Sensors & ransducers Vol. 70 Issue 5 May 204 pp. 65-70 Sensors & ransducers 204 by IFSA Publishing S. L. http://www.sensorsportal.co Multitarget Direction Measureent Based on Bistatic Radar Yubao ou School of Electrical & Inforation Engineering unan International Econoics University Changsha 40205 China el.: 86-073-88760386 E-ail: atlab_wjf@26.co Received: 5 March 204 /Accepted: 30 April 204 /Published: 3 May 204 Abstract: In recent years ultiple-input ultiple output (MIMO) radar has been widespread concern in the doestic and foreign researchers. Bistatic radar draws on the great success of MIMO technology in the counications field and it has any advantages over conventional radar. he direction angles estiations of bistatic MIMO radar are researched in this paper. o contrast traditional radar DOA estiates the direction vector of the bistatic MIMO radar is the Knonecker plot of the eission vector and reception vector that twodiensional direction angles is estiated. o solve this proble the principle of bistatic MIMO radar signal odel is in-depthly researched. By proposing Capon diensionality reduction ethod the two-diensional directions of the dual-based MIMO radar are estiated and coputer siulation is to verify the effectiveness of the ethod. Copyright 204 IFSA Publishing S. L. Keywords: Bistatic radar Capon diensionality reduction Multi-input ulti-output (MIMO) Direction-ofarrival (DOA) Direction-of-departure (DOD).. Introduction Multiple input ultiple output (MIMO) radar [] which has been proposed in recent years is new syste radar it draws on the hugely successful ultiinput ulti-output technology and has any advantages in the field of counication and it has been to show the potential and to ake an iportant contribution to the radar field. In recent decades foreign has introduced a variety of high-altitude highspeed aircraft and issiles low altitude cruise issiles (such as the oahawk cruise issile) stealth aircraft (such as F-7A B-2 F-35 the F-22 etc.) high-speed anti-radiation issiles (such as the AGM-88) and other high-perforance weapons and equipent and to create a surprise effect in the previous local wars and the tactics of odern warfare has been changed greatly. MIMO radar with its ulti-input ulti-output characteristics has a unique advantage in increasing space sapling antistealth and anti-electronic warfare and interference. In this coplex situation before the eney discovered and before the eney attack to ensure its own viability to grasp the electroagnetic power on the battlefield and to achieve effective onitoring and blocking of the eney target the radar is facing increasingly difficult tasks and challenges. For exaple the radar should have the ability to reote detection of weak target naely long-range ability to detect stealth targets. More radar syste has been need with a good position resolution to distinguish accurately the eney target position quickly and the eney is stroked in the positioning precision. herefore the target aziuth estiation perforance is an iportant part of a MIMO radar syste to detect the target inforation. Article nuber P_2052 65
Sensors & ransducers Vol. 70 Issue 5 May 204 pp. 65-70 Direction of arrival estiation (DOA) [2-5] is an iportant eleent of the array signal processing its application involves a nuber of ilitary and civilian econoic field of radar such as counications sonar seisic exploration radio astronoy and bioedical engineering. DOA estiation is to use a group sapling of source space in the tie doain and the airspace doain at the sae tie by a certain arrangeent of the sensors in different locations of the space and the orientation of the space source is estiated by analysis of the sensor array sapling data DOA estiation resolution accuracy robustness speed and can distinguish the target nuber etc. to deterine the aziuth estiation perforance is good or bad. MIMO technology [6-8] as an eerging ultichannel radar signal diversity spatial diversity has greatly enhanced the freedo of radar systes in signal processing. In order to guarantee the independence between the channels the transit signal of each transitting eleent are orthogonal each other. MIMO radar are coposition of the ultiple transit array eleents and the ultiple receive array eleent thus the target each array eission eleent excitation scattered field will be split in different spatial locations receiving array eleent while eission unit incentive scattered field of different locations will be superiposed on the sae receiver position the atched filter is applied to achieve signal sorting of different eission array eleent. hese characteristics have great potential in deterining the sapling proble of MIMO radar in space-related perforance. MIMO radar is increasing independent transit and receive eleent nuber a ore effective observation aperture have been produced with the two ultiplied cobination of the receiving and transitting antenna thus foring the equivalent observation aperture and the spatial sapling density. By DOA estiation theory these features can iprove the resolution and accuracy in target direction estiation and iprove the axiu nuber of resolvable target several ties. herefore DOA estiates has a very iportant theoretical and practical significance based on MIMO technology. Estiated (DOA) estiates the direction vector of bistatic MIMO radar (bistatic MIMO radar) to transit and receive direction vector of the Kronecker product with the direction of arrival of the traditional radar bistatic MIMO radar angle estiate [9 0] need to estiate the direction of arrival and direction of departure (DOD) the perspective of traditional radar estiates coplex and therefore Capon diensionality reduction ethod is proposed in bistatic MIMO radar transceiver angle estiation. 2. Bistatic MIMO Radar Model Bistatic radar technology [ 2] has been soe applied research in tactical radar and with the deands of odern warfare for radar and Rinpoche deand for radar itself battlefield survivability coupled with the rapid developent of digital circuit technology and bistatic radar technology is ore and ore attention in radar circles. he basic characteristics of bistatic radar is that sending equipent is separated fro receiving equipent (including antenna) the target and transceiver are located within the triangle plane. Copared with the onostatic radar bistatic radar structure configuration is ore coplex than the structure of the onostatic radar configuration. In order to achieve the correct positioning of the target there is the sending signals and receiving signals and postscript signal processing functions of the general base one in bistatic radar there is also the "three synchronization" proble of space tie and phase between sending and receiving. First sending and receiving bea space synchronous scanning. Since the basic features of bistatic radar sending and receiving beas are separated very far away then how to ensure that the process of scanning the airspace sending and receiving bea irradiation to the sae destination at the sae tie and with better data rate these ware a big technical probles. Second how to provide a unified high-precision tie reference for the transceiver which is the tie synchronization proble. ie synchronization issues are also to coplete the range and bistatic radar bea scanning synchronization depends on preise. Again for bistatic radar with pulse copression MI PD coherent signal processing etc. you ust also ensure that the phase synchronization between the transitter and receiver. he ain characteristics of bistatic radar can be suarized as the following three aspects: ) the sending and receiving devices separation; 2) the received signal is a non-backscattered echo objective; 3) triangulation positioning relationships. Forally based bistatic radar these three characteristics which leads to a bistatic radar at the tactical and technical perforance with soe unique advantages. igh easureent accuracy is achieved by using a narrow bea in precision-guided conventional radar the search capacity is often weak bistatic MIMO radar is using broad eission bea receiving siultaneous ultibea by using DBF technology the use of long accuulated technology to copensate for the power loss. Which shortens the search tie of the entire airspace but also to ensure the tracking and high-precision easureents on ultiple targets in a short period of tie the antisaturation attack will be greatly enhanced. Meanwhile the bistatic MIMO radar can siultaneously access the target relative to the point of view of the transceiver array and transceiver arrayto-target distance and the characteristics of inforation surplus and therefore has unique antideception jaing capabilities. Analysis of bistatic MIMO radar signal odel the array configuration is shown in Fig.. 66
Sensors & ransducers Vol. 70 Issue 5 May 204 pp. 65-70 eritian transpose. In the case of P target () can be odified to: Y = ArDΦ+ N (2) Fig.. Bistatic MIMO radar array configuration diagra. his is a narrow band of bistatic MIMO radar syste with M transit dense antennas and N receiving antennas densely. ransitter launches M quadrature encoder signals for the transitted signal vector s() t = s() t s2() t... sm () t. Assuing that the aperiodic autocorrelation cross-correlation signal sidelobes is very low even if the Doppler shift present. he encoding of the launching baseband signal is recorded as K ss C represents the -th transitted signal s s = K the use of binary sequences with zero correlation zone. And high Doppler frequency is still low in the zero correlation zone of the autocorrelation and cross-sidelobe. Doppler frequency is alost no effect on the orthogonality of the wavefor the wavefor approxiation retains the orthogonality of the target of large Doppler frequency. Doppler frequency can be caused by a variety of pulse and can be ignored. ere we assue that all targets are located in an adjacent range both the goal of zero correlation zone so the targets within the sidelobe can be ignored. Suppose a target is located in the (ϕ θ ) which ϕ is the desired launch target angle (defined as the DOD) and θ is the desired acceptance target angle (defined as of DOA). Data which is received by the target launch to reach the receiver array can be expressed as the following expression: 0... Y = a ( θ) a ( ϕ) β Ke + Zs K... 0 = a a ϕ β Ke + N fdtl r t ( θ) ( ) fdtl r t Zs () where N at ϕ is the -th K eleent in the launch of the array steering vector. N N C is defined as the -th baseband transit = and ( ) signals to atch the filter noise vector ( ) is the where Ar = ar( θ)... ar( θp) D ( )... = diagat ϕ at( ϕp) β Φ=... β P Ke Ke fdtl fdptl It is assued that the different objectives have different Doppler frequency and all the P target in the sae range interval. 3. Capon Down-diensional Method 3.. heoretical Analysis he full nae of the Capon algorith is the Capon's iniu variance algorith. he ain bea is fored using a part of the degree of freedo in the direction of the desired user but also zero point is fored with the reaining degree of freedo in the direction of the interference signal. he advantage of the Capon algorith is to ake a iniu power which is contributed by noise and any interference fro the non-source direction but they can keep power unchanged in the direction of the source signal. he features of the Capon algorith are adaptive interference cancellation and the nuber of interference cancellation by the array geoetry constraints and resolving power depends on the array geoetry and signal to noise ratio. Between Direction-of-arrival (DOA) of the bistatic MIMO radar and Direction-of-departure (DOD) the angle estiation are discussed and a diension reduction Capon algorith is proposed. his algorith only requires a one-diensional search you can avoid the two-diensional Capon (2D-Capon) high coputational cost of the algorith and prove that the algorith can be very good in ters of perforance and be better than the two-diensional Capon algorith. Multiple-input ultiple-output (MIMO) radar [-3] is spreading a variety of wavefor siultaneously by using a plurality of antennas while receiving the reflected signal in a siilar anner. MIMO radar Direction-of-departure (DOD) and the direction of arrival (DOA) will be conducted in-depth study. he two-diensional Capon algorith (2D-Capon) is an ipleentation of the algorith of the MIMO radar DOA and DOD estiation; owever the two-diensional search requires a higher coputational coplexity. 67
Sensors & ransducers Vol. 70 Issue 5 May 204 pp. 65-70 Diensionality reduction in the bistatic MIMO radar angle is estiated Capon algorith [4 5] which obviously reduces the coplexity. Consider bistatic MIMO radar systes: its eission arrays and receiver arrays are unifor linear array respectively and n are ordered adjacent antenna spacing half between the wavelength of the transitting array and receiving array. In addition to assue that K non-related goals through the atched filter output at the receiving end can be expressed as: ( φ ) ( θ ) ( φ ) ( θ )... ( φ ) ( θ ) X = ar at ar 2 at 2 ar K at K B + W (3) where θk and φk are K target noral eission arrays and array of launch angle and acceptance L K angle. B C wavefor is caused by the Doppler frequency in the nuber of snapshots K target sources include phase and aplitude and the agnitude of the ain subject such as reflection coefficient of the transit gain and receive gain path fading and other losses. ar( φk) = exp( jπsin φk)...exp ( jπ( N ) sinφk) at( θk) = exp( jπsin θk)...exp( jπ( M ) sinθk) a r ( φ k ) and at ( θ k) are respectively receive steering vector φ k and transit steering vectorθ k. W is to receive an additive white Gaussian noise atrix in receiving end. indicates plot for Kronercker. For the signal odel (3) we can see that the covariance atrix R x can be the origin of estiated by Rx = XX / L so we can construct twodiensional Capon spatial spectru: f capon ( φθ ) = a a R a a ( φ) ( θ ) ( φ) ( θ ) r t x r t where the K peak of f ( ) capon (4) φ θ is the target DODs and DOAs. he two-diensional Capon requires a detailed two-diensional search due to the high coputational cost the efficiency of doing so is relatively low. 3.2. Capon Down-diensional Method Definition: V φ θ ) = [ a ( φ) a ] R [ a ( φ) a ] ( r t K r t It can also be expressed as ( φ θ) = t( θ) r( φ) M x r( φ) M t( θ) a ( θ) Q( φ) a ( θ) V a a I R a I a = t t ( φ ) ( φ ) ( φ ) Q = a I R a I r M x r M (5) Equation (5) is a quadratic optiization proble. In order to eliinate the zero solution we also e a θ = consider constraints ( ) joined here e = [ 0...0] t to have been. his optiization proble we can use the linear constraint iniu variance to rebuild get: ( ) ( ) ( ) ( ) in a θ Q φ a θ ste.. a θ = (6) φ t t t Using the Lagrange ultiplier then equation (6) becoes: e Q( φ) e φ e ˆ φ = argin = argax φ e Q ( φ) And then search [ 90 90 ] ( ) (7) Φ you can get Q φ the K largest peaks of each eleent(). K largest peaks correspond to the requireents bistatic MIMO radar direction of arrival (DOA). On equation (5) you can also consider this: ( φ θ) = r( φ) N t( θ) x N t( θ) r( φ) ar( φ) P( θ) ar( φ) P( θ) = I a ( θ) R I a ( θ) V a I a R I a a = N t x N t (8) Siilarly using the Lagrange ultiplier so the solution becoes: ( θ) ˆ θ = arg ax e P e θ 2 2 [ ] 2 0...0 N e = R (9) Search θ [ 90 90 ] we can get the K axiu peak of P ( θ ) () eleent which is corresponding to the double base MIMO radar waves away fro the direction (DOD). Bistatic MIMO radar Capon algorith based on DOD and DOA estiation diensionality reduction algorith ain steps: ) covariance atrix R x ; 2) φ Q( φ ) s () Each eleent of the K largest peaks is obtained by (7) which can be double base MIMO radar DOA estiates; 68
Sensors & ransducers Vol. 70 Issue 5 May 204 pp. 65-70 3) θ by (9) P ( θ ) s () eleent of K largest peak are obtained which corresponds to DOD estiates. 4. Siulation of Experiental ests Uniforly spaced linear array with one pair of bases MIMO radar transceiver arrays are arranged in half-wavelength only the transitting array into M = 8 to receive the array of N = 8. Eission the orthogonal ardard code wavefor at the transitter sensor and signal to noise ratio for the SNR = 0 db the nuber of snapshots L = 00. If we assue that there are four goals in the space ( ϕ θ ) = ( 0 20 ) ( ϕ2 θ 2) = ( 0 30 ) ( ϕ θ ) = ( ) ( ) ( ) 3 3 40 30 ϕ4 θ 4 = 40 50 this is the case under the siulation of W Gaussian white noise with zero ean. And we ake the signal to noise ratio changes fro 5 db and 0 db to 30 db interval 5 db. 00 Monte Carlo experients to define the angle estiated ean square error (RMSE) as follows: Fig. 3. Direction of departure (DOD). RMSE 2 Lc θ ˆ θ 0 (0) = = Lc where Lc is the nuber of Monte Carlo experients. Fig. 2 and Fig. 3 of this algorith four pairs based MIMO radar target DOA and DOD estiates. he value of the DOD and DOA can clearly been understood fro two ap peaks. his shows that in the case of Gaussian white noise the ulti-objective point of view the joint estiation can effectively been achieved by the ethod. Fig. 4. Joint estiation of DOA and DOD. Fig. 5. Angle search. Fig. 2. Direction of arrival (DOA). he figure is our bistatic MIMO radar algoriths for four goals DOA and DOD estiates. he peak value of DOD and DOA can be clearly understood fro Fig. 4. DOA and DOD is a perspective value atching ethod in Fig. 5. 5. Conclusion and Outlook Multi-objective plubing of coherent bistatic MIMO radar are researched based on the sae phased array syste in the paper. he ain features of bistatic radar is that space between stations is sall the rays fro ultiple transceivers array eleent to the target is approxiately parallel the 69
Sensors & ransducers Vol. 70 Issue 5 May 204 pp. 65-70 target relative to the transceiver array has the sae DOA and for narrow-band transit signal envelope delay between array eleents can be ignored. he angle estiation of Direction-of-arrival (DOA) and Direction-of-departure (DOD bistatic MIMO radar are discussed in the article and a less-diensional Capon algorith is proposed. he proposed algorith requires only one-diensional search it can avoid the high coputational cost of twodiensional Capon (2D-Capon) algorith and show that the algorith is better than the two-diensional Capon algorith in ters of perforance. References []. e Zi-Shu an Chun-Lin Liu Bo MIMO radar and its technical characteristic analyses Acta Electronica Sinica Vol. 33 Issue 2A 2005 pp. 244-2445. [2]. Xu ong-bo Wang uai-jun Lu Min Su Yi A new algorith on estiation of DOA using MIMO radar Journal of National University of Defense echnology Vol. 3 Issue 3 2009 pp. 92-96. [3]. Xu ong-bo Wang uai-jun Lu Min et al Research on field experient of MIMO radar iaging and DOA Systes Engineering and Electronics Vol. 32 Issue 4 200 pp. 754-758. [4]. Wang Ju-ing Jiag Sheng-Li e Jin Liu Zhong Subspace-based DOA estiation in ipulsive noise environents for MIMO radars Journal of Astronautics Vol. 30 Issue 4 2009 pp. 653-657. [5]. Wang Ju-ing Jiang Sheng-Li Liu Zhong. Craer- Rao bounds of DOA estiation for MIMO radars in copound-gaussian clutter Journal of Electronics & Inforation echnology Vol. 3 Issue 4 2009 pp. 786-789. [6]. Dai Xizeng Peng Yingning ang Jun Detection perforance of MIMO radar Journal of singhua University (Science and echnology) Vol. 47 No. 2007 pp. 88-9. [7]. Qu Jin-You Zhang Jian-Yun Liu Chun-Sheng he detection perforance of MIMO radar with arbitrarily correlated wavefors Journal of Circuits and Systes Vol. 4 Issue 2 2009 pp. 68-73. [8]. Wang Ju-ing Jiang Sheng-Li e Jin Liu Zhong Generalized likelihood ratio detector for airborne MIMO radars Journal of Electronics & Inforation echnology Vol. 3 Issue 6 2009 pp. 35-38. [9]. A. Ghobadzadeh A. A. adaion M. R. aban GLR approach for MIMO radar signal sapling in unknown clutter paraeter in Proceedings of the International Syposiu on elecounications ehran 2008 pp. 624-628. [0]. A. Sheikhi A. Zaani Y. Norouzi Model-based adaptive target detection in clutter using MIMO radar in Proceedings of the CIE International Conference on Radar Shanghai 2006 pp. -4. []. A. Sheikhi A. Zaani Coherent detection for MIMO radars in Proceedings of the IEEE Radar Conference Boston MA 2007 pp. 302-307. [2]. L. Z. Xu J. Li Iterative generalized likelihood ratio test for MIMO radar IEEE ransactions on Signal Processing Vol. 55 Issue 6 2007 pp. 2375-2385. [3]. J. Li P. Stoica L. Z. Xu et al On paraeter identifiability of MIMO radar IEEE Signal Processing Letters Vol. 4 Issue 2 2007 pp. 968-97. [4]. L. Z. Xu J. Li P. Stoica Adaptive techniques for MIMO radar in Proceedings of the 4 th IEEE Workshop on Sensor Array and Multichannel Signal Processing Waltha MA 2006 pp. 258-262. [5]. I. Bekkeran J. abrikian arget detection and localization using MIMO radars and sonars IEEE ransactions on Signal Processing Vol. 54 Issue 0 2006 pp. 3873-3838. 204 Copyright International Frequency Sensor Association (IFSA) Publishing S. L. All rights reserved. (http://www.sensorsportal.co) 70