1.Introduction

Generally,electronic countermeasure(ECM)jamming methods can be classified as either active jamming or passive jamming based on whether they have an internal emission source or not.

It is worth remarking that although there are many methods and techniques to jam the radar system,active deception jamming is one of the most complicated methods,which makes the radar system easily deceived.Predominately,this active deceptive jamming is usually implemented by using a digital radio frequency memory(DRFM)system[1,2],where high-speed sampling digital memory is used to intercept the radar signals and to generate replicas that simulate false targets(FTs)to deceive the radar system.The DRFM technique is widely used and has become the core controlling unit in various jamming systems.This active deception jamming produced by DRFM using an FT generator(FTG)can obtain the processing gain in the radar receiver,which can produce FT information.These FTs can confuse the radar receiver,hindering its ability to identify the true targets from the false ones and make a wrong judgment on the battlefield situation,even result in a waste of weapons for a combat system and exposion of the fire deployment.These deceptive FTs cannot be easily suppressed,and true targets are difficult to be detected.

Therefore,without FTs suppression methods,radar systems will be easily deceived.Henceforth,more concern should be considered to their probable abilities in the electronic counter-countermeasure(ECCM),which is essential for the survival and operational performance of radar systems in hostile environments.These ECCM methods mainly suppress the active deceptive jamming signals by exploiting the differences between the deceptive jamming and radar signals as well as the features of the FT.For conventional radar systems,the ECCM ability of radar is limited to a single angle of observation which may result in obtaining insufficient information.Consequently,a multistatic radar system certainly becomes the major development trend,which is an integrated radar system with several spatially separated transmitting and receiving facilities,where the information of each target from all receivers is merged and jointly processed[3,4].More precisely,in the multistatic radar,the information of physical targets will be received by the widely separated receivers,which will be different from each other due to the difference in target responses including radar cross sections in different scattering directions,path loss effect and the variable interval of view angles.However,because the jamming signals are all generated by the same source(FTG),the de-ceptive FTs received by each receiver are highly correlated and almost identical.As a result,the multistatic radar system provides more flexibility and improved performance against active deception jamming.

Several anti-jamming strategies have been proposed to combat active deceptive jamming in radar systems.In view of the fact that the repeat jammer lags at least one pulse behind the echo,due to the orthogonal pulse block,ECCM schemes against range deceptive jamming have been proposed in[5–7].It is worth noting that the matched filter(MF)and correlation based methods are the conventional signal processing procedures in the above mentioned pulse diversity methods,with the obvious consequences resulting from the high range sidelobes.Certainly,the performance would degrade considerably in the presence of strong jamming power.Moreover,the pulse diversity technique is not applicable when the radar stations cannot alter the transmitting signals.

In[8,9]suppression methods based on transmitting pulses with random initial phases in the pulse repetition interval(PRI)domain against velocity deception jamming have been addressed.However,these schemes are limited to the constraint of the single target scenario.

For multistatic radar, the strategies take advantage of thedifference in a special location and velocity between physicaland false targets to combat the active deceptive jamming.

1.3 标本采集与处理 充分暴露宫颈阴道部，用棉球轻轻蘸去过多的黏液；用HPV取样器尖端置入子宫颈口，顺时针方向旋转至少6-8周，以保证收集足够量的宫颈管内脱落细胞；取样后将取样器立即放入HPV保存瓶中，室温保存待检。

In[12],the frequency diverse array multi-input-multiout-put (FDA-MIMO)radar to suppress the deceptive jamming in the joint transmit-receive domain has been investigated.Moreover,a method based on subarray FDA signal processing was proposed to counteract deceptive ECM signals in[13].However,the suppression was based on the fact that the FDA radar steering vector is dependent on both range and angle.Thus,the FTs can be suppressed owing to the difference in either range or angle.Thus,these methods work well only when the angle of the FT is different from that of the true target,or when they are in close angle but far from each other in the range.However,the discrimination performance will deteriorate severely when the ranges of the false and true target are too close.

ECCM suppression methods based on motion feature compensation have been examined in[10,11]against active phantom tracks.These methods mainly focus on the concepts of the data processing.The assumption that phantom tracks have a different measure error from the true ones is strongly required,and perform well only when the range and the angle errors of phantom tracks are large.

专注不等于固步自封，一直以来，国威将前瞻适用的技术应用于裁切设备的研发，早在2009年就推出了67小型切纸机，填补了当时国内数码切纸机的空白。在林孝国的诠释中，数码切纸机就是“傻瓜机”，客户不需要掌握特别的技艺就能进行操控。而对于最近出现频率颇多的“智能化”，国威也早有考虑，今年6月，公司将名称从原来的“浙江国威印刷机械有限公司”更改为现在的“浙江国威智能设备有限公司”或正是其布局之一。在明年召开的PRINT CHINA 2019上，我们或将看到国威推出的智能化的无人操作的设备。

In this paper,we suppress the multiple FTs in multistatic radar based on the difference in the propagation time delays of the FTs for different receivers.The signal models of FTs and PTs are given first.Then,the time delays of the multiple FTs are analyzed and compared in each receiver based on cross-correlation test of the received signals in different receivers.A feature space is established with a time delay ratio among the various receivers,where FTs are approximately coincident with one position,while PTs acquire small dispersion.On this basis,the local-density based spatial clustering of applications with noise(LDBSCAN)clustering approach[18,19]is utilized to discriminate between the FTs and the PTs.

In[15],a suppression method based on coherent canceling in mutltistatic radar has been proposed.This method first accomplishes FT alignment and then cancels the multiple FTs.Even though this approach can work quite effectively in certain cases;it nevertheless has many drawbacks.The method requires several steps to suppress the FTs,as an additional target(referred as secondary target) will be embedded into the receiver and needs to be rejected.Therefore,the method resorts to the binary integration to accomplish the secondary target rejection in all receivers,which leads to a system and computational complexity.Moreover,the method aligns the FTs of different receivers and then cancels the FTs by subtracting the signal in different receivers.Thus,if the pulses being reflected off the true targets happen to have similar time delays(close in range),then subtracting will erroneously also suppress the true targets.Also,if the alignments of the FTs are not good due to the existing of internal noise,the FTs will not be cancelled and detected as physical targets(PTs),which have fatal results in the battle field,if the PTs are erroneously removed or if the FTs are wrongly detected.

The hierarchical clustering method is utilized to discriminate deception targets in a defined amplitude ratio feature space across different receivers in[16].A method for distinguishing active FTs based on signal fusion has been addressed in[17]by making use of the difference in the spatial scattering property.These methods assume that,Doppler shift and the amplitude and phase error make no difference to the received signals.Nevertheless,in practical radar systems,the differences of these features in the receiving signal always exist.

The design and initial assessment of low-cost multistaticradar that utilizes digital components was reported by Derham et al. in [14]. This design has described the development and initial evaluation of a short-range,prototype multistatic radar that works in the cooperative mode.Furthermore,the advantages and limitations of a low-cost digital radar design were highlighted and further system development potential summarized.

2.Signal model

A multistatic radar system with one transmitter and the ith receivers is considered to sense the surveillance area.The first radar operates as both a transmitter and a receiver,while the remaining radar stations are used as receivers.In the monitoring area,it is assumed that there are M extended PTs.To protect these PTs,an FTG is utilized to generate P FTs.It is assumed that all of the receivers can receive the deception jamming.

The transmitted signal is assumed to be a linear frequency modulation(LFM),which can be represented as

where a(t)denotes the basic pulse shape,φ is the initial phase,fcis the carrier frequency and k denotes the chirp rate.Although a(t)can be any valid pulse shape,it is usually selected to be a rectangle window,i.e.,a(t)=1 for-TP/2≤t≤TP/2 and a(t)=0 otherwise,where TP represents the pulse width.B=kTpis the signal band width.Without loss of generality,we omit φ for brevity in the following derivations.

On the other hand,when xi,j＜0,R,i[z]is given by

据彝良县外宣办介绍，此次滑坡塌方量达1万立方米以上，并阻断小河形成堰塞湖；油房小学教学楼全部被掩埋，据初查18名学生被埋在垮塌的教学楼内；学校附近2户农户房舍被掩埋，其中1户农户1家3口全部逃离，另1户1人被掩埋。

(iii)Reachability distance of an object p w.r.t.object o:it is denoted by RD(p,o),and it is defined by

The FTG implemented by the DRFM system creates coherent returns that emulate the amplitude,frequency,and temporal characteristics of actual radar returns,making it difficult to disseminate FTs from the PTs.More precisely,in DRFM,the radar signal is usually first downshifted in frequency and subsequently sampled by a high-speed analog-to-digital converter(ADC).The stored samples are manipulated in amplitude,frequency,and phase,and then processed by a digital-to-analog converter(DAC),up converted and emitted back to the victim radar[1].

该车加速没反应的故障并没有在维修中重现。万般无奈下只能多次试车，数次试车后故障再现。同时，显示屏上出现的画面中，在动力电池符号旁有一个“惊叹号”(图2)，这引起了维修人员的警觉，说明该车动力电池可能出现故障，接下来重点检查转向高压电池组。

Assuming that the DRFM is sufficient to follow the radar signal and the quantization bits are large enough to ignore both the quantization error and the processing time[20].Also,let the FTG signal be Qp(t),which generates the pth FT.Thus,the received deceptive jamming signal J(t)can be expressed as

where P is the number of the FTs,βpiis the modulated amplitude of the qth FT,fd,piis modulated Doppler frequency of the qth FT,and τpiis the propagation time delay of jamming for the ith radar.The deception signal Qpis the time delay of frequency modulated transmitted signal s1(t),which can be represented as

Furthermore,the discrete-time signal SR,i[z]for the ith receiver is achieved by sampling(5)at the time instant ti=z/fs,z=1,...,Z(fsis the sample frequency)and can be expressed as

It is worth remarking that the parameters of PTs which include amplitudes,time delays,and Doppler frequencies are different from different widely separated receivers because of the electromagnetic responses in different receivers possess diversity.Nevertheless,the FTs for all receivers are all produced by the same source(FTG).Hence,they are coherent FTs,of which their modulated amplitudes and Doppler frequencies are all the same except for their time delays.

首先，需要对收集的资料进行质量控制，保障其可靠性。收集的资料必须经过当地城市规划主管部门审批，主要是建筑工程规划审批表、建设用地规划许可证、建筑工程规划许可证、审批规划方案或详规图及电子版、审批单体建筑施工图及电子版、勘测定界图电子版、审批建设工程施工变更资料等。这些资料由建设单位提供图纸复印件并加盖建设单位公章，电子版应与纸质图纸保持一致。

3.Coherent clustering discrimination method

In this section,the clustering analysis approach is utilized to discriminate between the PTs and the FTs.Before the clustering analysis,this method first estimates the propagation time delays τpiof the FTs based on the cross correlation function between different receivers[15],and then compensates these times to the reference receiver.

3.1 Time delays estimation

Therefore,the FTs time delays are estimated by the cross correlation function.Specifically,by taking SR,i[z]as the reference signal,the cross-correlation function CCFi,j[z]between SR,i[z]and SR,j[z]can be expressed as where⊗ denotes the convolution operation and[]∗represents the conjugate of a complex number.

Consequently,the time compensation xi,jbetween the ith receiver and the jth receiver(i/=j)can be obtained by extracting the maximum value of CCFi,j[x],in which the maximum value is the integration of all FTs.More precisely,for every target in the different receivers,the propagation time delays are different.Thus,the cross-correlation can achieve signal integration of one target at most,which will not be the maximum value under the presence of multiple FTs.Accordingly,the maximum value of CCFi,j[x]can be expressed as

循此，论文不揣谫陋，探讨“绿色原则”向《民法典》“物权编”辐射过程中生发的上述课题，以助推“物权编”编纂的顺利展开。

where|.|refers to the modulus of a complex number.

Thus,with the estimated time xi,jthe compensation of the signal CCFi,j[x]to the reference signal SR,i[x]can be obtained by shifting this estimated time,which can be explained as below.

In the case when xi,j＞0,R,i[z]is given by

Consequently,without the effects of the deceptive jamming,the received signal at the ith receiver is given by

It is worth noting that,after compensating the FTs time delays,the ratio across different receivers to the reference receiver will be approximately the same(time alignment),and therefore this characteristic can be used for FT association.However,the PT time delay is not well compensated for,thus it will possess small dispersion,which will not concentrate on one point.Before calculating the time delays,the corresponding relation of the detected targets in different receivers needs to be determined,which is defined as the target association.For every target in the reference receiver,the corresponding target should be found in the rest receivers.Fortunately,since the distance between receivers is generally not very long,the corresponding target is in the adjacent several range cells.All the FTs are created by the same deception signals,as a result of which their appearing time sequence,the intervals and the amplitude fluctuation are all the same in every receiver.This characteristic can be used for target association.For every possible target association situation,its correlation coefficient can be calculated.

Based on these characteristics,it is apparent that the clustering analysis approach is suitable for discriminating between the PTs and the FTs.The clustering result expects that all FTs gathering into one cluster introduces slight dispersion or coincidence into one point,while that of PT introduces a large dispersion in one cluster.

Noteworthy,in the discrimination method the time synchronization between the receivers is necessary,which is not a challenge and quite simple to be realized.In this paper,we suppose that the time synchronization has been accomplished[21,22].Furthermore,we assume that the positions of the transmitter and receivers(multistatic radar geometry)are known previously,even if the geometry condition does have an effect on the discrimination performance.

3.2 Clustering method

Clustering is an unsupervised technique,which is used to decompose the input data into individual meaningful groups based on the similarities between these objects.Different kinds of clustering may provide different clustering results.Therefore,it is imperative to identify the type of clustering based on the geometrical form of the data.In density-based clustering methods,the clusters are formed as regions in the data space in which the objects are of similar density,and separated by regions with different object densities.These regions may have an arbitrary shape and the points inside a region may be randomly scattered.In this paper,LDBSCAN is commonly used to reveal the data set,which has a much-skewed density distribution.

In order to implement the LDBSCAN algorithm in our dataset some terminologies must be defined.

关于旅游产业概念的界定，由于研究的角度不同，所得出的结论也自然不同，但共识性的因素主要是围绕着“旅游资源”、“旅游设施”、“旅游服务”等。所以对旅游产业概念可以表述为：“以旅游者为对象，借助旅游资源和旅游设施，满足旅游者在旅游活动中的吃、住、行、游、娱、购等需求，来实现旅游者精神和物质追求的综合性产业”[3]。

(i)k-distance of an object(p):it is denoted by kdist(p),and it represents the distance(d(p,o))between p and o∈D,which satisfy two conditions:i)There are at least k objects∈D,such that,d(p,)≤d(p,o).ii)There are at most k-1 objects,∈D,such that,d(p,)＜d(p,o).

学术英语智能写作评估的新研究——评《基于语类的二语研究论文智能写作评估：从设计、评估到优化》………………………………………………………………………………………… 钟家宝 钟兰凤（4.108）

(ii)k-distance neighbourhood of an object(p):Given the k-distance of p, the k-distance neighbourhood of p containsevery object whose distance from p is not greater than the k-distance,i.e.,

These objects q are called the k-nearest neighbours of p.It is simplified by Nk(p).

where M is the number of targets,vi(t)is the Gaussian thermal noise in the ith receiver,αmidenotes the amplitude(the complex parameter accounting for the target radar cross section and the channel propagation effects for the bistatic path,i.e.,the total path of the transmitter to the targetand the target to the ith receiver) of the mth target in the ith receiver,τmiis the propagation time delay of the mth target in the ith receiver(which equals the total path length of the transmitter to target range R1and the mth target to the ith receiver range Rmi,i.e.,τmi=2(R1+Rmi)/c,where c is the speed of light),and fd,miis the Doppler frequency of the mth target in the ith receiver throughout the same path fd,mi=

(iv)Local reachability density(LRD)of an object p:it is used to identify clusters in the dataset,and it is defined by

Intuitively,the LRD of an object p is the inversion of the average reachability distance based on the Min Ptsnearest neighbors of p.

(v)Local outlier factor(LOF)of an object p:it represents the degree of p to be outlier,and it is defined by

In LDBSCAN,each object is clustered by the value of LOF and LRD.Obviously,the point p becomes further from its nearest cluster when the LOFMPts(p)becomes higher and the ratio of LRDMPts(p)and LRDMPts(o)gets larger.It might be notable that the LOF represents the degree that the object is being outlying and the LOF of the main objects p in a cluster is approximately equal to 1,thus if LOF(p)≤LOFUB,p is called a core point and it should belong to a particular cluster.

(vi)Directly local density-reachable(DLDR):the object p is DLDR from the object q w.r.t.MPts and pt if

(vii)Local density-reachable(LDR):a point p is LDR from point q if there is a chain of points from p to q and each point is DLDR from its previous point.Here,the parameter pct is used to control the fluctuation of localdensity.Nevertheless,in general,it is not symmetric if q is not the MinPts-nearest neighbour of p.

④明确规定供水企业、用水户、管水部门的权利、义务以及相关法律责任，使供、用、管水三方能够从各自角度出发自觉执行计划用水管理办法。

GivenMPts, pt and LOFUB, the clusters can be identified by selecting any core point p from the dataset,and then all points that are LDR from p obtain the cluster.It is worth noting that,the result of LDBSCAN is influenced by choice of the LOFUB,pt,and MPts,and thus the right values of these parameters must be properly determined.More details about selecting these parameters can be found in[19,23]and references therein.However,choosing these parameters depends on what we want from the algorithms and in our situation;we choose them according to theoretical analysis and experimental viewpoints since most of the data sets that have been experimented have listed their parameter values.Thus,we make sure that all the clusters on the similar density are clustered,which guarantees no change of the clustering result can be given.

Based on the final clustering results, we measure the distance between the points inside each cluster.If the distance is less than or equal to some threshold η,this implies the cluster contains FTs.Else,the cluster contains PTs.For sufficient discrimination between FTs and PTs,the threshold is chosen to be very small,i.e.η=0.002.

The whole structure and understanding for the proposed ECCM technique can be summarized in Fig.1.

Fig.1 Structure of the proposed ECCM technique

4.Simulation results

This section is aimed at evaluating the performance of the proposed method.To this end,we consider a multistatic radar consisting of one transmitter,which also can be used as a receiver and two sufficiently separated receivers,i.e.,N=3.All the receivers are assumed to have the same antenna gain.The transmitted LFM signal is chosen to have a pulse width of Tp=10 μ s and a signal band width of B=5 MHz.The carrier frequency is 2 GHz.It is assumed that two PTs are present at ranges equal to R11=80 km and R12=100kmto the receiver1 with radial velocities equal to v11=900 m/s,v12=800 m/s respectively.For receiver 2 these PTs appear at ranges equal to R21=90 km and R22=125 km with velocities equal to v21=750 m/s,v22=650 m/s respectively,whereas the PTs appear at ranges equal to R31=88 km and R32=112 km with velocities v31=850 m/s,v32=750 m/s respectively to receiver 3.

Furthermore,it is assumed that all of the receivers are affected by the deception jamming.The DRFM FTG generates six FTs to deceive the multistatic radar system,of modulated ranges of 68 km,85 km,93 km,105 km,113 km,and 120 km.Also,their modulated velocities are all assumed to be the same,i.e.,300 m/s.Moreover,the signal-to-noise ratio(SNR),the jamming-to-signal ratio(JSR)and the jamming-to-noise ratio(JNR)are defined as below:

In the simulation,we set the clutter(noise)power as 2σ2=1,SNR=JNR=0 dB and JSR=5 dB.

4.1 Single simulation

Firstly,the cross-correlation function CCF1,2[x]for SR,1[x]and SR,2[x]is plotted in Fig.2 in order to estimate the FTs time delays.The result indicates that the maximum value is much greater than the other values and obtained when x1,2=1 340.This value is used to compensate the time delays of receiver2 where all FTs approximately have the same time delays when compared to receiver 1.The above step is also used for receiver 1 and receiver 3,and the results are plotted in Fig.3.

采样间距的设置对于数据采集精度的控制有着直接影响。若采样间距设置较大，后期数据处理及存储均会受到影响，而设置较小的数据采样间距，则会影响三维激光扫描仪扫描精度。本实验选用的三维激光扫描仪型号为VZ4000，每站的扫描时间控制在6 min左右最为适宜。

Fig.2 Cross-correlation function CCF1,2[x]

Fig.3 Cross-correlation function CCF1,3[x]

Secondly,receiver1 and receiver2 matched filter output in the presence of target and deceptive FTs before compensating the time delays are shown in Fig.4.Whereas receiver 2 matched filter output after compensating the FTs time delays is depicted in Fig.5.

3.融入高校专业背景。想要在高校导视系统设计中应用文化因素，我们可以在设计的过程中融入高校各学科的专业背景。在我国，高校的建设一般都有其学科侧重，并设置了国家或是省级重点专业，因此，在文化因素运用于高校导视系统设计的过程中，我们可以很好的将高校专业背景以及学科文化运用到高校导视系统的设计之中，如此一来，观者就能够更好的通过校园导视系统来感受高校的校园文化以及学术氛围，进而促进高校校园文化的传播。

Fig.5 Matched filter output after compensating the jamming delay times of receiver 2

It is evident to see that the selected FTs appear in the same positions(68 km,93 km and 120 km),while the PTs have different positions when compared with receiver 1(79.95 km and 102.5 km).This coincides with previous theoretical analysis and confirms that the clustering algorithm can utilize their differences to discriminate the FTs.It is worth noting that due to the existence of internal noise,the FTs might not always be matched at the exact time,but it is enough to support the visibility of the clustering algorithm.

细颗粒含量和加水量对灰岩料压实质量的影响……………………………………黄 恩，张 靖，王庆玲，等（5.53）

Thirdly,receiver 3 matched filter outputs before and after compensating the FTs time delays are illustrated in Fig.6.Similarly,it is obvious that this figure confirms the achieved results in receiver 2.

Fig.6 Matched filter output before and after FTs time delays compensation

Fourthly,the LDBSCAN algorithm is used to discriminate between PTs and FTs,where the parameters MPts,pt and LOFUB are selected to be 2,0.2 and 1.5,respectively.Then,the final results are obtained by measuring the distance between the points inside each cluster and compared to the threshold η.The threshold η for better discrimination is chosen to be η ≤ 0.002.

2016年10月，阿里巴巴整合优酷土豆、阿里影业、阿里音乐、阿里体育、阿里游戏等业务，成立阿里巴巴文化娱乐集团（即阿里大文娱），由俞永福出任首任董事长，杨伟东担任大优酷事业群总裁。而“终于可以休个假”的古永锵，变成了阿里大文娱战略投资委员会主席。

Accordingly,any points inside each cluster are accepted as PT only if the distance between any neighbor points is greater than the threshold.On the contrary,if the distance between any neighbor points inside each cluster is less than or equal to the threshold,the cluster is judged as the FT cluster.Fig.7 shows the final results.It is evident to see that from Fig.7,the proposed method successfully discriminates PTs from FTs and therefore the FTs can be completely suppressed.

Fig.7 Distribution of targets after the clustering approach

4.2 Monte Carlo simulations

Monte Carlo simulations are given to evaluate the performance of jamming suppression introduced by this method,and also to compare the proposed algorithm with some existing methods.

First,Monte Carlo simulation with 500 trails is given for analyzing the FTs suppression performance with the same parameters as above.It is worth noting that,if the FTs time delays are not well estimated,the suppression performance will deteriorate dramatically.Therefore,the probability of successfully preserving the PTs is mainly relying on the cross-correlation function output.

Accordingly,the root mean square error(RMSE)of estimating the FTs time delays is analyzed versus different numbers of JNR with various values of the JSR in Fig.8.As per simulation results in Fig.8,when the JNR is below 3 dB and JSR=0 dB,the proposed method cannot correctly estimate the FTs time delays,which implies that the method will not successfully suppress the FTs.Also,it is evident to see that when the JNR is greater than 3 dB,the obtained RMSE is approximately zero in all JSR values,which indicates that this method could successfully estimate the FTs time delay.The JNR is commonly higher to obtain better deceptive performance,which shows that the proposed method can sufficiently discriminate PTs from FTs,and therefore the deceptive FTs suppression can be performed.This is because higher JNR will lead to stable concentration of FTs and thus,better time delays compensation.

Fig.8 Estimated time delays RMSE versus JNR

Second,the probability of successfully preserving the PTs a function of the JNR is analyzed and compared with the coherent canceling method [15] in Fig. 9. In the simulation scenario,500 Monte Carlo simulations are conducted with the same parameters as above,but the false alarm probability Pfais set to be 10-4.It is seen that the probability of successfully preserving the PTs of the proposed coherent clustering method guarantees better performance when JSR=1 dB and JNR=3 dB or so.In contrast,the coherent canceling method with the same JSR and JNR suffers from severe performance degradation.It is worth noting that even if we increase the JSR to 3 dB,this method still does not achieve satisfactory performance.However,when SJR=10 dB,the performance slightly improves,but still cannot reach the performance of the proposed method.This is due to the fact that small change in FTs passions will lead to fatal results.

Fig.9 Probability of successfully preserving the PTs versus JNR

5.Conclusions

The multistatic radar system has valuable improved merits over active deception jamming because of its spatial diversity.In this paper,coherent clustering is applied to discriminate deceptive FTs from PTs,based on the diversities of the PTs and the consistency(the same source)of FTs after compensating the time delays of FTs.Accordingly,we implement the LDBSCAN algorithm to decompose the input data into an individual cluster based on their similarities.Then,we measure the distance between the points inside each cluster to discriminate PTs from FTs,in which PTs possess dispersion characteristic,but FTs are approximately concentrated in one possession.The numerical results have con firmed the feasibility and the effectiveness of the proposed method.

Acknowledgment

The authors express their gratitude to the editor and anonymous reviewers for their helpful comments and suggestions that improved this paper.

References

[1]BERGER S D.Digital radio frequency memory linear range gate stealer spectrum.IEEE Trans.on Aerospace and Electronic Systems,2003,39(2):725–735.

[2]ROOME S J.Digital radio frequency memory.Electronics&Communications Engineering Journal,1990,2(4):147–153.

[3]BAKER C J,HUME A L,BAKER C J.Netted radar sensing.IEEE Aerospace and Electronic Systems Magazine,2003,18(2):3–6.

[4]CHERNYAK V S.Fundamentals of multisite radar systems:multistatic radars and multistatic radar systems.London:CRC Press,1998.

[5]AKHTAR J.Orthogonal block coded ECCM schemes against repeat radar jammers.IEEE Trans.on Aerospace and Electronic Systems,2009,45(3):1218–1226.

[6]ABDALLA A,YUAN Z,RAMADAN M,et al.An improved radar ECCM method based on orthogonal pulse block and parallel matching filter.Journal of Communications,2015,10(8):610–614.

[7]ABDALLA A,YUAN Z,TANG B.ECCM schemes in netted radar system based on temporal pulse diversity.Journal of Systems Engineering and Electronics,2016,27(5):1001–1009.

[8]ZHANG J,ZHU D,ZHANG G.New antivelocity deception jamming technique using pulses with adaptive initial phases.IEEE Trans.on Aerospace and Electronic Systems,2013,49(2):1290–1300.

[9]XIONG W,WANG X,ZHANG G.Cognitive waveform design for anti-velocity deception jamming with adaptive initial phases.Proc.of the IEEE Radar Conference,2016:1–5.

[10]RAO B,ZHAO Y L,XIAO S P,et al.Discrimination of exoatmospheric active decoys using acceleration information.IET Radar Sonar&Navigation,2010,4(4):626–638.

[11]RAO B,XIAO S,WANG X,et al.Maximum likelihood approach to the estimation and discrimination of exoatmospheric active phantom tracks using motion features.IEEE Trans.on Aerospace and Electronic Systems,2012,48(1):794–819.

[12]XU J,LIAO G,ZHU S,et al.Deceptive jamming suppression with frequency diverse MIMO radar.Signal Processing,2015,113(C):9–17.

[13]ABDALLA A,WANG W Q,YUAN Z,et al.Subarray-based FDA radar to counteract deceptive ECM signals.EURASIP Journal on Advances in Signal Processing,2016,2016(1):1–11.

[14]DERHAM T E,DOUGHTY S,WOODBRIDGE K,et al.Design and evaluation of a low-cost multistatic netted radar system.IET Radar Sonar&Navigation,2007,1(5):362–368.

[15]WANG B,CUI G,ZHANG S,et al.Deceptive jamming suppression based on coherent cancelling in multistatic radar system.Proc.of the IEEE Radar Conference,2016:1–5.

[16]ZHAO S,ZHANG L,ZHOU Y,et al.Discrimination of active false targets in multistatic radar using spatial scattering properties.IEEE Sensors Journal,2016,16(8):2500–2508.

[17]ZHAO S,ZHANG L,ZHOU Y,et al.Discrimination of active false targets in multistatic radar using spatial scattering properties.IET Radar Sonar&Navigation,2016,10(5):817–826.

[18]XU R.Survey of clustering algorithms.IEEE Trans.on Neural Networks,2005,16(3):645–678.

[19]DUAN L,XU L,GUO F,et al.A local-density based spatial clustering algorithm with noise.Information Systems,2007,2007(32):978–986.

[20]GRECO M,GINI F,FARINA A.Radar detection and classification of jamming signals belonging to a cone class.IEEE Trans.on Signal Processing,2008,56(5):1984–1993.

[21]WANG W Q.GPS-based time&phase synchronization processing for distributed SAR.IEEE Trans.on Aerospace and Electronic Systems,2009,45(3):1040–1051.

[22]JOHNSEN T.Time and frequency synchronization in multistatic radar consequences to usage of GPS disciplined references with and without GPS signals.Proc.of the IEEE Radar Conference,2002:141–147.

[23]BREUNIG M M,KRIEGEL H P,NG RT,et al.LOF:identifying density-based local outliers.ACM Sigmod Record,2000,29(2):93–104.