A p-noma empowered pass-d2d communication system design method

Abstract

This application provides a P-NOMA-enabled design method for a PASS-D2D communication system, relating to the field of communication system technology. This application extends PASS to a PASS-D2D cooperative framework, designs a two-timeslot transmission protocol, and introduces MRC and LC merging strategies respectively. The closed-form solution for the ergodic rate of each scheme is derived, and the accuracy of the analytical results is verified through simulation. D2D cooperation can effectively alleviate the bottleneck of weak users and improve system throughput. The combination of P-NOMA and cooperative merging can further release gains, with P-LC typically performing better in interference-sensitive areas. Power allocation, user spacing, and overlap ratio jointly determine the upper limit of cooperative gains and the interference-dominant range. Reasonably controlling the near-end overlap ratio helps avoid rate degradation caused by excessive interference coupling. In summary, this application provides an interpretable theoretical basis and numerical support for the parameter design and deployment optimization of PASS and its cooperative extension systems.

Description

Technical Field

[0001] This application relates to the field of communication system design technology, and in particular to a P-NOMA-enabled PASS-D2D communication system design method. Background Technology

[0002] With ultra-large-scale antenna arrays (ELAA), high-frequency communication, and near-field propagation characteristics becoming crucial components of future wireless networks, electromagnetic waves no longer conform to the traditional far-field plane wave assumption. Their propagation is significantly influenced by spherical waves, projected aperture, and antenna spatial position, resulting in stronger spatial non-uniformity and user correlation in channel gain. Against this backdrop, achieving flexible multi-user access and reliable communication while maintaining high spectral efficiency has become a key challenge in near-field system design. Pinching-Antenna Systems (PASS) reshape the electromagnetic energy propagation path and radiation aperture by introducing a flexible and controllable antenna structure in space, providing new means for signal focusing, interference suppression, and link enhancement in near-field communication. Compared to traditional array gain methods, PASS can achieve more directional energy distribution with limited hardware complexity, constructing robust line-of-sight links for users. However, relying solely on spatial domain manipulation is insufficient to fully unleash the system's potential in multi-user access and spectrum reuse.

[0003] Meanwhile, device-to-device (D2D) communication offers significant advantages in shortening transmission distances, improving spectrum reuse efficiency, and enhancing edge user performance by allowing terminals to directly establish communication links under the coordination of cellular networks. In near-field environments, D2D links typically have shorter propagation distances and more stable channel conditions, providing ideal support for cooperative communication and information forwarding. However, D2D cooperative performance is highly dependent on the decodeability of information in the forward transmission phase; once the link from the source node to the cooperative user is restricted, its gain will be significantly weakened. To further improve the spectral efficiency of multi-user systems, NOMA actively superimposes multi-user signals in the power domain and uses successive interference cancellation (SIC) for decoding at the receiver, achieving multi-user multiplexing under the same time-frequency resources. Introducing NOMA into PASS-D2D systems is expected to simultaneously utilize spatial focusing gain, power domain discrimination capability, and cooperative diversity advantages, thereby significantly improving system capacity. However, in near-field scenarios, channel differences between users are strongly influenced by spatial location. Summary of the Invention

[0004] The purpose of this application is to solve the technical problem in the prior art where the channel difference between users in near-field scenarios is strongly affected by spatial location.

[0005] To address the aforementioned technical problems, this application provides the following technical solution:

[0006] A P-NOMA-enabled PASS-D2D communication system design method includes the following steps:

[0007] A clamp-on antenna PASS with an activatable dielectric waveguide is deployed, wherein the dielectric waveguide of the PASS is deployed at a height of [missing information]. Location, and allow along The axis flexibly activates the programmable amplifier PA, and the position of PA is set to... , The position is represented as ,in Two users are randomly distributed on a side with a length of Within the rectangular area, the centers of the rectangles are denoted as . and ;

[0008] In the first time slot, the base station (BS) serves two users using the Partial Non-Orthogonal Multiple Access (P-NOMA) principle. and Two user signals partially overlap in transmission resources, making They represent the allocations to and Based on the overlap ratio, the entire resource is divided into three regions: non-overlapping regions. Service Only Non-overlapping regions Service Only Overlapping areas Serving two users simultaneously;

[0009] In the second time slot, the cooperating users decode and forward the demand signal and the D2D signal through the device-to-device (D2D) link to achieve D2D cooperative forwarding. The cooperating nodes generate the forwarding signal using a power domain superposition method. ,in , This indicates the demand signal being forwarded. Indicates a D2D signal;

[0010] The receiver uses maximum ratio combining (MRC) or linear combining (LC) to superimpose observations from two time slots to obtain diversity gain.

[0011] Preferably, two collaboration modes are also included:

[0012] Pattern Sch1: by Support In the first time slot, R2 decoding in the overlapping region The signal is decoded at its own layer on R1∪R2; in the second time slot, As a decode-forward DF relay forwarder The message;

[0013] Pattern Sch2: by Support In the first time slot, Additional decoding The signal layer on the overlapping region R2; in the second time slot, As a DF relay forwarding The news.

[0014] Preferably, the MRC merging strategy includes:

[0015] In the MRC-Sch1 scheme, the system and rate are...

[0016] In the MRC-Sch2 scheme, the system and rate are...

[0017]

[0018] in, Indicates the subcarrier range of the corresponding region. Let SNR be the SNR of the D2D signal in the second time slot, where... This reflects the constraints of DF cooperation at the relay point.

[0019] Preferably, the LC merging strategy includes:

[0020] In the LC-Sch1 scheme, the signal transmitted in the second time slot is ,in , ;

[0021] In the LC-Sch2 scheme, the signal transmitted in the second time slot is ,in and .

[0022] Preferably, in the LC-Sch1 scheme, to offset the first time slot exist The interlayer interference generated at this location requires a cancellation factor designed to be...

[0023] ,

[0024] and take And obtained from the remaining power ;

[0025] In the LC-Sch2 scheme, to reduce interference, [the following is taken]: The power factor required for component calculation is

[0026]

[0027] and take , .

[0028] Preferably, in the LC-Sch1 scheme, after amplitude matching and phase alignment, the total system rate is:

[0029]

[0030] In the LC-Sch2 scheme, the total system speed is

[0031] .

[0032] Preferably, in the P-NOVA, the user in the first time slot SNR in different resource regions

[0033]

[0034] user The SNR in the first time slot is

[0035] .

[0036] Preferably, in the PASS, PA is sent to the user. The channel coefficient is expressed as

[0037] ,

[0038] The LoS channel of the inter-user line-of-sight link is represented as follows:

[0039] ,

[0040] in, Let be the free space loss constant. At the speed of light, For carrier frequency, For PA to user Free space propagation distance, This represents the waveguide propagation length from the feed point to the PA. The absorption coefficient is... Let the waveguide internal phase be from the feed point to PA. For the guided wave wavelength, is the effective refractive index of the dielectric waveguide.

[0041] Preferably, in the P-NOVA, the first time slot is superimposed and coded by the base station for the two users. ,in The power allocation factor for the near-end user layer. For the power allocation factor of the remote user layer, For users The expected information symbol and satisfy The total transmission power of the base station is The noise variance is The transmit signal-to-noise ratio can be defined as .

[0042] Preferably, the DF cooperation must satisfy the relay decoding constraint in the overlapping region R2. right The decoding constraints of the layer are

[0043] ,

[0044] right The decoding constraints of the layer are

[0045] .

[0046] Compared with the prior art, this application has the following beneficial effects:

[0047] (1) Under the condition of near-field spherical wave propagation, the PASS system structure, multiple access mechanism and D2D cooperative communication are modeled, and the applicability of OMA, NOMA and P-NOMA in PASS and PASS-D2D scenarios is analyzed. The advantages of P-NOMA in PASS-D2D communication are revealed, laying the foundation for subsequent performance analysis and system design.

[0048] (2) Based on the hybrid model of line-of-sight link and non-line-of-sight link, the achievable rate expression of PASS system is first derived, and the intrinsic relationship between the optimal clamping antenna position, aperture projection and propagation distance is explored to reveal the mechanism of the influence of antenna spatial deployment on system performance.

[0049] (3) Combining the partial overlap characteristics of P-NOMA with the D2D cooperation mechanism, two decoding schemes, MRC and LC, are constructed. The impact of overlap ratio, power allocation and channel link conditions on system performance is analyzed. The performance advantages of the proposed scheme in the PASS-D2D communication system are verified by theoretical analysis and simulation comparison. Attached Figure Description

[0050] Figure 1 It is a PASS-D2D communication model;

[0051] Figure 2 The relationship between traversal and rate of return (SNR);

[0052] Figure 3 The relationship between traversal and rate and the position of the clamping antenna;

[0053] Figure 4 The relationship between traversal, rate, power allocation, and overlap ratio;

[0054] Figure 5 The relationship between traversal and rate of motion and transmit power;

[0055] Figure 6 The relationship between traversal and rate, power allocation, and user spacing;

[0056] Figure 7 Heatmaps of traversal and rate under different overlap ratios. Detailed Implementation

[0057] The above content will be explained in conjunction with specific verification experiments:

[0058] I. P-NOMA-enabled PASS-D2D communication system model

[0059] 1.1 NOMA Enables PASS-D2D Communication System

[0060] First, consider a communication scenario where the downlink PASS provides services to two single-antenna users, one of whom is denoted as [user name missing]. Another user recorded as The system employs two time slots for information transmission and assumes all nodes operate in half-duplex mode to reduce self-interference and system implementation complexity. PASS's dielectric waveguides are deployed at high altitudes. Location, and allow along The axis flexibly activates PA. Let the position of PA be... ,user The position is represented as ,in Two users are randomly distributed on a side with a length of [missing information]. Within the rectangular area, the centers of the rectangles are denoted as . and ,like Figure 1 As shown. Therefore, PA to user The free space propagation distance can be expressed as

[0061]

[0062] In particular, due to the physical properties of the signal within the dielectric waveguide, a guided wave wavelength is introduced here. ,in For free space wavelengths, Let be the effective refractive index of the dielectric waveguide. Assume the coordinates of the feed point are... Then the waveguide phase from the feed point to PA can be written as: At the same time, an exponential factor is used. This represents the power attenuation of the waveguide along its length, where This represents the waveguide propagation length from the feed point to the PA. Parameter The absorption coefficient characterizes the degree of energy attenuation caused by dielectric loss, thus reflecting the waveguide material's ability to absorb signal energy. Under the above modeling assumptions, the PA to user... The channel coefficient can be expressed as

[0063]

[0064] in Let be the free space loss constant. At the speed of light, The carrier frequency is [value]. Furthermore, in the cooperative forwarding of the second time slot, a line-of-sight (LoS) link exists between users, and its channel can be represented as [expression].

[0065]

[0066] in This indicates the spatial distance between two users.

[0067] In the NOMA scheme, the first time slot is used by the base station to superimpose and encode the signals of the two users, and the transmitted signal is... ,in The power allocation factor for the near-end user layer. For the power allocation factor of the remote user layer, For users The expected information symbol and satisfy .

[0068] Assume the total transmission power of the base station is The noise variance is The transmit signal-to-noise ratio can be defined as Therefore, users The received signal in the first time slot can be expressed as:

[0069]

[0070] in It is additive white Gaussian noise, and Following the traditional NOMA decoding order, Near-end users typically enjoy better channel quality, requiring earlier decoding. To eliminate the main interference to itself, then decode after SIC. , Because the received signal power is weak, it usually does not have the capability to... The conditions for performing reliable SIC are therefore... Treat it as interference and decode it directly. Therefore, the SINR expression for the first time slot can be obtained as follows:

[0071]

[0072]

[0073] When a user in the first time slot successfully decodes information from another layer, that user can act as a DF relay for short-distance forwarding on the inter-user LoS link, thereby providing additional spatial diversity and link budget compensation for the target layer. In this system, due to the flexibility of the PASS system's PA, it is necessary to discuss this separately. Support and Support Two modes were explored to determine the optimal cooperative transmission method:

[0074] (1) MRC assisted Cooperative forwarding strategy (MRC-Sch1): Under typical NOMA settings, Due to its proximity to the radiation region and high channel gain, it is often possible to successfully decode the far-end layer in the first time slot. Therefore, it has the ability to The ability to perform DF forwarding. Let... The forwarding power in the second time slot is and define In order to carry additional transmission services while collaborating, Demand signal in the second time slot With D2D signal Power superposition is performed at this time. The received signal is

[0075]

[0076] in The additive white Gaussian noise in the second time slot, and The power allocation factor is and satisfies Due to the second time slot and Superimposed transmission on the same link, In decoding When Consider it interference, and thus obtain SINR is

[0077]

[0078] The signal-to-noise ratio (SNR) of the D2D signal received in this time slot can be written as:

[0079]

[0080] At the receiving end, the signal originates from the base station in the first time slot. The downlink, the second time slot comes from arrive The user-to-user LoS direct link. Here, an approximately optimal set union can be achieved through MRC, and its SNR can be calculated. However, the rate of D2D systems with DF cooperation is limited by relay decoding constraints, i.e. Reliable decoding is required in the first time slot. Therefore, the system and rate of the further Sch1 scheme are as follows:

[0081]

[0082] The coefficients reflect the time-domain multiplexing overhead of two-slot transmission. This result demonstrates that the Sch1 scheme can provide additional information to the remote layer via the inter-user LoS link and achieve cross-slot diversity gain through MRC.

[0083] (2) MRC assistance Cooperative forwarding strategy (MRC-Sch2): In a PASS system, the position of the clamping antenna and the waveguide phase will change the channel gain difference between the two users, which may lead to certain geometric configurations. Decline or The improved situation means that the traditional assumption that "near-end users are always stronger" no longer holds true in an instantaneous sense. At this point, to avoid... Degraded link quality leads to a significant decrease in system throughput. A cooperative role reversal strategy can be adopted: [The strategy involves...] In the first time slot, except for decoding its own layer In addition, further decoding of the near-end layer The signal-to-noise ratio used for forwarding in the second time slot, and thus for forwarding constraints, is... In the second time slot, Overlay sending With D2D signal ,but The received signal is ,in Similar to Sch1, In decoding When Considered interference, therefore the second time slot and The SNR is

[0084]

[0085] The receiver uses MRC to obtain diversity gain, and the combined signal-to-noise ratio is: Similarly, affected right Due to the decoding constraints, the system and rate of the Sch2 scheme can be written as follows:

[0086]

[0087] Therefore, it can be seen that the Sch2 scheme utilizes Cooperative signal forwarding, caused by changes in PA position or blockage The weakening of direct links improves the lower bound of system performance. Combining the two cooperation modes, NOMA-enabled D2D cooperative transmission can adaptively select the optimal cooperation scheme based on antenna deployment location and channel differences, effectively improving diversity gain and spectrum reuse gain.

[0088] In the two cooperative modes described above, if the receiver uses the MRC scheme to superimpose observations from two time slots, on the one hand, MRC only needs to use the signal-to-noise ratio (SNR) of each time slot for weighted combining, without relying on precise channel statistical modeling; on the other hand, in practical systems, cooperative nodes often struggle to perform high-precision phase pre-distortion calibration, while MRC does not require strict phase alignment, making it easier to implement. Therefore, MRC can obtain cross-time slot time diversity gain, effectively improving the receiving gain of weak signals. However, MRC also has some shortcomings. The inter-layer interference in the first time slot is treated as noise, and this interference will be included in the final received SNR during combining. It can be observed that MRC fails to actively eliminate the inter-layer interference generated by NOMA superposition in the first time slot. Therefore, when the transmitted SNR is high, the system performance is prone to entering an interference-limited state. To address this issue, this application proposes a power-domain linear combining (LC) method. The core idea is to utilize the second time slot to reduce inter-layer interference in the first time slot signal: The cooperating user designs and transmits a linearly superimposed coded signal with controllable amplitude and phase in the second time slot. Part of this signal is used to cancel interference components from the first time slot, another part is used to enhance the target layer signal, and a fixed portion of power is reserved for the D2D signal stream. Therefore, this application continues to discuss the LC-NOMA decoding strategy design under the Sch1 and Sch2 schemes.

[0089] (1) LC assist Cooperative forwarding strategy (LC-Sch1): In this mode, the first time slot is consistent with the MRC scheme, that is, the BS transmits the NOMA superimposed signal and then... Prioritize decoding of the remote layer However, in the second time slot, Sending a linearly superimposed signal with controllable amplitude and phase makes... Quantity reached Time and the first time slot The components exhibit opposite alignment, thereby achieving amplitude domain cancellation, and simultaneously making Components and the first time slot The components are aligned to achieve coherent enhancement and reserve power. Used for D2D signals Therefore, the signal transmitted in the second time slot is ,in , Correspondingly, The received signal is In order to offset the first time slot exist Interlayer interference generated at this point requires transmission in the second time slot. Amplitude matching is performed on the components. Because the interference intensity in the first time slot is... The decision is made, and the component corresponding to the second time slot is determined by... Therefore, the required offsetting factor can be designed as follows:

[0090]

[0091] in It is a small constant used in The value remains stable when the power is relatively weak. Taking into account the power constraint of the second time slot, we take... And obtained from the remaining power Regarding phase adjustment, right and Perform pre-rotation, so that arrive Time and the first time slot The proportions are relatively even, while arrive Time and the first time slot The components are aligned with each other, thus achieving cancellation and enhancement respectively. Under the above alignment, right The equivalent SINR is

[0092]

[0093] The second equation demonstrates the effect of the LC scheme in mitigating inter-layer interference. Furthermore, when... First decode After performing SIC, the equivalent SNR of the D2D signal stream is calculated as follows: Therefore, the total system rate of Sch1 under the LC decoding strategy can be written as:

[0094]

[0095] (2) LC assist Cooperative forwarding strategy (LC-Sch2): This strategy addresses situations where the PA (Portable Architect) may experience occlusion after movement, allowing the BS (Portable Architect) to... The effective link is in a poor state, while It has relatively stronger reception capabilities. This is to suppress inter-layer signal interference in the first time slot and improve the near-end user layer. Sch2's decoding capability Assisted by D2D link in the second time slot Receive signal. At this time... Design and send a linear superposition signal to enable exist To achieve coherence enhancement by aligning with the same position, and at the same time... exist The opposite alignment is used to achieve amplitude domain cancellation, i.e. ,in and To perform amplitude matching to reduce interference, for The power factor required for component calculation is

[0096]

[0097] and take , Under the condition of relative alignment, Second time slot pair The equivalent SINR is

[0098]

[0099] exist First decode After performing SIC on the second time slot, the signal-to-noise ratio of the D2D signal is: Therefore, the total rate of the Sch2 system under the LC scheme can be expressed as:

[0100]

[0101] In summary, the LC method utilizes the combined amplitude and phase control of the second time slot to cancel or significantly weaken the interlayer interference of the first time slot at the receiver, while simultaneously achieving coherent enhancement of the target layer. This aims to improve the performance limitations of traditional MRC in the interference-dominant region while maintaining the D2D multiplexing gain.

[0102] 1.2 P-NOMA Enables PASS-D2D Communication System

[0103] Building upon the previous steps, P-NOMA is further introduced to improve the system's schedulability and throughput performance under limited spectrum resources.

[0104] Unlike traditional power domain NOMA, which superimposes signals across the entire bandwidth, P-NOMA achieves signal superposition through partial resource overlap. and The signals only undergo non-orthogonal superposition on a portion of the resources, thus reducing the area occupied by cross-layer interference while preserving multiple access gain. Here, we consider that the first time slot is used for P-NOMA transmission by the BS, and the second time slot is used by cooperating users to decode and forward demand signals and D2D signals through a D2D link.

[0105] In the first time slot, the BS serves two users using the P-NOMA principle. and Two user signals partially overlap in transmission resources. Let They represent the allocations to and Based on the overlap ratio, the entire resource is divided into three regions: non-overlapping regions. Service Only Non-overlapping regions Service Only Overlapping areas Serving two users simultaneously, the scope is the same as above. For overlapping areas... ,set up and Assigned to respectively and The power coefficient. Furthermore, for all subcarriers, let the unit transmit power and noise variance be respectively... and The subcarrier signal-to-noise ratio is defined as Under the aforementioned PASS channel model, users in the first time slot SNR in different resource regions can be written as

[0106]

[0107] Among them, in the overlapping area Inside, After performing SIC on the overlapping region according to the P-NOMA decoding order, the target signal detection and decoding are completed. Similarly, The SNR in the first time slot is

[0108]

[0109] It is evident that the system degrades to single-signal transmission in non-overlapping regions, while inter-layer interference arises in overlapping regions due to power domain superposition. Furthermore, the PA position of the PASS and waveguide attenuation affect the reception quality of each user in different regions. Therefore, two D2D cooperative modes are further considered. In the second time slot, to achieve D2D cooperative forwarding, the cooperative nodes use power domain superposition to generate the forwarding signal. ,in , This indicates the demand signal being forwarded. This represents a D2D signal.

[0110] (1) MRC assisted Cooperative forwarding strategy (P-MRC-Sch1): In this scheme, In the overlapping region of the first time slot Decoding Signal, and in The upper layer completes its own layer decoding. In the second time slot, As a DF relay forwarding The message. Because DF cooperation must satisfy relay decoding constraints, therefore in superior, right The decoding constraints of the layer are

[0111]

[0112] exist At the end, the second time slot forwarding uses the MRC scheme with the first time slot to obtain cross-time slot diversity gain. The combined equivalent signal-to-noise ratio is The second time slot decoding The signal's SNR is .

[0113] Considering the bandwidth proportions of different regions and the transmission overhead of two time slots, the total achievable rate of the system can be written as:

[0114]

[0115] in Indicates the subcarrier range of the corresponding region. is the SNR of the D2D signal in the second time slot. Where... This reflects the constraints of DF cooperation at the relay point.

[0116] (2) MRC assistance Collaborative forwarding strategy (P-MRC-Sch2): Due to the adjustability of PASS's PA deployment, it is geared towards... The downlink may be at a disadvantage in certain deployment configurations. To improve the system's robustness in unfavorable locations, Sch2 allows... Additional decoding in the first time slot In overlapping areas The signal layer above satisfies the DF forwarding conditions of the second time slot. The decoding constraint at this time is... In the second time slot, Forward information, The direct observations from the first time slot and the relayed observations from the second time slot are merged using MRC, thereby achieving... Get it , .in Correspondingly, the achievable total rate of the system under Sch2 is

[0117]

[0118] The above results indicate that the P-NOMA overlap ratio By changing The bandwidth allocation is adjusted to modify the inter-layer interference structure. Simultaneously, the PA location and waveguide attenuation term affect the PASS downlink gain, thus impacting the performance advantage range of the Sch1 and Sch2 schemes. By introducing DF cooperation in the second time slot and employing MRC combining, the system can utilize cross-time slot diversity to improve the effective reception quality of the target layer, while also... Power multiplexing is achieved between cooperative forwarding and D2D signals.

[0119] To further reduce the impact of inter-layer interference, the linear combining concept is also introduced into the P-NOMA transmission structure. While forwarding the target layer, the cooperating user uses amplitude matching and phase pre-rotation to achieve amplitude domain cancellation between the cancellation component introduced in the second time slot and the interference component in the first time slot at the receiving end, and to make the target layer component and the target component in the first time slot superimposed in phase, thereby achieving the effect of interference suppression and target signal enhancement at the cross-time slot level.

[0120] (1) LC assist Cooperative forwarding strategy (P-LC-Sch1): The SNR of the first time slot still adopts...

[0121]

[0122]

[0123] During the second time slot, the base station remains silent. As a relay auxiliary Signal transmission is performed. Let the D2D signal power ratio be... Then the remaining power It is further divided into an interference cancellation section and a target signal enhancement section. This is for overlapping areas. The first time slot is Place and Aligned interference components have equivalent amplitude The magnitude of the offset term introduced through the D2D link is To achieve amplitude matching cancellation, the required coefficient is defined as follows:

[0124]

[0125] Further and obtained Regarding phase design, Pre-rotate the transmitted signal in the second time slot so that Quantity reached The time relative to the interference component in the first time slot is aligned in opposite directions, thus achieving cancellation. Simultaneously, it makes... Quantity reached The time slot is aligned with the target component in the first time slot to achieve coherence enhancement. Under the above conditions of amplitude matching and phase alignment, exist The SNR can be written as

[0126]

[0127] This structure is for the first time slot. Interlayer interference is canceled out. Then in After decoding, the D2D signal is decoded using SIC. Simultaneously, due to DF cooperation, the first time slot relay... Must be able to Successfully decoded Layer, whose constraints are

[0128] .

[0129] Considering the combined bandwidth ratio and the two-slot overhead, the total system rate under this scheme can be expressed as:

[0130]

[0131] In the non-overlapping regions No interference cancellation is required.

[0132] (2) LC assist Cooperative forwarding strategy (P-LC-Sch2): This scheme is mainly aimed at... In situations with poor channel conditions or obstructions, the BS will... The link is in a poor condition, and It has stronger reception capabilities. To suppress the first time slot in... Up by The resulting interlayer interference and enhancement Reliable decoding of the layer, the second time slot is provided by Assisted by D2D link The received signal is processed using the LC principle to achieve interference cancellation and target enhancement at the receiver. This is achieved through matching... In the first time slot and The amplitude of the alignment interference can be compensated for by designing a cancellation coefficient.

[0133]

[0134] Further take and order Under phase pre-rotation conditions, the second time slot arrives. of The component is aligned with the target component in the first time slot, while The component is aligned opposite to the interference component in the first time slot, thus forming coherent enhancement and amplitude cancellation. At this time... exist The SNR is

[0135]

[0136] Similarly, due to the use of DF cooperation, the first time slot Need to Additional decoding Layer, whose constraints are This is consistent with the above. Therefore, the total system speed under this scheme is...

[0137]

[0138] 2. PASS system rate analysis and clamp antenna deployment design

[0139] 2.1 NOMA-enabled PASS-D2D system rate closed-form solution derivation and performance analysis

[0140] To highlight the impact of PASS guided wave attenuation and two-time-slot transmit signal-to-noise ratio on performance, the constants related to system parameters are combined and defined as follows: and ,in and The normalized transmit signal-to-noise ratios of the two time slots are respectively characterized. This reflects the exponential attenuation effect of waveguide propagation. Based on this, the SNR of the near-end user's own signal under the NOMA scheme in the first time slot is denoted as... The SINR value before the near-end user performs SIC on the far-end layer is: The SINR of the remote user decoding its own signal is The second time slot activates only the D2D link and performs cooperative forwarding, while introducing a power allocation factor for the D2D signal. At this point, the SINR of the forwarded desired signal can be written as: The SNR of the D2D signal is To transform the ergodicity and rate from expected forms into computable probability integrals, this application employs a mathematical identity: for any nonnegative random variable... ,have .

[0141] Under a uniform planar distribution, the traversal rate of the first time slot needs to be adjusted accordingly. Two-dimensional regional integral. Definition and utilize It is possible to obtain the correct answer. Closed-form result

[0142]

[0143] in In the NOMA-based MRC-Sch1 scheme, the system and rate can be written as:

[0144]

[0145] Taking the expectation of the above expression, it can be written as: ,in , ,as well as , Assuming the distance between the two users remains constant, It can be approximated as and A directly related constant value. Then, using the CCDF identity, we obtain...

[0146]

[0147] For the first probability term, by ,right definition ,but Equivalent to .because Evenly distributed within a square area, along... The one-dimensional integral in the direction is

[0148]

[0149] for First, Conditionalization: Any real number ,Depend on achievable

[0150] This can be further deduced

[0151]

[0152] because Follows an exponential distribution, let The conditional expectation can be written in the form of a Laguerre weight function.

[0153]

[0154] Therefore, adopt The Gauss-Laguerre quadrature yields finite sums and approximations.

[0155]

[0156] at last We use the Q-point Chebyshev-Gauss quadrature. Let... , , , You can get The finite and closed expressions are

[0157]

[0158] Thus, the sum rate of NOMA's MRC-Sch1 scheme is... The mathematical closed-form solution can be derived.

[0159] In NOMA's MRC-Sch2 scheme, the direction of cooperation is opposite, i.e. Support Its sum rate is

[0160]

[0161] in , Taking its expectation, it can also be decomposed into...

[0162] ,in and Similar to the previous scheme, this section mainly analyzes...

[0163] Using the traversal logarithmic rate identity, we obtain...

[0164]

[0165] because exist The square region contains a strict lower bound. There is an upper bound, therefore the above integral can be performed on a finite interval. Precise calculations are performed.

[0166] For the first probability, by achievable

[0167]

[0168] For the second probability, first... Conditionalization and Gauss-Laguerre quadrature are used, expressed as follows:

[0169] Thus, a finite sum representation is obtained.

[0170]

[0171] Finally, for use The Chebyshev-Gaussian product is calculated at the point. Let... , , , ,but

[0172]

[0173] Based on the above analysis, the sum rate of the NOMA-based MRC-Sch2 scheme is... The mathematical closed-form solution can be derived.

[0174] In the LC scheme, the "amplitude domain cancellation" of the cross-layer interference in the first time slot and the "coherent enhancement" of the target signal are performed in the second time slot. Therefore, this scheme mainly targets the data contained in the rate expression. The partial derivation is performed in a closed loop, and the remaining terms can be directly derived from the corresponding results of the MRC scheme without further explanation.

[0175] In the LC-Sch1 scheme, the second time slot is... Align the phase and control the amplitude so that exist The component in the same direction as the first time slot achieves sign reversal alignment, thereby completing amplitude domain cancellation. exist The component at the point is superimposed in phase with the component in the first time slot to achieve coherence enhancement. Therefore, the SNR corresponding to the system bottleneck term can be written as: ,in , Accordingly, the traversal rate term to be calculated is: .

[0176] Here, the complementary distribution function identity is used, and it is noted that... Thus obtain

[0177]

[0178] in The derivation is completely consistent with the MRC-Sch1 scheme. (Changes) ,in We can obtain the one-dimensional integral form.

[0179]

[0180] In a given hour, , where defined ,when Then, it can be further equivalent to Therefore, conditional probability can be written as a one-dimensional integral in the coordinate domain.

[0181]

[0182] make Further To obtain the expected result

[0183]

[0184] At this time, it is available The Gauss-Laguerre quadrature yields a finite sum expressed as:

[0185]

[0186] in For Laguerre nodes and weights. Finally, for use The Chebyshev-Gaussian product is calculated at the point. Let... , , , ,

[0187] but It can be written as a computable finite and closed structure

[0188]

[0189] Similarly, in the LC-Sch2 scheme, the same method can be used to derive...

[0190]

[0191] Based on the above derivation, it can be seen that the closed-form analysis steps of the LC scheme are highly consistent with those of the MRC scheme: the direct link related terms in the first time slot can still be obtained through coordinate domain integration and... The inner closed-form integral is simplified; the collaborative bottleneck term is transformed into a probability product integral using the complementary distribution function identity. Furthermore, the exponential weight integral caused by D2D random distance is handled using Gauss-Laguerre quadrature, and the outer layer is completed using Chebyshev-Gauss quadrature. Integrating, we obtain finite and closed expressions.

[0192] 2.2 Design and Performance Analysis of P-NOMA-Enabled Decoding Scheme for PASS-D2D System

[0193] In the P-NOMA-enabled PASS-D2D collaborative framework, the system no longer uses the traditional power domain NOMA "full bandwidth superposition" transmission mode, but instead divides the normalized transmission resources into three complementary regions through partial superposition: and These are non-overlapping regions, each carrying only one load. and Data layer transmission, This is the overlapping region, used to carry the superimposed transmission of two user layers. Let... They respectively represent belonging to and The overlap ratio, then the bandwidth proportions are respectively , and In this structure, non-overlapping regions naturally possess interference-free single-user transmission characteristics, while DF-type cooperative bottleneck terms only appear in overlapping regions. To maintain consistency in symbol definitions, the following definition is used: and These are system parameters. User locations still follow a uniform distribution within a planar rectangle, and the D2D link distance is determined by the geometric relationships in the coordinate domain.

[0194] In the MRC scheme, the second time-slot D2D link is used to provide additional cooperation gain to the destination user. The destination performs maximum ratio combining (S2D) of the two time-slot observations in the corresponding resource area, that is, additively combining the received SNR / SINR of the first time slot with the D2D layer gain of the second time slot. The regionalized transmission structure of P-NOMA determines that combining only applies to resource areas related to the cooperation layer: in non-overlapping areas, since there is no superposition interference, the receiver does not need to perform S2D, and the traversal rate term can be directly written as... In the form of, and the conclusion of coordinate domain meanization from the previous section can be reused to transform the two-dimensional uniform distribution integral into a one-dimensional integral; in the overlapping region In the above scenario, near-end users need to decode the far-end layer first to ensure relay feasibility, while far-end users treat the near-end layer as interference and decode it directly. Furthermore, the destination end performs MRC merging of the target layer in two time slots. The forwarding layer rate is determined by the DF constraint. Taking Sch1 as an example, exist Top Layer decoding is by express, The merging of the two time slots is by Therefore, The bottle constraint on the top is Similarly, Sch2 simply swaps the relay and destination ends. Therefore, the key difference between P-NOMA and NOMA is that the constraint terms only exist in... Up, and and There is no interference above, which structurally reduces the constraints of DF constraints on the global sum and rate.

[0195] In the mathematical derivation, the ergodicization process of the non-bottleneck terms in the MRC scheme remains consistent with the previous section: for any Category items, first in Directional utilization The antiderivative is obtained by integrating the function within the closed form, and then... The direction can be solved by one-dimensional integration, so the derivation will not be repeated; the corresponding one-dimensional integral result will be directly used for calculation. In contrast, the bottleneck term needs to be addressed. The structure is still transformed into a structure using the complementary distribution function identity. The probability integral makes The internal correlation is decomposed into the product of relay events and destination events. Since the destination probability includes D2D random distance... Further By conditionalizing and incorporating the exponential weighted integral structure to introduce the Gauss-Laguerre quadrature, a finite and closed computational framework can be obtained. Finally, the outer layer... The integral is discretized into a finite sum using Chebyshev-Gauss quadrature. This derivation is similar to the NOMA scheme in the previous section, but it exhibits the regional scaling characteristics of P-NOMA in the threshold function and the upper bound of the integral; that is, all thresholds related to superposition are included. Scaling, and the bandwidth weight of the bottleneck term is changed by Explicitly given. Based on the above results, the traversal and rate of the P-NOMA-based Sch1 and Sch2 schemes under MRC can be divided into: bandwidth-weighted summation of the traversal rates of each region, Bottleneck terms, and D2D signal traversal rate.

[0196] In the LC scheme, the D2D transmission in the second time slot achieves amplitude domain cancellation and coherence enhancement at the destination through phase pre-spinning and amplitude control, thereby enhancing the overlapping region. The system suppresses cross-layer interference introduced by the first time slot superposition and uses the remaining power to enhance the equivalent signal energy of the target layer. Since the cooperative constraint of P-NOMA also only occurs... Therefore, the main differences in the derivation of the LC scheme compared to the MRC scheme lie in... In the bottleneck term, it can be seen that the analytical framework of the LC scheme can reuse the coordinate domain meanization and quadrature strategy of the MRC scheme to the greatest extent: the non-overlapping area still maintains single-user interference-free transmission, and its ergodic term is consistent with the MRC scheme. In the overlapping area, the relay end's decodeability probability is also not affected by the LC design, because LC only acts on the D2D transmission and destination observation structure of the second time slot, while the relay end decoding event is determined by the superposition structure of the first time slot. Therefore, the only thing that needs to be recalculated in the LC scheme is... The bottleneck section. Specifically, taking Sch1 as an example, The bottleneck term can still be written as ,in Similar to MRC, the difference is... The LC scheme achieves both cancellation and enhancement. After applying the complementary distribution function identity, the bottleneck term is transformed into... The domain probability integral, and the upper bound is still given by The boundedness is determined, therefore the outer integration interval remains unchanged. In the probability calculation at the receiving end, for After conditionalization, Can be equivalently converted to about The threshold event is then obtained by integrating the area of ​​the coordinate domain. And then Perform an exponentially weighted average. Since the structure of the exponential distribution remains unchanged, this averaging process can still be achieved through variable substitution. Introducing the Gauss-Laguerre quadrature yields a finite sum form, and finally the outer layer... The integral is discretized into a finite sum using the Chebyshev-Gauss quadrature, thus constructing a finite-sum closed-form solution for the sum rate. Similarly, for Sch2, only the destination end is... Swap to And replace the upper bound of the outer integral with the one given by The upper bound of the SNR at the upper relay end is determined That's all.

[0197] From a performance mechanism perspective, since DF constraints only act on... The system can and This allows for the continuous acquisition of interference-free rate contributions, thereby mitigating the drag on global performance and rate from bottlenecks in scenarios such as limited relay decoding, geometric asymmetry, or link obstruction. Simultaneously, Not only did it change the contribution weights of each region, but it also entered The threshold function alters the decodeable probability at the relay end and the over-threshold probability at the destination end, thus its impact on ergodicity and rate typically presents a non-monotonic trade-off. Compared to MRC, LC further alters the signal structure at the receiver through interference cancellation and coherent enhancement in the second time slot, enabling more efficient utilization of the D2D power budget in interference-dominated overlapping regions, potentially improving the effective SINR corresponding to the bottleneck term. However, this gain depends on the feasibility of phase alignment and amplitude matching, and is related to... , , The power allocation strategy and D2D link quality are coupled together.

[0198] Based on this, the MRC scheme emphasizes the direct superposition of energies from two time slots, resulting in a simple structure and robust implementation; while the LC scheme, while maintaining the regional advantages of P-NOMA, achieves this through... The structured suppression of cross-layer interference improves the efficiency of cooperative gain utilization and can obtain a unified finite and closed computational expression under the same coordinate domain meanization and quadrature framework, providing rigorous support for subsequent parameter optimization and performance boundary comparison.

[0199] 3. Simulation Results

[0200] The simulation focuses on core parameters such as the deployment location of the clamped antenna, user geometry, power allocation, and cooperation strategies. The study investigates the trend of the system traversal rate with varying transmit power, user spacing, and key parameters. For the PASS system, the results are used to verify the modeled coupling relationship between the waveguide-feed-clamped radiation link and user location, and to reveal the impact mechanism of different deployment locations on the service capabilities of near-end / far-end users. For the PASS-D2D system, the study further evaluates whether D2D forwarding can effectively alleviate the bottleneck of weak user links under a two-timeslot cooperation framework, and the performance differences exhibited by different decoding and combining methods in different regions.

[0201] Figure 2This paper demonstrates the trends in ergodicity and rate of OMA, NOMA, and P-NOMA in the PASS system as the transmit SNR increases. It also compares two typical propagation scenarios: an ideal scenario where both near-end and far-end users are in line-of-sight links, and a more general scenario where near-end users maintain line-of-sight while far-end users are in non-line-of-sight conditions. Overall, the curves for all three schemes monotonically increase with SNR. This is because increasing transmit power directly enhances the effective receive SNR at the user end, resulting in a higher data rate. Notably, the relative relationships between different multiple access methods are significant: regardless of whether it's a "LoS-LoS" or "LoS-NLoS" scenario, the ergodicity and rate of P-NOMA are consistently higher than those of NOMA and OMA. This advantage is more stable in the mid-to-high SNR range, indicating that the proposed partially overlapping transmission mechanism not only improves spectrum utilization efficiency but also more effectively suppresses the interference costs caused by multi-user overlap.

[0202] OMA employs fully orthogonal resource partitioning, which avoids co-channel interference but also means that the time-frequency resources available to each user are strictly divided, resulting in a strict upper limit on the system's spectral efficiency. Traditional NOMA uses superimposed coding transmission, relying on SiC to separate multi-user signals, theoretically offering stronger multiplexing potential. However, its performance is highly dependent on the interference structure and decoding reliability: when weak user link conditions are insufficient or interference is too strong, the success rate of SiC decreases, and residual interference accumulates, significantly offsetting the multiplexing gain. In contrast, P-NOMA's core advantage lies in the flexibility of partial overlap. By adjusting the signal overlap ratio, it achieves a continuously controllable trade-off between "multiplexing gain" and "interference management." In other words, when the system is in a region more suitable for superimposed transmission, P-NOMA can expand the overlap to obtain higher spectral efficiency; while when the link is unbalanced or decoding risks increase, it can reduce the overlap area to reduce interference for weak users, thus avoiding performance being dominated by residual interference. Therefore, P-NOMA can achieve higher and more robust ergodicity and rate in both types of propagation scenarios.

[0203] Comparing the two figures, it can be seen that when the remote user switches from LoS to NLoS, the overall rate level of all three schemes decreases. This is determined by the additional losses and fading caused by non-line-of-sight propagation. In particular, the decline is often more pronounced in the NOMA scheme in this scenario. This is because the remote user is already in a weak link state, and under superimposed transmission, it also needs to withstand co-channel interference from the near-end user's signal. The equivalent reception quality is more sensitive to interference. Once the decoding and cancellation process is not ideal, residual interference will reduce its rate, thus affecting the system and rate. Conversely, P-NOMA reduces the interference pressure on the remote user on the overlapping resources by reducing the superposition ratio or adjusting the proportion of the superposition area, thus showing stronger robustness to NLoS degradation. In addition, the comparison of different user spacing conditions in the figure shows that as the user spacing increases, the system ergodicity and rate will degrade more significantly, and this degradation is more easily amplified in schemes that rely on SIC. The direct impact of increased spacing is a decrease in received power and an exacerbation of link imbalance, reducing the decoding reliability of weak users and making SiC more susceptible to the combined effects of path loss and random fading. Once SiC fails to achieve stable success, residual interference will manifest in the mid-to-high SNR region, diminishing the multiplexing advantage of NOMA and even resulting in less than ideal growth efficiency in some areas. P-NOMA maintains relatively smooth performance changes with increased spacing because it can actively reduce overlap and interference levels, thereby avoiding performance limitations imposed by SiC failures.

[0204] Figure 3 The results of ergodicity and rate variations of three multiple access schemes (OMA, NOMA, and P-NOMA) in a PASS system as a function of the clamping antenna position are presented, revealing the coupling relationship between clamping antenna deployment and multiple access technology. Since the radiation point of PASS is determined by the position of the clamping antenna on the waveguide, changes in the clamping position will simultaneously alter the equivalent propagation distance and received power of the two users, thereby directly affecting the rates of the two users and the overall system rate.

[0205] Firstly, observing the OMA curve reveals that its optimal clamping position occurs near the midpoint of the geometric positions of the two users. This physically means that OMA avoids co-channel interference through orthogonal resource allocation, and the overall system rate is primarily determined by the individual link quality of each user. When the clamping antenna is located at the midpoint between the two users, the two links achieve relatively balanced geometric gain, preventing the weaker user from significantly dragging down the overall rate due to excessive distance. If the clamping antenna is biased towards one user, the equivalent propagation distance of the other user increases, and the received power decreases. Even with the influence of resource orthogonality, this will still lead to a significant reduction in its rate, thus decreasing the overall rate.

[0206] In contrast, the optimal clamping positions of NOMA and P-NOMA exhibit stronger bias characteristics, especially in the results with a fixed overlap ratio in the left figure, where the optimal points are often closer to the near-end user. The fundamental reason for this phenomenon is that the SiC constraint under superimposed transmission becomes a critical bottleneck for the system. For NOMA, the near-end user not only needs to decode its own signal but also typically needs to decode and eliminate the far-end user's signal to suppress interference and achieve a higher effective rate. When the clamping antenna is closer to the near-end user, the channel gain of the near-end user is significantly improved, and its receiver has a stronger ability to separate superimposed signals, especially with a higher probability of successful decoding of the far-end user's signal, thus reducing the rate loss caused by residual interference. Conversely, if the clamping antenna is too close to the far-end user, although the far-end user link will be strengthened, the decreased received power of the near-end user will directly weaken the reliability of SiC, and residual interference will quickly become apparent at medium to high SNR, leading to a more significant decrease in system performance and rate. Therefore, under fixed overlap ratio and fixed power allocation conditions, making it easier for the near-end user to complete SiC is often more effective in improving system performance and rate than further enhancing the far-end link.

[0207] Furthermore, P-NOMA typically maintains a higher overall performance level than NOMA and OMA on the same coordinate axis, but its curve is smoother and its sensitivity to clamping position is often lower than that of traditional NOMA. This is because the "partial overlap" of P-NOMA introduces additional degrees of freedom, allowing the system to avoid being subjected to strong interference coupling across the entire range. When the clamping position deviates from the optimal point, the full-bandwidth superposition of traditional NOMA makes interference loss almost unavoidable, while P-NOMA reserves the OMA portion for each user through non-overlapping resources, making the sum rate not entirely determined by the success of SIC. Therefore, P-NOMA not only achieves higher peak rates at the optimal point but also has better robustness at non-optimal locations, which is particularly important for scenarios with dynamic user changes in real-world deployments. The right figure further illustrates the effect of the overlap ratio on the optimal clamping position. Under the same power allocation conditions, when the overlap ratio changes, the trade-off between spectrum reuse and interference management is redistributed, leading to a shift in the optimal clamping position. Intuitively, a larger overlap ratio means a higher proportion of superimposed resources, and the system is closer to a state of strong multiplexing and strong interference; a smaller overlap ratio means the system is closer to a state of weak multiplexing and low interference. Under this mechanism, the curve trend of OMA reflects a shift: when the effective overlap ratio increases and a more significant shared resource effect is introduced, the optimal point of sum rate is no longer entirely determined by the balance of two independent links, but is affected by the equivalent gain of the shared resource segment. Therefore, its optimal clamping position may shift slightly from the strict midpoint to the side that is more conducive to the overall shared gain. At the same time, NOMA is more sensitive to the overlap ratio because its interference comes entirely from the superimposed segment. The larger the overlap ratio, the stronger the constraint of SIC, and the more likely the system is to be dominated by the near-end user; when the overlap ratio decreases, the SIC-related loss is weakened, and the contribution of improving the quality of the far-end link to the total rate is more prominent. Therefore, the optimal clamping position may show a trend of moving towards the far-end user side. Thus, the overlap ratio changes the dominant factors of system sum rate, making the optimal deployment position no longer fixed, but adaptively adjusted with parameters.

[0208] Figure 4Further analysis of the performance differences of different multiple access technologies in the PASS system from the perspective of power allocation and P-NOMA overlap ratio reveals the following: the left figure shows the trend of sum rate with the near-end user power allocation coefficient, while the right figure shows the variation of sum rate with signal overlap ratio under different power levels. In the left figure, the OMA curve remains almost horizontal because, under the premise of resource orthogonality, changes in power allocation do not introduce additional co-channel interference coupling; the sum rate is mainly determined by the single-user signal-to-noise ratio on the resources occupied by each user. In a non-interfering structure, adjusting the near-end user power only changes the rate allocation ratio between the two users to a certain extent, but the improvement in sum rate is limited, thus exhibiting weak sensitivity. In contrast, the NOMA curve steadily increases with the increase of the near-end power coefficient, indicating that under this simulation setting, the system sum rate benefits more from the enhancement of the near-end link. On the one hand, with higher transmit power, near-end users receive significantly better signal quality, thus directly increasing their rate contribution. On the other hand, near-end users decode superimposed signals more reliably, enabling them to complete the necessary interference cancellation process more stably and reducing the impact of residual interference on the system.

[0209] Unlike the monotonic increase of NOMA, P-NOMA exhibits a non-monotonic pattern of first decreasing and then increasing in the left figure, a key characteristic of partial overlap mechanisms. When the signal overlap ratio is small or at a fixed level, an increase in the near-end power factor amplifies the interference component of the near-end signal to the far-end user on overlapping resources. Since far-end users are already more constrained by path loss, their effective reception quality is more sensitive to additional interference. Therefore, as the power factor increases from low to medium range, the rate decrease for far-end users may exceed the gain from the rate increase for near-end users, ultimately leading to a decrease in the sum rate. As the power continues to increase, the rate contribution of near-end users increases rapidly, and a stronger near-end link allows the system to mitigate the interference effect through the structural advantages of P-NOMA, i.e., no interference on non-overlapping resources, while on overlapping resources, a reasonable overlap ratio prevents the interference term from being linearly amplified. At this point, the near-end rate increase becomes the dominant factor again, and the sum rate recovers accordingly.

[0210] As can be observed in the right figure, the NOMA curve remains almost constant as the overlap ratio changes. This is because traditional NOMA is equivalent to "full overlap" superposition transmission, and its resource structure does not include an adjustable overlap dimension. Therefore, its performance is mainly determined by power allocation and channel conditions, and does not change with the overlap ratio. In stark contrast, the P-NOMA curve shows a clear upward trend with the overlap ratio, and the rate of increase varies at different near-end power levels, indicating that the overlap ratio is an effective way to improve the rate of P-NOMA. When the near-end power is low, appropriately increasing the overlap ratio can increase the proportion of reusable resources, enabling the system to obtain higher spectral efficiency without significantly worsening interference, thus gradually increasing the sum rate. When the near-end power is high, the multiplexing benefits and interference from increasing the overlap ratio are simultaneously enhanced. However, because the near-end link is stronger and the system is more capable of suppressing or withstanding interference, the overall performance still shows an increase. However, the selection of the optimal overlap ratio needs to take into account the anti-interference capability of the far-end user. In other words, the right figure reveals the core design logic of P-NOMA: the overlap ratio is not necessarily better the larger it is, but rather a more suitable degree of multiplexing should be sought under given power allocation and channel conditions to obtain a higher sum rate.

[0211] Figure 5 System ergodic and rate results for the PASS-D2D cooperative system under different transmit powers are presented, covering the baseline direct transmission scheme BAS, two cooperative merging strategies MRC and LC, and P-MRC and P-LC after introducing the P-NOMA mechanism. Sch1 and Sch2 decoding / forwarding designs are compared, and corresponding analytical results are provided to verify the accuracy of the theoretical derivation. Overall, the sum and rate of all schemes increase with... The increase in transmit power leads to a boost because it simultaneously improves the equivalent reception quality of the first time slot PASS link and the second time slot D2D link, resulting in a stable throughput gain.

[0212] First, BAS, as the baseline curve without cooperative enhancement, is at the lowest level in both Sch1 and Sch2, with a relatively small growth slope. This indicates that when the system relies solely on single-hop PASS links, the total rate is easily limited by weak user links, especially when the user spacing is large or the quality of the remote user link is weak, where the contribution of weak users significantly reduces the total performance. After introducing D2D cooperation, the performance of both MRC and LC schemes improves, and they achieve a more significant power gain compared to BAS. This shows that the direct user link in the second time slot can effectively provide additional information support for weak users, thereby alleviating the constraint of the remote link bottleneck in the single-hop PASS downlink on system throughput. In particular, the cooperative gain does not only appear in the high-power range, but is also evident in the low-to-medium power range. This means that the D2D link provides not only additional power but, more importantly, changes the information acquisition path for weak users, enabling the system to achieve a more stable reliability improvement over a wider SNR range.

[0213] Secondly, comparing MRC and LC reveals that their differences under different protocols exhibit significant limitations. When the system bottleneck is primarily driven by minimum constraints at the relay / forwarding stage, the gap between MRC and LC is compressed, resulting in relatively close curves. However, when the bottleneck gradually shifts to the terminal merging stage or interference suppression becomes the dominant factor, LC often demonstrates better effective interference control and energy utilization efficiency, thus gaining an additional rate advantage. Combining this with the previous discussion of the "min" region, it can be observed that MRC favors simple energy-level merging, suitable for noise-constrained situations or where the two components are relatively independent; LC, on the other hand, achieves a better balance between interference and useful components through more refined signal-level weighting, thus having an advantage in regions where interference is more significant or where the information in the two time slots is more correlated.

[0214] Further observation of the Sch1 and Sch2 curves reveals a stable performance gap for the same scheme under both protocols, indicating that the decoding order and forwarding design significantly impact the release of cooperative gains. Overall, the superior protocol typically corresponds to a better interference structure and a more reliable decoding link, allowing the second time-slot D2D signal to be more effectively used to cancel or suppress residual interference in the first time slot, thereby increasing the effective rate for weaker users and driving up the overall rate. Conversely, the weaker protocol is more susceptible to SIC failures or residual interference accumulation; even with increased transmit power, some gain is affected by the interference term, resulting in a lower overall curve. In other words, the difference between the Sch1 and Sch2 schemes essentially reflects the system's different control capabilities over the interference propagation chain under different decoding paths. Furthermore, P-MRC and P-LC exhibit the highest overall rate across the entire range, showing a more significant upward shift compared to traditional cooperative NOMA schemes. This demonstrates the strong complementarity between the partial overlap mechanism introduced by P-NOMA and D2D cooperation: on the one hand, partial overlap prevents the superimposed transmission from being strongly coupled across the entire bandwidth, reducing the system's dependence on strict SIC success from the source and mitigating the penalties for residual interference pairs and rates; on the other hand, the D2D link provides additional information components for weak users, further enhancing their decoding and interference suppression capabilities on overlapping resources, thus truly converting the design freedom brought by adjustable overlap into rate gains. Therefore, P-LC typically achieves the highest performance upper bound because it simultaneously utilizes the dual advantages of partial overlap to reduce interference coupling and signal-level combining to improve effective SINR; P-MRC, on the other hand, achieves a similar performance improvement with lower complexity.

[0215] Figure 6 The rate response of the PASS-D2D system within a two-dimensional adjustable parameter space is further presented, with the horizontal and vertical axes representing the power allocation coefficient and user spacing, respectively. Four sub-graphs correspond to the MRC and LC strategies under Sch1 and Sch2, respectively. As the user spacing increases, the ergonomics and rate of all four schemes decrease. This is because both the D2D link and the weak user link are simultaneously affected by increased path loss, gradually weakening the gain provided by the second time slot forwarding, making the system more susceptible to being dominated by the "weak link bottleneck." Conversely, when the user spacing is small, the D2D link quality is better, and the cooperative information can more fully compensate for the detection error and residual interference in the first time slot. Therefore, the overall rate curve rises, exhibiting more significant optimization potential.

[0216] From a power allocation perspective, the rate surface is not simply monotonic, but exhibits different sensitivities under different protocols. Sch1-MRC and Sch1-LC typically show a superior "ridge" in the medium power allocation range, indicating that insufficient near-end power weakens its supporting role in cooperative links, while excessive near-end power exacerbates interference to far-end users during the first time slot overlap, thus lowering system performance and rate. In contrast, the two Sch2 graphs show a lower overall rate level and a steeper surface, indicating that Sch2, under this setting, is more likely to trigger minimum constraints or residual interference accumulation at relays, making the system more sensitive to power allocation and distance changes, resulting in lower performance.

[0217] Further comparison of MRC and LC reveals that the LC surface is smoother in some regions, especially with larger user spacing or more offset power distribution. LC is less prone to rapid rate drops compared to MRC. This indicates that LC has a stronger ability to suppress residual interference through signal-level weighted combining, and can maintain a more robust effective SINR in the "min" region. When the system is mainly constrained by the forwarding link itself, the difference between MRC and LC is compressed, and the surface shape tends to be consistent.

[0218] Figure 7 A rate heatmap of the P-NOMA-enabled PASS-D2D system in a two-dimensional overlap parameter space is presented, with the horizontal axis representing the overlap ratio of near-end users. The vertical axis represents the overlap ratio of remote users. The four subgraphs correspond to the MRC and LC merging strategies under Sch1 / Sch2, respectively. The overall system rate varies with... Increased and then decreased significantly, while for The sensitivity to [something] is relatively weak, with only a certain sloping gradient appearing in Sch2-LC. This indicates that under a given power allocation and two-slot cooperative structure, the performance bottleneck is mainly due to [something]. The interference coupling strength of the relevant overlapping segments is determined, and The overlap ratio has a greater impact on resource allocation and available bandwidth for weak users; its marginal effect is usually smaller than that of other factors. The resulting interference effect.

[0219] when When the signal is smaller, most system resources are in non-overlapping transmission states, limiting the degree of co-frequency overlap. Multi-user interference in the first time slot is significantly reduced, and the near-end interference pressure experienced by weak users in the overlapping segment is lowered, thus reducing the risk of SIC failure and residual interference. Simultaneously, D2D forwarding in the second time slot is more easily used to compensate for detection errors of weak users or to cancel residual components, allowing end-to-end gain to be more directly translated into throughput improvement. Therefore, the left side of the heatmap generally shows a higher rate level. With... As the number of overlapping signals increases and the proportion of superimposed resources rises, the system tends to have stronger multiplexing, but interference in overlapping segments also increases synchronously. Because the near-end signal arrival power is stronger, its equivalent interference to far-end users is more significant. The effective SINR at the far end is highly sensitive to this interference term, leading to a decrease in the reliability of its first time slot decoding. Even with the introduction of D2D cooperation, compensation efficiency will be limited, and the system is more likely to enter intervals dominated by residual interference or minimum constraints, thus resulting in interference along the edges. Rapid degradation of direction. This indicates that within this collaborative framework, a larger overlap ratio in P-NOMA is not necessarily better; excessively large overlap can lead to problems. This would make it difficult for the benefits of reuse to offset the costs of disruption.

[0220] In comparison, The main adjustment involves regulating the proportion and sharing level of non-overlapping resources outside the overlap area for remote users. Since the D2D link provides an additional information path for weak users, it does not entirely depend on the decoding performance of the overlap segment. The resulting fine-tuning of bandwidth share typically brings only limited marginal changes, manifesting as an insignificant vertical gradient. The sloping gradient observed in Sch2-LC indicates that the Sch2 decoding path may make weak users more dependent on the interference levels of specific resource segments, while the signal-level weighting of LC further amplifies the differences in quality between different resources. The changes began to have a noticeable impact on performance. Overall, the heatmap patterns of MRC and LC under the same cooperative scheme are similar, indicating that the overall trend in this group of parameter scans is dominated by the overlapping interference structure in the first time slot, and the merging strategy brings more constant-level gains.

[0221] In summary, this application focuses on the downlink transmission of PASS powered by clamp-on antennas and its D2D cooperative extension, aiming to systematically reveal the intrinsic relationship between "waveguide feeding—clamp-on radiation—user spatial distribution—multiple access interference coupling" from three levels: theoretical modeling, transmission mechanism, and closed-loop performance analysis. First, a channel model considering waveguide attenuation and geometric propagation loss is established for the PASS system, and the rate expressions and key parameter influence laws under three multiple access strategies—OMA, NOMA, and P-NOMA—are presented. The results show that P-NOMA, by introducing adjustable partial overlap, ensures a trade-off between spectral efficiency and interference management, achieving a higher and more robust sum rate under different link visibility conditions. Meanwhile, the clamp-on antenna position plays a decisive role in system performance; OMA tends to achieve optimal performance at the user's geometric midpoint, while the SIC-dependent superposition transmission scheme is more biased towards the near-end user side to ensure decoding reliability. Furthermore, this application extends PASS to a PASS-D2D cooperative framework, designs a two-slot transmission protocol and introduces MRC and LC merging strategies respectively, derives the closed-form solutions for the ergodic rates of each scheme, and verifies the accuracy of the analytical results through simulation. Numerical results show that D2D cooperation can effectively alleviate the bottleneck of weak users and improve system throughput. The combination of P-NOMA and cooperative merging can further release gains, with P-LC typically performing better in interference-sensitive areas. Meanwhile, power allocation, user spacing, and overlap ratio jointly determine the upper limit of cooperative gains and the interference-dominant region. Reasonably controlling the near-end overlap ratio helps avoid rate degradation caused by excessive interference coupling. In summary, this application provides an interpretable theoretical basis and numerical support for the parameter design and deployment optimization of PASS and its cooperative extension systems.

Claims

1. A design method for a P-NOMA-enabled PASS-D2D communication system, characterized in that, Includes the following steps: A clamp-on antenna PASS with an activatable dielectric waveguide is deployed, wherein the dielectric waveguide of the PASS is deployed at a height of [missing information]. Location, and allow along The axis flexibly activates the programmable amplifier PA, and the position of PA is set to... , The position is represented as ,in Two users are randomly distributed on a side with a length of Within the rectangular area, the centers of the rectangles are denoted as . and ; In the first time slot, the base station (BS) serves two users using the Partial Non-Orthogonal Multiple Access (P-NOMA) principle. and Two user signals partially overlap in transmission resources, making They represent the allocations to and Based on the overlap ratio, the entire resource is divided into three regions: non-overlapping regions. Service Only Non-overlapping regions Service Only Overlapping areas Serving two users simultaneously; In the second time slot, the cooperating users decode and forward the demand signal and the D2D signal through the device-to-device (D2D) link to achieve D2D cooperative forwarding. The cooperating nodes generate the forwarding signal using a power domain superposition method. ,in , This indicates the demand signal being forwarded. Indicates a D2D signal; The receiver uses maximum ratio combining (MRC) or linear combining (LC) to superimpose observations from two time slots to obtain diversity gain.

2. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 1, characterized in that, It also includes two collaboration modes: Pattern Sch1: by Support In the first time slot, R2 decoding in the overlapping region The signal is decoded at its own layer on R1∪R2; in the second time slot, As a decode-forward DF relay forwarder The message; Pattern Sch2: by Support In the first time slot, Additional decoding The signal layer on the overlapping region R2; in the second time slot, As a DF relay forwarding The news.

3. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 2, characterized in that, The MRC merger strategy includes: In the MRC-Sch1 scheme, the system and rate are... In the MRC-Sch2 scheme, the system and rate are... in, Indicates the subcarrier range of the corresponding region. Let SNR be the SNR of the D2D signal in the second time slot, where... This reflects the constraints of DF cooperation at the relay point.

4. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 2, characterized in that, The LC merging strategy includes: In the LC-Sch1 scheme, the signal transmitted in the second time slot is ,in , ; In the LC-Sch2 scheme, the signal transmitted in the second time slot is ,in and .

5. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 4, characterized in that, In the LC-Sch1 scheme, to offset the first time slot exist The interlayer interference generated at this location requires a cancellation factor designed to be... ， and take And obtained from the remaining power ; In the LC-Sch2 scheme, to reduce interference, [the following is taken]: The power factor required for component calculation is and take , .

6. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 4, characterized in that, In the LC-Sch1 scheme, after amplitude matching and phase alignment, the total system rate is: In the LC-Sch2 scheme, the total system speed is 。 7. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 1, characterized in that, In the P-NOVA, the user in the first time slot SNR in different resource regions user The SNR in the first time slot is 。 8. The design method for a P-NOMA-enabled PASS-D2D communication system according to claim 1, characterized in that, In the PASS, PA is connected to the user. The channel coefficient is expressed as ， The LoS channel of the inter-user line-of-sight link is represented as follows: ， in, Let be the free space loss constant. At the speed of light, For carrier frequency, For PA to user Free space propagation distance, This represents the waveguide propagation length from the feed point to PA. The absorption coefficient is... Let the waveguide internal phase be from the feed point to PA. For the guided wave wavelength, is the effective refractive index of the dielectric waveguide.

9. A design method for a P-NOMA-enabled PASS-D2D communication system according to claim 1, characterized in that, In the P-NOVA, the first time slot is encoded by the base station for both users. ,in For the power allocation factor of the near-end user layer, For the power allocation factor of the remote user layer, For users The expected information symbol and satisfy The total transmission power of the base station is The noise variance is The transmit signal-to-noise ratio can be defined as .

10. A design method for a P-NOMA-enabled PASS-D2D communication system according to claim 1, characterized in that, The DF cooperation must satisfy the relay decoding constraint in the overlapping region R2. right The decoding constraints of the layer are ， right The decoding constraints of the layer are 。

Images