuming XPD of 5 dB, the BER performance significantly decreased compared to only broadcasting. However, as XPD increased to 10 dB and 15 dB, the BER performance approached that of broadcasting only. At XPD of 10 dB, we can achieve the BER of ${10}^{-2}$ with Eb/No of 30 dB, and it is expected that further BER improvement by implementing FEC.

Figure 8 shows the throughput of the proposed method compared to only broadcasting. When XPD is 5 dB, more information is transmitted with SNR between 10–25 dB compared to the conventional method. In addition, when XPD is 10–15 dB, an additional throughput improvement of about 15% can be achieved.

Next, we will demonstrate the frequency diversity of FSS. Figure 9 shows the simulation results when FSS spread length of the additional data is varied from 4 to 16. In this simulation, we assume XPD is 15 dB.

Observing Figure 9, with increasing the spread length, the frequency diversity can be achieved significantly. In particular for BER of ${10}^{-3}$, we can achieve about 7 dB gain. This is because the impact of frequency selective fading could be distributed over more subcarriers by increasing the spreading code length.

Finally, to demonstrate the effectiveness of adaptive modulation (AM), we evaluated the system performance across different modulation schemes with the XPD value fixed at 15 dB. The results, shown in Figure 10, present the total throughput for data transmitted from both the vertical and horizontal antennas as the modulation scheme transitions from QPSK to 16QAM. The modulation switching based on SNR can be summarized as follows:

Table 2. Modulation switching table for adaptive modulation.

Modulation	QPSK	16QAM
SNR [dB]	$0\le \mathrm{SNR}<24$	$24\le \mathrm{SNR}$

clearly indicates that QPSK is employed for SNR values below 24 dB, while 16QAM is used for SNR values of 24 dB or higher, reflecting the trade-off between data rate and signal robustness. These findings suggest that integrating SNR-based feedback into the system and adopting adaptive modulation with QPSK and 16QAM can significantly improve the overall performance, particularly in challenging channel conditions.

5. Conclusions

This paper proposed a novel method to improve the transmission rate in data communication by utilizing polarization properties to suppress interference without affecting broadcast waves. Simulation results show that the proposed method not only enhances the transmission rate in data communication, but also achieves improved BER within the tolerance range. Next, we demonstrated that the BER performance can be further improved by employing FSS. Finally, the effectiveness of adaptive modulation in improving the system performance was clarified: the XPD value was fixed at 15 dB and the system throughput was evaluated for various modulation schemes, and the results showed that QPSK is used when SNR is less than 24 dB, and 16QAM is employed when SNR is greater than 24 dB. The throughput was confirmed to be improved using QPSK when the SNR is less than 24 dB. In addition, the use of FSS as the spreading code is expected to significantly reduce the amount of control information in adaptive modulation because the received SNR is more uniform.