Unmanned aerial vehicle positioning method and system based on sensing collaborative fusion technology

Abstract

The invention discloses an unmanned aerial vehicle positioning method and system based on a communication and sensing collaborative fusion technology, and the method comprises the steps: 1, building an unmanned aerial vehicle communication and sensing integrated system platform which comprises an integrated transmitter and an integrated receiver; step 2, in the integrated transmitter, carrying out communication and inductance integrated design on a radar waveform to form a radar wave beam; meanwhile, in the integrated transmitter, a communication symbol is generated, and a communication beam is formed; step 3, fusing the formed radar beam and communication beam to form an integrated waveform, and performing radio frequency modulation on the integrated waveform; 4, the integrated receiver carries out equipment demodulation on the integrated waveform, and after filtering, target detection and channel estimation are carried out on the unmanned aerial vehicle; and 5, according to a target detection result and a channel estimation result, performing optimization processing on the trajectory of the unmanned aerial vehicle. According to the method, an integrated waveform generation model is provided, and code tracking precision, compatibility and the like are evaluated. And earnings brought by communication and sensing integration are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the positioning technology of unmanned aerial vehicles, and particularly relates to an unmanned aerial vehicle positioning method based on a sensing and cooperation fusion technology and a system thereof. BACKGROUND

[0002] In recent years, with the development of unmanned aerial vehicle technology, unmanned aerial vehicles have shown great potential in many fields, such as logistics, agriculture, disaster relief and national defense. However, due to the low cost, simple operation and convenient carrying, some unmanned aerial vehicles have strong load capacity and are easy to be used by illegal personnel, seriously disrupting the order of air traffic control. Especially in the situation of anti-terrorism, security and stability, if unmanned aerial vehicles are used by illegal personnel, they will bring serious hidden dangers and threats to social security and national defense construction. In complex urban scenarios, due to the characteristics of high mobility, miniaturization and easy sheltering by buildings, traditional radar-based detection methods are difficult to achieve accurate and rapid detection in high-rise building environments. This not only limits the detection ability of high-power radars on long-distance targets, but also makes it difficult to meet the modern countermeasures requirements.

[0003] By using wireless communication to integrate unmanned aerial vehicles and mobile communication networks, and integrating the sensing data of multiple nodes in the network, the cooperative sensing area is far beyond that of a single radar. However, in the process of integrating unmanned aerial vehicles and cellular networks, many challenges are faced, such as mobility, interference and degradation of communication quality. Based on this problem, unmanned aerial vehicle and base station cooperative detection has been proved to be an effective method, which can significantly improve the overall performance of the system through the cooperative work of multiple unmanned aerial vehicles. In addition, the sensing and integration technology uses a unified transceiver and spectrum resource to realize load saving and spectrum reuse, while providing efficient cooperative detection and sensing data sharing capabilities, supporting positioning, ranging, speed measurement, imaging, detection, identification and other functions, which can effectively alleviate the problem of insufficient detection capability of single radar sensors and improve the overall performance and service capability of the network.

[0004] In addition, with the large-scale commercialization of 5G around the world, the problem of scarcity of low-frequency spectrum resources has become increasingly apparent, and spectrum resource management departments, communication technology companies and related research institutions around the world have turned their attention to the development and research of higher frequency bands. Among them, millimeter waves have attracted much attention and attention from the industry, and have become the focus of global B5G technology development. B5G will further deepen the mobile Internet and expand the scope of Internet of Things based on 5G, and will be deeply integrated with sensing and positioning and artificial intelligence to realize the intelligent connection of all things. Therefore, communication and sensing integration is an important development direction in the future, which further combines with artificial intelligence technology to realize the intelligent connection of all things.

[0005] Future B5G / 6G networks will adopt higher frequency band millimeter wave communication, resulting in large propagation loss, weak diffraction and diffraction ability, and relatively limited coverage. Although large-scale MIMO technology can solve the problem of millimeter wave propagation loss, directional beam transmission still cannot ensure reliable coverage in dense communication scenarios. Therefore, how to break through the coverage economy problem of high frequency communication is a difficult problem that B5G communication perception integrated network needs to solve. SUMMARY

[0006] The technical problem to be solved by the present application is: how to reduce the propagation loss, enhance the diffraction and diffraction ability, and provide a UAV positioning method and system based on the sensing coordination fusion technology.

[0007] To solve the above technical problems, the present application adopts the following technical solutions:

[0008] The UAV positioning method based on the sensing coordination fusion technology, characterized by comprising the following steps:

[0009] Step 1: Establish a UAV sensing integrated system platform, including an integrated transmitter and an integrated receiver;

[0010] Step 2: In the integrated transmitter, the radar waveform is designed for sensing integration to form a radar beam; at the same time, in the integrated transmitter, a communication symbol is generated and a communication beam is formed;

[0011] Step 3: The radar beam and the communication beam formed are fused to form an integrated waveform, and the integrated waveform is modulated by radio frequency;

[0012] Step 4: The integrated receiver demodulates the integrated waveform, and after filtering, performs target detection and channel estimation on the UAV;

[0013] Step 5: According to the target detection result and the channel estimation result, the UAV trajectory is optimized.

[0014] In the joint processing design of the radar beam and the communication beam in step 3, a low-speed moving code division orthogonal frequency division multiplexing joint communication and sensing 6G model communication MTC system is adopted, in which MTC user 1 and MTC user 2 simultaneously perform bidirectional communication and radar sensing through a line-of-sight link; each user is equipped with a double-array JCS transceiver composed of a transmitting array and a receiving array, which respectively generates a transmitting beam and a receiving beam;

[0015] The JCS signal processing system is arranged on the MTC user, an SINR threshold is arranged on the direct sequence-spectrum spreading switching module, the DSSS switch decides whether to use the CD-OFDM signal or the OFDM signal by comparing the signal-to-noise ratio threshold with the signal-to-noise ratio obtained by the communication channel estimation; under the condition of low signal-to-noise ratio, the DSSS switch is switched to the DSSS module, the original symbol is first expanded by the DSSS codebook matrix, and then is modulated by the OFDM modulator to generate the CD-OFDM signal; under the condition of high signal-to-noise ratio, the DSSS switch is empty, that is, the DSSS codebook matrix is set as the unit matrix, therefore, the original symbol is directly modulated by the OFDM modulator;

[0016] The user 1 first demodulates the communication signal of the user 2, takes the radar echo signal as a small interference, then restores the communication signal of the user 2 by using the known CSI, and removes the communication signal from the superimposed signal.

[0017] The isolation shielding plate and the electric leakage elimination module are arranged between the transmitting array and the receiving array.

[0018] The i-th frequency domain CD-OFDM symbol vector received by the user 1 is

[0019]

[0020] The code channel matrix is decomposed by using the Hermite matrix property, and the decomposed communication signal is

[0021]

[0022] After the demodulated communication symbol is obtained, the communication signal is removed from the received superimposed signal , and the i-th radar echo signal is represented as

[0023]

[0024] The average error propagation power (AEPP) of the CD-OFDM JCS system can be represented as

[0025]

[0026] The IPN variances of the CD-OFDM and OFDM JCS signal processing are

[0027]

[0028] Wherein, γ CD = η C γ OF . Since the CD-OFDM JCS processing enjoys the CDM gain η C . ​

[0029] The step 4 includes the following specific steps:

[0030] Step 4.1: a sparse low-rank channel model is established, a pilot is inserted into the modulated signal, an OFDM integrated signal is generated through serial / parallel conversion, inverse fast Fourier transform, cyclic prefix addition and digital / analog conversion and the like and is transmitted;

[0031] Step 4.2: a low-complexity channel estimation algorithm is established to obtain the trajectory information of the unmanned aerial vehicle.

[0032] In step 4.1, an observation matrix φ ∈ R M×N An observation is performed on a discrete coefficient signal x ∈ R N×1 with a sparse degree of K and a length of N, and an observation result y ∈ R N×1 ; the sparse signal is represented as x = Ψs, wherein Ψ and s represent a sparse basis and a sparse coefficient respectively; meanwhile, is defined as a perception matrix; the compressive sensing technology solves an underdetermined equation to reconstruct the original signal under the condition that the observation vector y and the perception matrix A are known; in the OFDM communication perception integrated system, the pilot transmitted by N subcarriers is represented as

[0033] Y = XH + Z = XFh + Z

[0034] wherein Y = [y1, y2, …, y N ] T , X = diag(x1, x2, …, x N ) is an N-dimensional diagonal matrix composed of the pilot signals; Z is an additive white noise; F is an N × P-dimensional Fourier transform matrix, and P is a multipath number; the channel impulse response h = [h1, h2, …, h P ] T ; let the perception matrix be A = XW, and the input / output relationship can be rewritten as

[0035] Y = Ah + Z

[0036] At the receiving end, the channel impulse response h is reconstructed by using a sparse signal reconstruction algorithm since the perception matrix A and the received signal Y are known.

[0037] The downlink channel transmission system is represented as

[0038]

[0039] wherein v l represents a beamforming vector related to the downlink communication link, and s lis the unit power owned by user l, s0 represents the covariance matrix of the dedicated radar signal When the base station transmits x, it simultaneously receives the uplink communication signal and the target reflection, d k represents the downlink signal of user k, and h k represents the uplink channel between user k and the base station;

[0040] Next, the echo signal of the MIMO radar is modeled: it is assumed that the radar channel is composed of a line-of-sight path, and the transmitting and receiving ULAs at the BS are both half-wavelength antenna spacings; it is assumed that the target to be detected is located at an angle of θ0, and the target reflection is given by where β0 is the complex amplitude of the target, mainly determined by the path loss and the radar cross section; based on the given uplink communication signal and the target echo, the signal received by the FD BS is represented as:

[0041]

[0042] where n represents Gaussian white noise, z represents unwanted signal-related interference, and z is represented as two parts, the first part corresponds to clutter reflected from the surrounding environment, and the second part is SI caused by the operation of the base station under consideration:

[0043]

[0044] The complete received signal at the base station is represented as:

[0045]

[0046] The radar SINR is used as the performance index of the sensing function, and the radar SINR is:

[0047]

[0048] where represents the interference channel, which is defined as the sum of I interference channels and the SI channel;

[0049] Similarly, by applying another set of receive beamforming transmitters We obtain the received SINR corresponding to user k:

[0050]

[0051] In the present application, multiple variables are combined, and design is performed under two criteria: 1) transmission power minimization; and 2) overall sum rate maximization, which respectively correspond to the power efficiency and spectrum efficiency improvement of the ISAC system;

[0052] 1) Transmission power minimization is represented as:

[0053]

[0054] where τ rad is a constant minimum SINR threshold required for successful completion of sensing operation, and denote the minimum SINR requirements for uplink user k and downlink user l, respectively;

[0055] 2) The sum and rate maximization expressions are given by:

[0056]

[0057] subject to γ rad ≥ τ rad ,

[0058]

[0059] A low-complexity channel estimation algorithm is adopted, which is given by:

[0060] The RIS-aided massive MIMO wireless communication system includes K users, a BS, and a RIS and a RIS smart controller; the RIS smart controller is in high-speed wired connection with the RIS and the BS, so that the BS can control the RIS in real time; wherein the UE sends a signal to the BS; as a MIMO system, the BS and each UE respectively adopt N r and N t antennas, and the RIS has N s units, and then the uplink channel between the UE and the RIS is represented as where each entry of G independently obeys the same complex Gaussian distribution The uplink channel between the RIS and the BS is represented as where each entry of H independently obeys the same complex Gaussian distribution The uplink channel between the UE and the BS is represented as where each entry of B independently obeys the same complex Gaussian distribution Considering a uniform linear array, the array response is given by where (g) T denotes the transpose operation;

[0061] The channel is modeled as G k = β k a(N s , θ k )a H (N t , θ uk ) and H = γa(N r , θ gr )aH (N s ,θ gt ), where β k ,γ:CN(0,1) represent the complex channel gains, θ gr ,θ gt ,θ k , denote the normalized departure and arrival angles of G k and H, respectively;

[0062] RIS as a phase shifter for the incident signal is usually modeled as a diagonal matrix, where each diagonal entry is independently controlled by a phase, let be the phase of the RIS element, we write the vectorized RIS element response as Let P = diag(p) be the phase shift matrix; finally, the signal received at the base station during the p-th time slot is in the form of:

[0063]

[0064] where x k,p ,k = 1,K,K, p = 1,K,P represent the pilot sent by the k-th user in the p-th time slot, is additive white Gaussian noise, let [x k，1 x k，2 ,..., x k，P ] T be the pilot sequence of the k-th user, assuming that the users are distinguished by x k1 ≠ x k2 | k1≠k2 , and where is known and power;

[0065] An iterative method is used to implement channel estimation: first, the OMP method is used to search for channel directions to quickly obtain angle information, and by adjusting the number of grids, the algorithm resolution and overhead can be easily traded off; assuming that the angle of the channel falls exactly on the point of the grid, the sparse representation of Y[p] is obtained, which is further rewritten in vector form, and through K iterations by the OMP method, the recovered Y[p] can be easily obtained, and in turn the angle estimation of the unmanned aerial vehicle can be obtained; continue to obtain the estimation of the equivalent gain of the entire cascaded channel; finally, according to the estimated channel parameters, the pilot of each user can be recovered, and by comparing the estimated pilot sequence with the known pilot sequence , the K users can be identified.

[0066] An unmanned aerial vehicle positioning system based on the sensory coordination fusion technology adopts the method, and the system performs the method of steps 1-5.

[0067] The application with the technical scheme has the following beneficial effects:

[0068] (1) Integrated waveform design

[0069] In-depth analysis of the built-in mechanism of integrated waveform generation and fusion, and suppression of waveform distortion caused by non-independent components by using carrier multiplexing time code. Further, the waveform or frame structure is designed according to the needs of communication and sensing, the integrated waveform design mechanism is studied, the influence of multi-path signal waveform on parameter estimation and performance limit is described, the waveform modulation technology of ubiquitous distributed nodes is explored, the waveform design scheme is analyzed and improved, the integrated waveform generation model is proposed, and the code tracking accuracy, compatibility and the like are evaluated. The benefits brought by integrated communication and sensing are improved.

[0070] (2) High-reliability integrated communication and sensing channel coding and estimation

[0071] For pilot design and channel estimation, the conditions required to be met by the pilot structure with extremely low peak-to-average ratio are analyzed theoretically, and the pilot insertion density under different fading environments is designed and optimized; a low-complexity channel estimation algorithm is designed by using the transmission characteristics and channel sparsity of centralized Massive MIMO; an equivalent generalized linear channel model for multi-user transmission is established, and a frequency domain equalization method and its low-complexity implementation structure are studied, so that the connection number or spectral efficiency is significantly improved.

[0072] (3) Beam management technology for integrated communication and sensing

[0073] Through the high-precision hierarchical beam codebook design assisted by sensing, a performance optimization scheme for beam alignment and beam tracking is proposed. Further, the trade-off relationship between communication rate, sensing accuracy and other sub-indices is analyzed, the channel state information feedback under the integrated communication and sensing performance index is used to excavate the key technology of integrated beam management, and an intelligent and efficient integrated communication and sensing beam management scheme is explored to comprehensively guarantee the timeliness of information and sensing transmission in complex and high-dynamic scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 To map the TP to the vTP through the environmental object / scatterer;

[0075] Figure 2 Common problem analysis diagram for ISAC transmission system;

[0076] Figure 3 Integrated communication and sensing system model diagram;

[0077] Figure 4 CD-OFDM JCS signal processing schematic diagram;

[0078] Figure 5 Signal processing process diagram of user 1;

[0079] Figure 6 Block diagram for sensing channel coding and echo transmission principle;

[0080] Figure 7 System diagram for OFDM communication sensing integration;

[0081] Figure 8 System model diagram for centralized Massive MIMO system;

[0082] Figure 9 Process flow block diagram for the built sensing integration system platform. DETAILED DESCRIPTION

[0083] An unmanned aerial vehicle positioning method based on sensing coordination fusion technology, comprising the following steps:

[0084] Step 1: Establish an unmanned aerial vehicle sensing integration system platform, including an integrated transmitter and an integrated receiver;

[0085] Step 2: In the integrated transmitter, the radar waveform is designed for sensing integration to form a radar beam; at the same time, in the integrated transmitter, a communication symbol is generated and a communication beam is formed;

[0086] Step 3: The radar beam and the communication beam formed are fused to form an integrated waveform, and the integrated waveform is subjected to radio frequency modulation;

[0087] Step 4: The integrated receiver demodulates the integrated waveform, and after filtering, performs target detection and channel estimation on the unmanned aerial vehicle;

[0088] Step 5: According to the target detection result and the channel estimation result, the trajectory of the unmanned aerial vehicle is optimized.

[0089] 6G needs a sub-centimeter-level positioning solution to meet various types of application scenarios in the future. To achieve this level of positioning accuracy, we should have a more in-depth understanding of the propagation environment of wireless signals. By obtaining the radio frequency map of the propagation environment, we can try to obtain the corresponding UE position. In this way, the multipath nature of the propagation channel will be of some help.

[0090] However, multipath-aided positioning techniques face various challenges. In the initial phase of environment perception, the reflection of the Transmission Point (TP) location with respect to all scatterer reflection surfaces in the map can be obtained. By decomposing the multipath channel into multiple LOS channels from multiple anchors, which are obtained by mirroring the TP (Transmission Point) on the surface of the scatterer at each path, called Virtual Transmission Point (vTP). The problem is that the number of vTPs grows linearly with the number of reflection planes and exponentially with the number of allowed reflections. This problem is particularly prominent in scenarios with a large number of outdoor scatterers. At the same time, the UE is not clear how to match each measurement parameter vector (composed of angle, time delay, and Doppler) to the vTP, which can result in a large positioning error. In general, the matching between observations and visible vTPs is a combinatorial problem with exponential complexity.

[0091] Multipath parameter estimation mainly refers to estimating the J main multipath component parameters of each transmit beam received by the UE, including time delay τ j , Doppler v j , channel path coefficient β j and the angle of arrival , i.e. the elevation and azimuth angles of the jth path. For all j, all these parameters are collected into a vector, denoted as θ j . Assuming that the transmitted signal on the mth beam is S m (t), the received signal is represented as follows:

[0092]

[0093] In the formula, X j (t i ; θ j ) is the received signal of the jth path, which contains the influence of beamforming at the transmitting end and additive white Gaussian noise at the receiving end. The joint estimation of these space-time-frequency parameters brings a complex non-convex optimization problem. In addition, the entanglement of path parameters also limits the accuracy and reduces the resolution of estimated parameters, thereby affecting their distinguishability.

[0094] At the same time, in the 6G ISAC system, the perception capability introduces several new Key Performance Indicators (KPIs), as shown in Table 1.

[0095] Indicator Description Coverage Distance and field of view angle limitations at which the system can detect objects Resolution Ability to distinguish multiple objects in terms of distance, angle, velocity, etc. Accuracy Difference between sensed and actual values in terms of distance, angle, velocity, etc. Detection and false alarm probabilities Probabilities of detecting the presence and absence of an object, respectively Availability Percentage of time the system can provide sensing services on demand Refresh rate Refresh rate of positioning data

[0096] Table 1.

[0097] In the UAV integrated sensing and communication system platform, the core is the waveform joint design of radar beam and communication beam in step 3. The KPIs of sensing and communication are in conflict with each other, which is a big problem. Especially, the design goals of the two waveforms are different - the communication waveform design focuses on the improvement of spectral efficiency, while the sensing waveform design focuses on the improvement of sensing resolution and accuracy.

[0098] Frequency Modulated Continuous Wave (FMCW) is usually used in radar systems, but it is not suitable for data transmission at the required rate of communication services. Some studies have modified the FMCW waveform to make it more suitable for communication systems. In such studies, the rising part of the chirp wave is used for communication systems, and the falling part of the chirp wave is used for sensing systems. A "trapezoidal frequency modulation continuous wave" (TFMCW) is introduced, which multiplexes the sensing and communication cycles in the time domain. Although these techniques can effectively multiplex communication data and sensing signals, the problem of low spectral efficiency still exists due to the presence of chirp sensing signals. Spreading the single-carrier waveform radar and communication signals in the code domain is another major research direction. In such waveforms, radar performance is affected by the sequence autocorrelation performance. Long spreading codes can provide good autocorrelation, but will reduce the spectral efficiency of communication. In addition, more complex algorithms are needed for Doppler estimation. It is necessary to further develop waveform design for 6G ISAC requirements, and to seek a balance between communication and sensing performance.

[0099] In ISAC systems, the integrated hardware scheme that shares baseband and radio frequency hardware can reduce overall energy consumption, reduce system size, and shorten the time delay of information exchange between systems, which is beneficial to the mutual promotion of communication and sensing in the distortion correction and compensation of hardware. Common problems in ISAC systems caused by hardware sharing, such as Figure 2 as shown.

[0100] The differences in communication and sensing evaluation metrics and algorithms result in vastly different hardware requirements for the two. Considering cost and scale, ISAC system hardware design will try to align with traditional communication architecture, taking into account the impact of hardware distortion on sensing performance. For example, the large capacity of the communication system depends on full-duplex isolation, while limited transmit-receive isolation is a major problem for OFDM waveforms in detecting static targets. Reasonable ISAC system radio frequency architecture and self-interference cancellation design is a key technical problem to be solved. To ensure performance, sensing requires the accumulation of coherent signals, so the sensing system is more sensitive to sampling jitter, frequency offset and phase noise. This feature of sensing in turn puts higher requirements on system synchronization and reliability. In short, when choosing ISAC waveforms, sensing algorithms and non-ideal distortion compensation schemes, the above hardware challenges must be considered.

[0101] At present, a code division multiplexing-based integrated sensing and communication signal processing technology is a potential applicable technology. The code division gain performance of the orthogonal codebook effectively suppresses the influence of electromagnetic environment interference on the bit error rate of communication demodulation. With the greatly reduced bit error rate, the communication signal characteristics can be recovered in combination with channel estimation, and the mutual interference between integrated sensing and communication signal reception signal processing can be further weakened by using a serial interference cancellation method, to achieve mutual gain effect of sensing and communication performance.

[0102] In the joint processing design of radar beams and communication beams in step 3, the present application considers a low-speed moving code division orthogonal frequency division multiplexing (CD-OFDM) joint communication and sensing (JCS) 6G model communication (MTC) system, wherein the MTC system performs synchronous and continuous radar sensing and unified communication spectrum and transceiver, as shown in Figure 3 MTC users 1 and 2 simultaneously perform bidirectional communication and radar sensing through line-of-sight (LoS) links. Each user here refers to each unmanned aerial vehicle, and each user (unmanned aerial vehicle) is equipped with a dual-array JCS transceiver composed of a transmitting array (TxA) and a receiving array (RxA), which respectively generate transmitting beams (TxBs) and receiving beams (RxBs). It is assumed that the transmitting power of user 1 and user 2 on each subcarrier is P1 and P2, respectively, and the antenna array is a uniform linear array (ULA). The number of isotropic antennas of TxA and RxA is represented by M and N, respectively. In order to solve the above minimum distance problem, in-band full-duplex (IBFD) operation is required to make RxA work continuously. In this case, the self-interference from TxA is large enough to destroy communication and echo reception. Therefore, an isolation shield plate and a leakage elimination module need to be set between TxA and RxA to reduce or even eliminate self-leakage interference.

[0103] As shown in Figure 4As shown, User 2 and User 1 have the same JCS signal processing system. A SINR threshold is assumed on the Direct Sequence-Spectrum Spread (DSSS) switching module. The DSSS switch determines whether to use a CD-OFDM signal or an OFDM signal by comparing the SINR threshold with the SINR estimated from the communication channel. Under low SINR conditions, the DSSS module is switched on, and the original symbols are first expanded using the DSSS codebook matrix and then modulated by the OFDM modulator to generate a CD-OFDM signal. This signal has coded demultiplexing (CDM) gain, improving the processing SINR and reliability at the cost of higher computational complexity. Under high SINR conditions, the DSSS switch is empty, meaning the DSSS codebook matrix is ​​set to an identity matrix. Therefore, the original symbols are directly modulated by the OFDM modulator, as the OFDM signal can meet reliability constraints under high SINR conditions without increasing computational complexity.

[0104] like Figure 5 As shown, we take the signal processing of User 1 as an example; the processing of User 2 is the same. User 1 first demodulates the communication signal of User 2, treating the radar echo signal as a small interference. Then, it reconstructs the communication signal of User 2 using the known CSI and removes it from the superimposed signal. Compared with the traditional OFDM JCS system, code division processing leads to an increase in CDM gain, which enables the proposed CD-OFDM JCS system to improve reliability and operate under low signal-to-noise ratio conditions. Furthermore, the improved reliability can greatly suppress error propagation, further ensuring the sensing performance of the full-duplex radar.

[0105] The i-th frequency domain CD-OFDM symbol vector received by User 1 is

[0106]

[0107] Using code channel matrix Despreading using the Hermitian matrix property yields the despread communication signal as follows:

[0108]

[0109] After obtaining the demodulated communication symbols, from the received superimposed signal After removing the communication signal, the i-th radar echo signal can be represented as:

[0110]

[0111] The average error propagation power (AEPP) of a CD-OFDM JCS system can be expressed as:

[0112]

[0113] The IPN variance of CD-OFDM and OFDM JCS signal processing is respectively

[0114]

[0115] Wherein, γ CD = η C γ OF . Since CD-OFDM JCS processing enjoys CDM gain η C , the IPN variance of CD-OFDM JCS processing is smaller than that of OFDM JCS processing, which leads to a smaller bit error rate.

[0116] Consider an ISAC system as shown in Figure 6 , wherein the dual-function FD BS receives communication signals from K single-antenna uplink users and transmits an ISAC signal through the same time-frequency resource.

[0117] Step 4 includes the following specific steps:

[0118] Step 4.1: Establish a sparse low-rank channel model, insert a pilot into the modulated signal, generate an OFDM integrated signal through serial / parallel conversion, inverse fast Fourier transform (IFFT), add a cyclic prefix, and digital / analog conversion, etc. Process and transmit, the specific process is as follows:

[0119] Since the wireless propagation signal and the perception signal echo are often sparse under the communication-perception integrated scenario, the application designs a sparse low-rank channel model in step 4. Traditional communication systems usually use least square estimation and MMSE algorithm for channel estimation. Considering the sparse low-rank channel model under the communication-perception integrated scenario, it is assumed that an observation matrix φ ∈ R M×N The discrete coefficient signal x ∈ R N×1 with a signal length of N and a sparsity of K is observed, and the observation result is y ∈ R N×1 . The sparse signal can be expressed as x = Ψs, where Ψ and s represent the sparse basis and sparse coefficients respectively. At the same time, is defined as the perception matrix. The compressed sensing technology can solve the underdetermined equation to reconstruct the original signal under the condition that the observation vector y and the perception matrix A are known. In the OFDM communication-perception integrated system, the frequency domain pilot signal input-output relationship can be expressed as

[0120] Y = XH + Z = XFh + Z

[0121] Wherein: Y = [y1, y2, …, y N ] TFor the receiving vector, X = diag(x1, x2, ..., x N ) is an N-dimensional diagonal matrix composed of pilot signals; Z is additive white noise; F is an N×P Fourier transform matrix, and P is the multipath number; the channel impulse response h = [h1, h2, ..., h P ] T The above equation has a similar form to the compressed sensing expression. Let the sensing matrix be A = XW, then the input-output relationship can be rewritten as follows:

[0122] Y = Ah + Z

[0123] At the receiver, since the sensing matrix A and the received signal Y are known, sparse signal reconstruction algorithms such as orthogonal matching pursuit (OMP) and sparse adaptive matching pursuit (SAMP) can be used to reconstruct the channel impulse response h.

[0124] Figure 7 The paper presents a channel estimation process based on CS in an OFDM communication-sensing integrated system. At the transmitting end, pilot signals are inserted into the modulated signal. The OFDM integrated signal is generated and transmitted through serial-to-parallel conversion, inverse fast fourier transform (IFFT), addition of cyclic prefix, and digital-to-analog conversion.

[0125] We consider a downlink channel transmission system where the sensing signal x is formed by multiple antenna beamforming, simultaneously performing radar sensing and downlink multi-user communication, which can be expressed as:

[0126]

[0127] Among them, v l s represents the beamforming vector associated with the downlink communication link. l This represents the unit power of user l. s0 represents the covariance matrix of the dedicated radar signal. When the base station transmits x, it simultaneously receives the uplink communication signal and the target reflection. k This represents the downlink signal for user k. Also, use h... k This represents the uplink channel between user k and the base station.

[0128] Next, we model the echo signal of the MIMO radar under consideration. We assume the radar channel consists of a line-of-sight (LoS) path, and the transmit and receive ULAs at BS are both half-wavelength antenna spacing. Assuming the target to be detected is located at an angle θ0, the target reflection is given by the following equation. where β0is the complex amplitude of the target, mainly determined by path loss and radar cross section. Based on the given uplink communication signals and target echoes, we represent the signal received at the FD BS as

[0129]

[0130] where n represents the Gaussian white noise, z represents the unwanted signal-related interference. z can be represented as two parts, the first part corresponds to the clutter reflected from the surrounding environment, and the second part is the SI caused by the considered base station operation:

[0131]

[0132] The complete received signal at the base station is represented as

[0133]

[0134] The performance of the radar and communication systems depends largely on the corresponding SINR. In particular, when considering point target detection in MIMO radar systems, the detection probability of the target is usually a monotonically increasing function of the output SINR. Therefore, we directly adopt the radar SINR as the performance indicator of the sensing function. We obtain the radar SINR as

[0135]

[0136] where represents the interference channel, defined as the sum of I interference channels and the SI channel.

[0137] Similarly, by applying another set of receive beamforming transmitters We obtain the received SINR corresponding to user k as

[0138]

[0139] In this invention, we mainly focus on the general transceiver beamforming design without imposing any strict constraints or strategies to eliminate the interference involved in the system. In this invention, multiple variables are jointly designed under two criteria: 1) transmit power minimization; 2) sum-rate maximization, corresponding to the power efficiency and spectral efficiency improvement of the ISAC system, respectively. Specifically, for the first design criterion, we consider minimizing the total transmit power consumption while guaranteeing the minimum SINR requirements of uplink communication, downlink communication, and radar sensing. The corresponding problem is represented as

[0140]

[0141] where τ radis a constant minimum SINR threshold required for successful completion of the sensing operation, and respectively represent the minimum SINR requirements for uplink user k and downlink user l.

[0142] We also want to maximize the sum rate of all uplink and downlink users under a limited transmit power budget while guaranteeing the sensing performance by constraining the minimum radar SINR. Therefore, we formulate the problem as:

[0143]

[0144] subject toγ rad ≥τ rad ,

[0145]

[0146] It is observed that both (7) and (8) are non-convex problems, and their global optimal solutions are usually difficult to obtain by polynomial time algorithms. In addition, the optimization variables are tightly coupled, which makes the problem more complex and difficult to handle. We will derive the optimal receive beamformers in closed-form expressions, and then design an efficient algorithm based on the SCA technique to optimize the BS transmit beamforming and user transmit power.

[0147] Step 4.2: Establish a low-complexity channel estimation algorithm to obtain the trajectory information of the UAV.

[0148] Considering the complexity of pilot design and channel estimation in centralized Massive MIMO systems, we will analyze the conditions required to be met by the extremely low peak-to-average ratio pilot structure from a theoretical level, design and optimize the pilot insertion density in different fading environments. At the same time, the present application designs a low-complexity channel estimation algorithm to adapt to the Massive MIMO system with large-scale antenna arrays, establishes an equivalent generalized linear channel model for multi-user transmission, studies the frequency domain equalization method and its low-complexity implementation structure, in order to improve the spectral efficiency and performance of the system, and promote the performance and reliability of the centralized Massive MIMO system in complex environments.

[0149] As shown in Figure 8 , the RIS-assisted massive MIMO wireless communication system includes K users (UEs), a BS, and a RIS and a RIS intelligent controller. The RIS intelligent controller is in high-speed wired connection with the RIS and the BS, enabling the BS to control the RIS in real time. We consider the uplink transmission case, in which the UE transmits signals to the BS. As a MIMO system, the BS and each UE respectively employ N r and N t antennas, and the RIS has N sThe uplink channel between the UE and the RIS is then represented as where each entry of G independently obeys the same complex Gaussian distribution Similarly, the uplink channel between the RIS and the BS is represented as where each entry of H independently obeys the same complex Gaussian distribution The uplink channel between the UE and the BS is represented as where each entry of B independently obeys the same complex Gaussian distribution Consider a uniform linear array (ULA), the array response is given by where (g) T denotes the transpose operation.

[0150] To simplify the channel estimation of B, we can first turn off the power of the RIS using the RIS smart controller. The channel estimation of B is then simplified to a point-to-point traditional MIMO channel estimation problem, which can be solved by existing methods. Once B is estimated, we turn on the RIS. When channel estimation is performed for the RIS-aided MIMO wireless communication system, the received signal component from B can be treated as a known constant, which means the impact of B can be completely eliminated.

[0151] Therefore, in this work, we focus on the channel estimation of G and H. At high frequencies, e.g., terahertz (THz), scattering suffers severe attenuation (more than 20 dB), which results in the power of the NLoS component being negligible relative to the LoS. Therefore, by considering only the line-of-sight component, we model the channel as G k = β k a(N s , θ k )a H (N t , θ uk ) and H = γa(N r , θ gr )a H (N s , θ gt ), where β k , γ: CN(0, 1) represent the complex channel gain, θ gr , θ gt , θ k , denote the normalized angle-of-departure (AoD) and angle-of-arrival (AoA) of G k and H, respectively.

[0152] The RIS, as a phase shifter for the incident signal, is usually modeled as a diagonal matrix, where each diagonal entry is independently controlled by a phase. Let For the phase of RIS elements, we write the vectorized RIS element response as On this basis, in order to write conveniently, the phase shift matrix is recorded as P = diag (p). Finally, we get the signal received by the base station during the p th time slot Its form is:

[0153]

[0154] Where x k,p ,k = 1, K, K, p = 1, K, P represents the pilot sent by the k th user in the p th time slot, Is the additive white Gaussian noise. Let [x k，1 , x k，2 ,..., x k，P ] T Be the pilot sequence of the k th user, we assume that the users are distinguished by x k1 ≠ x k2 | k1≠k2 And Where Is the known and power.

[0155] To solve the above problems, we use an iterative method to realize channel estimation. First, use the OMP method to search for the channel direction to quickly obtain the angle information, and by adjusting the number of grids, the algorithm resolution and overhead can be easily balanced. Assuming that the angle of the channel falls exactly on the point of the grid, the sparse representation of Y[p] can be obtained, which is further rewritten in vector form, and through K iterations by the OMP method, the recovered Y[p] can be easily obtained, and the angle estimation of the unmanned aerial vehicle can be obtained. Continue to obtain the estimation of the equivalent gain of the entire cascaded channel. Finally, each user's pilot can be recovered according to the estimated channel parameters. By comparing the estimated pilot sequence With the known pilot sequence , the K users can be identified.

[0156] Step 5: According to the target detection result and the channel estimation result, the unmanned aerial vehicle trajectory is optimized; according to the waveform demodulated according to the foregoing process, each unmanned aerial vehicle can be identified, and the running trajectory such as the angle of the unmanned aerial vehicle is obtained, so that subsequent processing such as positioning and tracking is performed on it.

[0157] In order to verify the waveform design, channel estimation and beam management method and technology of time-space integration, the application develops and builds a time-space integration system, and carries out key technology verification and performance evaluation on the platform, promotes the application demonstration of the technology, including algorithm verification and hardware platform verification two aspects: waveform generation and beam management method integrated into the platform verification, phased array front end integrated into the system to verify its function and performance.

[0158] The built-in sensing integrated system platform can transmit integrated communication-sensing integrated waveform, realize multi-target sensing and multi-user communication, higher spectrum efficiency, higher communication rate and sensing accuracy. The built-in sensing integrated system platform needs to control the energy consumption and system stability of the transmitted signal as much as possible, and miniaturization, low energy consumption and high performance are the design criteria of the system platform. Based on the structural characteristics of the communication system and the radar system, the built-in sensing integrated system is shown in Figure 9 .

[0159] The built-in sensing integrated system platform has a time-sharing transmission and reception function: in the transmission mode, the transmitter generates radar detection waves and user downlink communication symbols, and obtains an integrated signal through waveform design and weighted superposition based on different criteria. After modulation and amplification by the radio frequency front end, the integrated signal is transmitted. In the receiving mode, the receiver includes a target echo signal processing and an uplink communication symbol processing subsystem. The target echo signal processing subsystem removes communication information through matched filtering, and then performs target detection and angle of arrival estimation to determine the target. The uplink communication symbol processing mainly performs communication channel estimation and uplink communication symbol detection. The channel state information and target state information obtained by the receiver signal processing are fed back to the transmitter to assist waveform design and beam management.

[0160] In order to verify the effectiveness, accuracy, robustness and real-time performance of the proposed built-in sensing integrated waveform design, channel estimation and beam management method and technology, the proposed waveform design and beam management method are written into the signal generation module and the beam forming module of the transmitting end respectively, and the proposed channel estimation method is loaded into the channel estimation module of the receiving end. According to the test results on the built-in system platform, the algorithm is evaluated, optimized and adjusted.

Claims

1. A UAV positioning method based on sensory collaborative fusion technology, characterized in that: Includes the following steps: Step 1: Establish an integrated UAV sensing system platform, including an integrated transmitter and an integrated receiver; Step 2: In the integrated transmitter, the radar waveform is designed using a sensing integration method to form a radar beam; at the same time, communication symbols are generated in the integrated transmitter, and a communication beam is formed. Step 3: Merge the existing radar beam and communication beam to form an integrated waveform, and perform radio frequency modulation on the integrated waveform; Step 4: The integrated receiver demodulates the integrated waveform, filters it, and then performs target detection and channel estimation for the UAV. Step 5: Optimize the UAV trajectory based on the target detection results and channel estimation results.

2. The UAV positioning method based on sensory collaborative fusion technology according to claim 1, characterized in that: In the joint processing design of radar beam and communication beam in step 3, a low-speed mobile code division orthogonal frequency division multiplexing joint communication and sensing 6G model communication MTC system is adopted. In the MTC system, MTC user 1 and MTC user 2 simultaneously conduct bidirectional communication and radar sensing through line-of-sight links; each user is equipped with a dual-array JCS transceiver consisting of a transmit array and a receive array, which generate transmit beam and receive beam respectively. A JCS signal processing system is set up on the MTC user. A SINR threshold is set on the direct sequence-spectrum spread switching module. By comparing the signal-to-noise ratio threshold with the signal-to-noise ratio estimated by the communication channel, the DSSS switch determines whether to use a CD-OFDM signal or an OFDM signal. Under low signal-to-noise ratio conditions, the system switches to the DSSS module, expands the original symbols through the DSSS codebook matrix, and then modulates them through the OFDM modulator to generate a CD-OFDM signal. Under high signal-to-noise ratio conditions, the DSSS switch becomes empty, that is, the DSSS codebook matrix is ​​set to the identity matrix. Therefore, the original symbol is directly modulated by the OFDM modulator. User 1 first demodulates User 2's communication signal, treats the radar echo signal as a small interference, and then uses the known CSI to reconstruct User 2's communication signal and remove it from the superimposed signal.

3. The UAV positioning method based on sensory collaborative fusion technology according to claim 2, characterized in that: An isolation shield and a leakage current elimination module are installed between the transmitting array and the receiving array.

4. The UAV positioning method based on sensory collaborative fusion technology according to claim 2, characterized in that: The i-th frequency domain CD-OFDM symbol vector received by User 1 is Using code channel matrix Despreading using the Hermitian matrix property yields the despread communication signal as follows: After obtaining the demodulated communication symbols, from the received superimposed signal After removing the communication signal, the representation of the i-th radar echo signal is as follows: The average error propagation power (AEPP) of a CD-OFDM JCS system can be expressed as: The IPN variances of CD-OFDM and OFDM JCS signal processing are respectively Where, γ CD =η C γ OF Because CD-OFDM JCS processing enjoys CDM gain η C .

5. The UAV positioning method based on sensory collaborative fusion technology according to claim 1, characterized in that: Step 4 includes the following specific steps: Step 4.1: Establish a sparse low-rank channel model, insert pilot signals into the modulated signal, and generate an integrated OFDM signal through serial-to-parallel conversion, inverse fast Fourier transform, addition of cyclic prefix and digital-to-analog conversion, etc., and then transmit it. Step 4.2: Establish a low-complexity channel estimation algorithm to obtain the trajectory information of the UAV.

6. The UAV positioning method based on sensory collaborative fusion technology according to claim 5, characterized in that: In step 4.1, an observation matrix φ∈R is used. M×N For a discrete coefficient signal x∈R with sparsity K and length N, N×1 Observations are performed, and the observation result is y∈R N×1 ; A sparse signal is represented as x = Ψs, where Ψ and s represent the sparse basis and sparse coefficients, respectively; meanwhile, Defined as the sensing matrix; compressed sensing technology reconstructs the original signal by solving underdetermined equations given the observation vector y and the sensing matrix A; in an OFDM integrated communication and sensing system, the pilot signal transmitted by N subcarriers has a frequency domain pilot signal input-output relationship expressed as... Y = XH + Z = XFh + Z Where: Y = [y1, y2, ..., y N ] T For the receiving vector, X = diag(x1, x2, ..., x N ) is an N-dimensional diagonal matrix composed of pilot signals; Z is additive white noise; F is an N×P Fourier transform matrix, and P is the multipath number; the channel impulse response h = [h1, h2, ..., h P ] T Let the perception matrix be A = XW, then the input-output relationship can be rewritten as follows: Y = Ah + Z At the receiving end, since the sensing matrix A and the received signal Y are known, the channel impulse response h is reconstructed using a sparse signal reconstruction algorithm. In the downlink channel transmission system, the sensing signal x is formed by multiple antenna beamforming, simultaneously performing radar sensing and downlink multi-user communication, as follows: Among them, v l s represents the beamforming vector associated with the downlink communication link. l It represents the unit power of user l, and s0 represents the covariance matrix of the dedicated radar signal. When the base station transmits x, it simultaneously receives the uplink communication signal and the target reflection, d. k This represents the downlink signal of user k, and at the same time, it is represented by h. k This represents the uplink channel between user k and the base station; Next, the echo signal of the MIMO radar is modeled: It is assumed that the radar channel consists of a line-of-sight path, and the transmit and receive ULAs at BS are both half-wavelength antenna spacing; it is assumed that the target to be detected is located at an angle θ0, and the target reflection is given by the following formula. Where β0 is the target's complexity amplitude, mainly determined by path loss and radar cross-section; based on the given uplink communication signal and target echo, the signal received by FD BS is represented as: in n represents Gaussian white noise, z represents unwanted signal-dependent interference, and z is represented by two parts: the first part corresponds to clutter reflected from the surrounding environment, and the second part is the SI caused by the considered base station operation. The complete received signal at the base station is represented as follows: Using radar SINR as a performance indicator for sensing functions, the radar SINR is: in The interference channel is defined as the sum of I interference channels and SI channels; Similarly, by applying another set of receiving beamforming transmitters We obtain the received SINR for user k as follows: In this invention, multiple variables are combined, and the design is carried out under two criteria: 1) minimizing transmit power; 2) maximizing overall efficiency and rate, which correspond to improving the power efficiency and spectral efficiency of the ISAC system, respectively. 1) Minimizing transmit power is expressed as: Where τ rad It is the constant minimum SINR threshold required to successfully complete the sensing operation. and These represent the minimum SINR requirements for uplink user k and downlink user l, respectively. 2) The overall sum and rate maximization are expressed as: subject toγ rad ≥τ rad , 7. The UAV positioning method based on sensory collaborative fusion technology according to claim 5, characterized in that: A low-complexity channel estimation algorithm is adopted, specifically: The RIS-assisted massive MIMO wireless communication system includes K users, a BS, a RIS, and a RIS intelligent controller; the RIS intelligent controller establishes a high-speed wired connection with the RIS and the BS, enabling the BS to control the RIS in real time; wherein the UE sends signals to the BS; As a MIMO system, the BS and each UE respectively adopt N r and N t One antenna, RIS has N s Each unit is then represented as the uplink channel between the UE and the RIS. Each entry in G independently follows the same complex Gaussian distribution. The uplink channel between RIS and BS is represented as Each entry of H independently follows the same complex Gaussian distribution. The uplink channel between the UE and the BS is represented as Each entry in B independently follows the same complex Gaussian distribution. Consider a uniform linear array, the array response is given by Given, where (g) T Indicates the transpose operation; Model the channel as G k =β k a(N s ,θ k )a H (N t ,θ uk ) and H=γa(N r ,θ gr )a H (N s ,θ gt ), where β k ,γ:CN(0,1) represents the complex channel gain, G k The normalized departure angle and arrival angle of H; As a phase shifter for the incident signal, the RIS is typically modeled as a diagonal matrix, where each diagonal entry is independently controlled by a phase, allowing... For the phase of a RIS element, we write the vectorized RIS element response as The phase shift matrix is ​​denoted as P = diag(p); finally, the signal received by the base station during the p-th time slot is obtained. Its form is: Where x k,p ,k=1,K,K,p=1,K,P represents the pilot signal transmitted by the k-th user in the p-th time slot, N[p]: It is additive white Gaussian noise, let [x k，1 x k，2 , ..., x k，P ] T Let x be the pilot sequence for the k-th user. k1 ≠x k2 | k1≠k2 To differentiate users, and in It is a known sum and power; Channel estimation is achieved using an iterative method: First, the OMP method is used to search for the channel direction to quickly obtain angle information. The algorithm resolution and overhead can be easily balanced by adjusting the number of grid points. Assuming the channel angle falls exactly on a grid point, a sparse representation of Y[p] is obtained. This representation is then rewritten into a vector form, and the recovered Y[p] can be easily obtained through K iterations using the OMP method, thus providing the angle estimate for the UAV. Next, the equivalent gain of the entire cascaded channel is estimated. Finally, the pilot signal for each user can be recovered based on the estimated channel parameters by using the estimated pilot sequence. With known pilot sequences By comparing the results, K users can be identified.

8. A UAV positioning system based on sensory collaborative fusion technology using the method described in any one of claims 1-7, characterized in that: The system executes steps 1 through 5.

Images