A Decision-Based Model and Algorithm for Maneuvering Target Tracking

Size: px
Start display at page:

Download "A Decision-Based Model and Algorithm for Maneuvering Target Tracking"

Transcription

1 WSEAS RANSACIONS on SYSEMS A Decision-Based Model and Algorith for Maneuvering arget racking JIAHONG CHEN ZHONGHUA ZHANG ZHENDONG XI YONGXING MAO China Satellite Maritie racking and Control Departent, Jiangyin, Jiangsu, 4 43 stone_cjh@sina.co; zhangzh_stc@sina.co; dong369@hotail.co; yx88@o.co Abstract: he decision-based odels for aneuvering target tracking were studied in this paper. Focusing on the proble of dissatisfied with single odel tracking and the optial odel-set is difficult to design of the ultiple-odel (MM) algorith, we odified the current odel, and proposed an adaptive single odel (ASM) to track angular otion. he united velocity-acceleration estiation-based and direct easureentbased decision algoriths for odel-switching were provided and discussed. At the sae tie, we discussed the question of the thresholds setup, and gave a significance test ethod based on hypothesis testing theory. Siulation results show that the tracking perforance of this new adaptive single odel is stationary, and is iproved uch ore than the current odel, also better than the interacting ultiple-odel (IMM) for strong aneuvering target tracking. Furtherore the coputational load of the new odel can be less than that of IMM. Key-Words: adaptive single odel, ultiple-odel, aneuvering target tracking, current odel, decisionbased algorith, odel-switching Introduction he key to successful target tracking lies in the effective extraction of useful inforation about the target s state fro observations. A good odel of the target will certainly facilitate this inforation extraction to a great extent. [-6] gave a coprehensive and up-to-date survey of the techniques for tracking aneuvering target, especially of the tracking odels. Interrelationships aong odels and insight to the pros and cons of odels were provided by []. In the history of the developent of aneuvering target tracking (M) techniques, single odel based adaptive Kalan filter free of decision cae into existence first. As one single odel can only atch one otion ode best, and the perforance of tracking tie-varying or strong aneuvering target is dissatisfied, so the decisionbased algorith was proposed, by [4], which belonged to the switching-odel approach in a broad sense. he followed approaches of ultipleodel (MM) algoriths have becoe quite popular for M, aong which the interacting ultipleodel (IMM) is ore preferable [5-0]. he universal aneuvering target tracking techniques based on single odel are adaptive Kalan filtering algoriths [,, 3], which faced the difficulties to find a suitable bank of paraeters to atch the syste otion ode precisely. he algoriths, which use statistical ethods with ulti-banks of paraeters to describe the otion ode approxiately, were called MM algoriths. o iprove the tracking perforance of MM will face to coplex odel-set designing and burdened calculating. It was shown that the syste is optial only when there is one odel in the odel-set and the odel also ust atch the otion ode copletely by [8]. Here we will study the decision-based algoriths and try to avoid the coplex designing of the odel-set and to find a ore suitable single odel for M. As surveyed by [4], for M, there are three ain classes of decision-based techniques: equivalent noise, input detection and estiation, and switching odel, which are based on decisions regarding to target aneuver. In a broad sense, algoriths in the equivalent noise or input detection and estiation approaches also belong to the switching odel approach since following different decisions these algoriths taking different actions, which ay be construed as filtering based on different odels. he fundaental proble is the detection of aneuver onset and terination. wo popular choices for the detection ter are the ISSN: Issue 8, Volue 8, August 009

2 WSEAS RANSACIONS on SYSEMS easureent residual z and the input estiate u, others are based of the or of there cobined functions. In the open literature available to us, one decision-based algorith usually uses only one detection ter, which can not be ore adaptive to the tie-varying aneuver generally. Moreover there are only two single odels in essence which are switched each other, e.g., between CV (constantvelocity) and CA (constant-acceleration) [4] or between CV and Singer (Singer Acceleration) Models [5], etc. he ethods can generally be called as estiation-based ethods where u and z or v (velocity) are all estiation-based. he precision of estiation is relative to the syste noise (process noise and easureent noise) and odel reliability. Generally the infection of noise can be decreased by odel filtering. We can also take the direct easureent as detection ter which will be siple. Furtherore the validity of odel is irrelative to the decision, whereas the noise infection will be increased. In this paper we intend to propose a united velocity-acceleration estiation-based decision algorith (EBDA) and a direct easureent-based decision algorith (MBDA) with a new single united aneuvering and nonaneuvering odel. he rest of the paper is organized as follows. Section presents a new adaptive odel. he united velocity-acceleration estiation-based decision algorith and thresholds setup are presented in Section 3 and 4. Section 5 gives the easureentbased decision algorith with the new odel. Coparison of the new algoriths and IMM for siulated different trajectories is given in section 6. Concluding rearks are provided in Section 7. he New Adaptive Single Model. he Singer Model and Current Model he CV odel is ore proper for non-aneuvering target, while the CA odel is ore suitable for the otion whose acceleration derivative (i.e., jerk) a(t) is an independent process (white noise) w(t) : a ( t) = w( t). he Singer odel assues that the target acceleration a(t) is a zero-ean first-order stationary Markov process with autocorrelation α τ R a ( τ ) = E[ a( t + τ ) a( t)] = σ e. Such a process a(t) is the state process of a linear tie-invariant syste: at () = αat () + wt () () where a = x, is the acceleration; σ = E[ a( t) ] is the instantaneous variance of the acceleration; α = / τ is the reciprocal of the aneuver tie constant τ and thus depends on how long the aneuver lasts. For exaple for an aircraft, τ 60s for a lazy turn and τ 0 0s for an evasive aneuver, for the issile τ aybe a little longer []; w( t ) is zero-ean white noise with constant power spectral density S w = ασ. As the aneuver tie constant τ increases or decreases the Singer odel corresponds to a otion in between of (nearly) CA and (nearly) CV. It should thus be clear that the Singer odel has wider coverage than CV and CA odels [, 6]. Another acceleration odel, called the current odel by its authors, proposed by [6], is in essence a Singer odel with an adaptive ean of acceleration that is ean-adaptive acceleration (MAA) odel: at () = αat () + αat () + wt () () With state vector x = [ x x x], where the superscript denotes the transpose of atrix, the state-space representation of this odel is: x () t = 0 0 () t 0 a() t 0 x + + w () t (3) 0 0 α α excluding the tie instants at which saples are taken since a ( t ) is assued piecewise constant. he discrete-tie equivalent is: α ( + α + e )/ α α xk+ = 0 ( e )/ α xk + α 0 0 e α ( /) ( + α + e )/ α α ( e ) / α ak + wk α e = Fαxk + Bαak + wk (4) where is the sapling interval. As α <<, the covariance of the noise is given by: w k /0 /8 /6 4 3 Qk = ασ /8 /3 / (5) 3 /6 / A key underlying assuption of the current odel given by [6] (but not so stated explicitly) is that ak+ = aˆ k k, that is to use the last estiation as the current ean. his is questionable and can be actually avoided [, 6]. Getting the expectation of the last ite of (4), we have: ISSN: Issue 8, Volue 8, August 009

3 WSEAS RANSACIONS on SYSEMS α α ak+ = E[ ak+ zk] = e E[ ak zk] + ( e ) ak α α = e aˆkk + ( e ) ak = aˆk + k where z k is the easureent vector. (6). Modification to the Current Model Application of the current odel can be found in the literatures. A weighted current odel adaptive Kalan filtering algorith is suggested by [7], while an iproved adaptive filtering algorith of noise variances self-adaptation was given by [8] based on the current statistical odel. Here we take into account the atching proble of the odel. he process of target tracking is the servo syste of radar driving antenna to point to the target with tie-varying angular velocity and acceleration controlling instruction. By the sense of physics, it is viable to use ean-adaptive acceleration odel to atch the relative angular otion of the target. Actually, the angular velocity and acceleration are often tie-varying for radar target tracking, especially for rectilineal otion with constant velocity the angular otion is not constant. As τ 0s and the sapling interval is sall ( 0.5s, it s possible), expand e α of the current odel (4) with aylor series: α 3 e = α + ( α) ( α) + (7)! 3! Substituting (7) into (4), ignoring the high order ite, we can obtain nearly current odel: 3 3 / α /6 α /6 xk+ = 0 α / x k + α /ak + wk 0 0 α α (8) It s not difficult to find that the result of (6) is also correct for (8). As the new odel should be universal and adaptive, we odified the odel of (8) by iporting the adaptive factor λ : 3 3 λ( / α /6) α /6 xk+ = 0 λ( α /) xk + α / ak + wk 0 0 λ( α) α (9) Denotes (9) as: x k+ = Fk x k + Bkak + w k (0) where λ only can be 0 or, α is also a fixed value (in this paper we choose 0.) or zero. Analyzing the odel of (9), we can see that: ) When α and a k are not equal to zero, and λ =, the syste is nearly equivalent to MAA odel; ) When α 0 and λ = but a k = 0, the syste is nearly equivalent to Singer odel; 3) When α = 0 and λ =, the syste is nearly equivalent to CA odel but for rando noise ite; 4) When α = 0 and λ = 0, the syste is nearly equivalent to CV odel but for rando noise ite. We call the universal odel of (9) as iproved adaptive single odel (ASM). It s not difficult to find that if the paraeter α and λ varying in real tie, this new odel can atch constant-velocity, constant-acceleration and tie-varying acceleration otion, so it has strong adaptability. 3 United Estiation-Based Decision Algorith 3. Adaptive Rules o atch the target otion exactly, ASM should be switched in real tie between MAA, CA and CV according to the easureents and states estiation. We will design the adaptive rules with united velocity-acceleration estiations by distinguishing three kinds of situations. he ean (average) and odified standard deviation (MSD) of acceleration and velocity in continuous steps are: µ a, = aˆ k ik i () i= / σ a, = ( ( aˆ k ik i µ a, ) ) i= µ v, vˆ k ik i i= / σv, = ( ( vˆ k ik i v ) ), i= k () = (3) µ (4) where denotes current step of filtering and estiating. Here can not be too large because of aneuver will take place at any tie, so we use the odified standard deviation instead of the standard deviation by statistical theory. he adaptive rules for odel-switching are followed as: ) If the syste is under MAA odel currently ( λ =, α 0 ) (a). Meeting the condition σ a, < δ, i.e., when the MSD of acceleration of the final steps is sall enough, the odel should be switched to CA odel, which eans λ =, α = 0. ISSN: Issue 8, Volue 8, August 009

4 WSEAS RANSACIONS on SYSEMS (b). Meeting the condition µ a, < δ and σ v, < δ 3, i.e., when the MSD of velocity of the final steps is sall enough and the ean of acceleration is near to zero, the odel should be switched to CV odel, which eans λ = 0, α = 0. ) If the syste is under CA odel currently ( λ =, α = 0 ) (a). he (b) of () is still applicable. (b). Meeting the condition σ a, > δ 4, i.e., when the MSD of acceleration of the final steps is large enough, the odel should be switched to MAA odel, which eans λ =, α 0. 3) If the syste is under CV odel currently ( λ = 0, α = 0 ) Meeting the condition σ v, > δ 5, i.e., when the MSD of velocity of the final steps is large enough, the odel should be switched to CA odel, which eans λ =, α = 0. If the conditions are not et, the odel should be hold. 3. State Estiation in Spherical Coordinate-Syste Many studies of target tracking are focused on describing the otion ode in inertial Cartesian coordinate-syste (CS). But radar syste can only provide the bearing (or aziuth) b and elevation e easureents, and possibly soe radar can also provide range r or range rate (Doppler) r (in spherical CS). o study target otion in the Cartesian CS is the direct ethod, but the coplex nonlinear odel ust be introduced because of no x, y and z easureents in Cartesian CS. All kinds of approxiation and transforation will bring on bias of angular controlling instruction. We consider that studying independently of radar tracking in spherical CS fro state estiation in Cartesian CS is doable, because in spite of any way the radar and target oving, the essence of tracking is to ascertain the relative bearing and elevation angular otion of the target proptly and exactly. An angular error tracking filter and odel was researched by Ekstrand B. in [9], where only the nearly constant-velocity (CV) odel was used to estiate the inertial angular rate of the line-of-sight (LOS), whereas no optial angle estiation directly. Here we will estiate the angle directly. Because the range r, bearing b and elevation e in spherical coordinate-syste of the radar syste can be considered as independently, and the state and easureent equations of the three channels are the sae, so we can filter and estiate the three channels independently with the sae algorith. he state equation of the syste is just like (9), whereas the easureent equation is: zk = Hxk + vk (5) = [ 0 0] xk + vk where z k denotes the easureent and v k is the easureent noise, the covariance of vk is Rk. We can see that the odel is linear, despite it is tievarying, so the optial Kalan filter can be used here. o take the bearing b channel for exaple, denotes x = b b b, than the EBDA for state estiation ainly consists of two steps: value F k, B k by the adaptive rules and estiate the state with Kalan filter. he coplete steps for the state estiation algorith are concluded as in able. We proposed the new odel based on angular otion of radar target tracking, which can atch the unifor otion and variable otion together. It is suitable for range channel at the sae tie too, which is just the exhibition of the strong adaptability of the new odel algorith. able he EBDA of ASM for State Estiation. value Fk and Bk by the adaptive rules. odel-conditioned filtering (if switching filter fro k ) Predicted state: xˆk+ k = Fkx ˆkk + Ba k k Predicted easureent: z ˆk + ˆ k = Hx k + k Predicted covariance: Pk+ k = FP k k kfk + Qk Measureent residual: z k+ = zk+ z k+ k Residual covariance: Sk+ = HPk+ khk + Rk+ Filter gain: Kk+ = Pk+ kh Sk + Updated state: xˆk+ k+ = xˆk+ k + Kk+ z k+ Updated covariance: Pk+ k+ = Pk+ k Kk+ HPk+ k 4 hresholds Setup he key to successful perforance iproving of ASM tracking lies in the exact setup of the thresholds δ, δ, δ 3, δ 4 and δ 5. Besides experient and experience, we will analyze in view of probability and statistics theory. 4. hresholds of Variable Motion ake the statistic of acceleration for exaple. If the syste oves with constant acceleration, we can ISSN: Issue 8, Volue 8, August 009

5 WSEAS RANSACIONS on SYSEMS assue the acceleration has Gaussian distribution N( µ a, σ a ). he first oent (ean) µ a varies by the setup, whereas the second oent (covariance) σ a is relatively fixed, and depends on the perforance of the syste, such as sapling interval, easureent precision, etc.. We also can say that σ a is deterined by the syste, and can be obtained approxiately by experient sapling. If σ a, σ a obviously under MAA odel, we can consider the acceleration is nearly constant reasonably; on the other hand, if σ a, > σ a obviously under CA odel, odify the odel to MAA is also believable. Definition : assue F(y) is the distillation function of rando variable Y, p is a real nuber, 0 p, if: P{ Y yp} = F( yp) = p (6) where P{ } stands for probability. We call y p the p percentile of the distribution. o ascertain δ and δ 4 is a proble of hypothesis testing. ) Inferior (lower) liit δ We establish the original hypothesis H 0 : σ a, σ a against the alternative hypothesis H: σ a, < σ a. Definition : ype error is the error ade when rejecting H 0 while H 0 is true. ype error is the error ade when accepting H 0 while H is true. o control the risk of istaking σ a, < σ a when σ a, σ a actually, i.e., the ype error is less than β (significance level). According to the Gaussian distribution of a, the rando variable ( a ˆk ik i µ a, ) σ a i= is chisquare distributed ( χ ( ) ), so the rejection region W (rejecting H 0 and accepting H ) is: { ( aˆk k,, aˆk k ) : ( ˆ ak ik i µ a, ) σ a i= σ a, χ ( β ) } a ( ) = σ (7) χ β ( ) is the β percentile of χ ( ) distribution. If β is sall, rejecting H 0 and accepting H is credible by event with sall probability should not take place in one experient. χ Here we take β = 0. and = 5, query the distribution table, χ 064, so when 0. ( 4 ) =. σ a, 0. 56σ a, (7) coes into existence. hat is to say we can set δ = σ a /. ) Superior (upper) liit δ 4 Here we should control the risk of istaking σ a, > σ a when σ a, σ a actually, so we establish the original hypothesis H0: σ a, σ a against the alternative hypothesis H: σ a, > σ a. he rejection region W of the rando variable ( a ˆk ik i µ a, ) is: σ a i= aˆk k aˆk k ˆ ak ik i µ a, σ a i= ( ) σ a, = > χ β ( ) } σ a { (,, ) : ( ) (8) β, then χ 0 = 7.779, so when If = 0..9(4) σ a, >. 395σ a, (8) coes into existence. hat is to say we can set δ4 =. 395σ a. If we set δ 4 = σ a, then β < , so the reliability of H : σ a, > σ a is very high ( > ). In brief, we can set δ near to σ a / and δ 4 near to σ a fro the above analysis. In fact the setup of δ and δ 4 should also accord with the experient results. wo questions ust be paid attention to. he first is that σ a / and σ a are only reference values because can not be too large by the realtie requireent. he deterinistic value ust be experienced typical experients. Second, σ a, is relative to the sapling interval, also sae with σ a, so when is varied the thresholds should be varied accordingly. 4. hresholds of Unifor Motion he ethods of setup of δ 3 and δ 5 reseble to the above δ and δ 4. Assuing the syste is under CV odel, and the distribution of velocity is N( µ v, σ v ), so δ 3 can be near to σ v /, and δ 5 can be near to σ v. he atching perforance of CV odel is inferior to that of CA and MAA odel except that the syste is actually oving with constant velocity. Switching the CA and MAA odel to CV odel ISSN: Issue 8, Volue 8, August 009

6 WSEAS RANSACIONS on SYSEMS should be cautious, so we appended the constraint of acceleration ean: µ a, < δ. he physical eaning of this condition is obvious. If the acceleration noise is N(0, σ a ), we establish the original hypothesis H0: µ a, = 0 against the alternative hypothesis H: µ a, 0. he rando variable µ a, 0 σ a is standard Gaussian distribution N(0,), so the confidence interval W0 should eet: µ a, σ a < u β / (9) where u β / is the percentile of Gaussian distribution, β = 0. as the above tests, so we have: u β / u σ a µ a, < σa = σa = σa = (0) δ is near to σ a /. 36. Different fro the above tests, accepting H 0 should be credible here. Here we should also control the ype error, i.e., control the risk of istaking µ a, = 0 when µ a, 0 actually. As can not be too large, according to the hypothesis testing theory we can increase β. In order to be ore credible of H 0 : µ a, = 0, δ can be set between ( ~ )σ a 0 conservatively by experient. By the way, the short sapling interval and suitably large can iprove the stability of statistical data, and is of advantage to atch otion ode in real-tie. 5 Direct Measureent-Based Decision Algorith and Discussion 5. Adaptive Rules he first-order difference of position ( ) is z,k relative to the target velocity. he second-order difference of position ( z,k ) is relative to the target acceleration, and the third-order difference of position ( z 3,k ) is relative to the target acceleration derivative. Denote: z, k = zk+ z k () z, k = z, k+ z, k () z3, k = z, k+ z, k (3) Regarding of the randoicity of easureent, the average (ean) of the n th-order difference in steps is used here (denote as z,k, z,k and z 3,k ). We give the thresholds of velocity η and η, and of acceleration η3 and η4. ) If the syste is under MAA odel currently ( λ =, α 0 ) (a). Meeting the condition z 3, k < η 3 and z, k < η, i.e., when the acceleration and its derivative of the final steps are sall enough siultaneously, the odel should be switched to CV odel, which eans λ = 0, α = 0. (b). Meeting the only condition z 3, k < η 3, i.e., when the acceleration derivative of the final steps is sall enough, the odel should be switched to CA odel, which eans λ =, α = 0. ) If the syste is under CA odel currently ( λ =, α = 0 ) (a). Meeting the condition z 3, k > η 4, i.e., when the acceleration derivative of the final steps is large enough, the odel should be switched to MAA odel, which eans λ =, α 0. (b). Meeting the condition z 3, k < η 3 and z, k < η, i.e., when the acceleration and its derivative of the final steps are sall enough siultaneously, the odel should be switched to CV odel, which eans λ = 0, α = 0. 3) If the syste is under CV odel currently ( λ = 0, α = 0 ) (a). Meeting the condition z, k > η and z 3, k > η 4, i.e., when the acceleration and its derivative of the final steps are large enough siultaneously, the odel should be switched to MAA odel, which eans λ =, α 0. (b). Meeting the only condition z, k > η, i.e., when the acceleration of the final steps is large enough, the odel should be switched to CA odel, which eans λ =, α = 0. he estiation steps and thresholds setup of this MBDA are siilar to that of EBDA in able. 5. Discussion he new algoriths of ASM can switch to the better atching odel based on the analysis of the state estiation or the direct easureents. If the real target trajectory is taken as an unknown coplex curve, than the new ASM will fit the ode coprehensively with first degree, quadratic and cubic curves. It is ore adaptive than the traditional single odels or switching-odel approaches only with first degree and quadratic curves. If the atching odel of ASM is appropriate, than the tracking perforance will be iproved especially ISSN: Issue 8, Volue 8, August 009

7 WSEAS RANSACIONS on SYSEMS for the strong aneuvering targets. he siulation results have shown this well. We give a decision-based switching-odel approach for M, which is different fro the traditional switching-odel approaches. First, the nearly n th-order ( n =,, 3 ) odels were used to fit the otion ode, the fitting perforance of the new algorith should be better. Second, the switched odels have a unifor expression, odelswitching are realized by revaluing the paraeters. hird, the union of velocity, acceleration and jerk is taken as the detection ite rather than that only one of the is used in the traditional switching-odel approaches. Forth, the physical sense of the detection ter and the thresholds are obvious and understandable. One iportant assuption is that if the syste oves with constant acceleration (or constant velocity), the acceleration (or velocity) has Gaussian distribution. In the new algoriths, if the ode variation was detected at tie k, the odel should be switched fro k. When is sall enough (< τ ), switching fro k or k has no obvious difference. Switching fro k can lighten the coputational load. On the other hand if τ is saller, the aneuver acceleration can be considered as noise. For EBDA, considering the statistical reliability of the detection ters, the step can not be too sall. With the acceleration to be estiated, > 3. In view of real-tie coputation, can not be too large. Generally when <= 0, the coputational load is nearly approxiate to the IMM algorith. For the coputational syste with high perforance, if is very sall, than can be larger. For MBDA, the third-order difference should be calculated, then to copute the average. he covariance of the n th-order difference is in (direct) n n σ σ z proportion to (it can be proved about z, is the variance of easureent z ). should be large enough to reduce the infection of noise, > 5 generally. Sae as the EBDA also can not be too large ( < τ ). he coplexity of the two algoriths corresponds to each other. Fro the siulation results we can see that the tracking perforance of MBDA+ASM is also satisfied, which shows the strong adaptability of ASM. 6 Siulation In this section, we copared the perforance of ASM with EBDA to the classic MAA and IMM algoriths by siulation. MBDA based of ASM is also siulated. 6. Design of Siulation With the origin O at the radar center of the three axis (range, bearing, elevation), we assue the target oves in a plane with fixed z = he state vector is [ x x y y ] in Cartesian CS, whereas [ r r r ], [ b b b ] and [ e e e ] in spherical CS. Siulation tie is 0~00s. he odel-set of IMM has one CV odel and two C (Constant-urn) odels. here is always one odel in the odel-set which is atched to the syste otion ode copletely. In this point, the result of IMM algorith in this siulation is preferable to the adaptive IMM algorith. he transition probability atrix is: p ij = (4) and the odel initialized probability is: µ 0 = [ ] (5) he design of MAA and ASM siulation is a little ore coplex. o copare the perforance exactly, we use the easureents which were created in Cartesian CS containing process noise and easureent noise, after Cartesian-to-Spherical transforation, and obtain the easureents in Spherical CS. When finished the siulation, we transfor the estiating result in Spherical CS to Cartesian CS, and then copare to the true trajectory. In this siulation = 0.05, = 5 for EBDA, = 0 for MBDA. Pay attention that the filtering noise atrix in Spherical CS is relative to the noise in Cartesian CS but not equal to. he perforance of MAA and ASM siulation is very conservative and will be better in practice engineering in this point. We give the index of root ean square ( RMS ), noralized position error ( NPE ) and coputational load ( CL ). he CL of the algorith is the ean value which is evaluated in ters of calculating tie with respect to Monte Carlo runs. he forulas of calculating RMS and NPE are: M RMS( x, k) = ( [ ~ x ~ x ]) k k M (6) N k = j= RMS( x) = RMS( x,k) (7) N ISSN: Issue 8, Volue 8, August 009

8 WSEAS RANSACIONS on SYSEMS N / NPE = ( [ xk + yk + z k]/ N) / k= N ( [( xk zx, k) + ( yk zy, k) k= / ( zk zz, k) ] / N) M NPE NPE M j = + (8) = (9) where j =,,, M is the Monte Carlo runs, M = 00 ; k =,,, N is the siulation steps, N = 000 ; ~ xk ( j ), ~ y k, and ~ z k are respectively the estiation position error of k tie and the j th siulation in Cartesian CS; x k ( j ), y k, and z k are respectively the true position; z x, k, z y, k, and z z, k are respectively the easureents of three directions. Four different trajectories (Fig. ) of the target are tracked in the siulation. y.5 x rajectory rajectory rajectory 3 rajectory x Fig. he target trajectories () rajectory : x0 = [ ], constantly turns with ω( t ) = 5 / s. rajectory : x0 = [ ], 0~40s constantly turns with ω( t ) = 5 / s (turn left ), 55~75s constantly turns with ω( t ) = 7 / s (turn right), other tie oves in straight line with constant velocity (unifor otion). rajectory 3: x0 = [ ], 5~45s constantly turns with ω( t ) = 9 / s (turn right ), 60~75s constantly turns with ω( t ) = 7 / s (turn left), other tie oves in straight line with constant velocity. rajectory 4: x0 = [ ], 0~5s constantly turns with ω( t ) = 9 / s (turn right ), 40~55s and 60~85s constantly turns with ω( t ) = 7 / s (turn left), other tie oves in straight line with constant velocity. his is a trajectory like the figure Siulation Results Siulation results are given by able to able 5. able and able 3 gave the RMS of x and y directions respectively with ASM, IMM and MAA. able 4 gave the NEP and able 5 gave the CL. able Root Mean Square of x Coordinate () RMS ( x ) EBDA IMM MAA rajectory rajectory rajectory rajectory able 3 Root Mean Square of y Coordinate () RMS ( y ) EBDA IMM MAA rajectory rajectory rajectory rajectory able 4 Noralized Position Error () NPE EBDA MBDA IMM MAA rajectory rajectory rajectory rajectory able 5 Coputational Load (s) CL EBDA MBDA IMM MAA rajectory rajectory rajectory rajectory he filtering paraeters of ASM and MAA are identical and invariant in this siulation. he aneuverability (in Cartesian CS) is stronger and stronger fro trajectory to 4. Fro able, able 3 and able 4, the tracking perforance of EBDA and MBDA based of ASM is iproved obviously to MAA, although fro able 5 we can see that the coputational load of EBDA and MBDA is larger than that of MAA. Here we have 000 steps of one Monte Carlo siulation, the tie of one cycle of ASM is less than 3s (which is relative to the ISSN: Issue 8, Volue 8, August 009

9 WSEAS RANSACIONS on SYSEMS coputer perforance), and to 50s sapling interval the resource is well enough. For the tracking perforance of EBDA, to trajectory, IMM is a little better than EBDA, to trajectory and 3 the perforances of the two algoriths are near to each other, whereas to trajectory 4 EBDA is better than IMM. EBDA is also better than MBDA. Fro able, able 3 and able 4 we can conclude that when the aneuverability is weak or the otion ode varies little, IMM is a little better, but when the aneuverability is strong or the otion ode varies uch, ASM is better. Furtherore to say relatively, the tracking perforance of ASM is ore stationary, and that of IMM varies ore obviously. In the ASM algorith the adaptability of α is not considered. We can see fro (8) that if the acceleration a is approxiately constant then the coefficient α has less effect. In fact, to trajectory, the aneuver tie of the relative angular otion should be longer, that is α could be saller. We take α = 0. 0 for another siulation. he results show that the perforance of ASM ( NPE = 0.64 by EBDA) is nearer to IMM but not so apparent. 5 Conclusion his paper studied and proposed a new iproved adaptive single odel and two algoriths for aneuvering target tracking based on the relative angular otion of aerocrafts and radar, which is also suitable for range tracking. o say relatively, the tracking perforance of ASM is ore stationary, and that of IMM varies ore obviously according to the aneuverability and the design of odel-set. he united velocity-acceleration estiationbased decision algorith and direct easureentbased decision algorith based of ASM see to be valid for M. he physical concept of these new odel algoriths is clear, and they are siple and valuable for practical engineering, especially for radar tracking. Here we ephasized particularly on steadily tracking and exactly estiating of the state in spherical CS, which about the estiation of the state in Cartesian CS is not the goal of this paper, we will study in another one. References: [] X. R. Li and V. P. Jilkov, A survey of aneuvering target tracking-part I: dynaic odels, IEEE ransactions on Aerospace and Electronic Systes, Vol.39, No.4, 003, pp [] X. R. Li and V. P. Jilkov, A survey of aneuvering target tracking-part II: ballistic target odels, Proc. 00 SPIE Conf. on Signal and Data Processing of Sall argets, San Diego, CA, USA, 00, pp [3] X. R. Li and V. P. Jilkov, A survey of aneuvering target tracking-part III: easureent odels, Proc. 00 SPIE Conf. on Signal and Data Processing of Sall argets, San Diego, CA, USA, 00, pp [4] X. R. Li and V. P. Jilkov, A survey of aneuvering target tracking-part IV: decisionbased ethods, Proc. 00 SPIE Conf. on Signal and Data Processing of Sall argets, Orlando, FL, USA, 00, pp [5] X. R. Li and V. P. Jilkov, A survey of aneuvering target tracking-part V: ultipleodel ethods, IEEE ransactions on Aerospace and Electronic Systes, Vol.4, No.4, 005, pp [6] C. Z Han, H. Y. Zhu and Z. S. Duan, Multisource inforation fusion, singhua University Press, 006. [7] W. D. Blair, G. A. Watson and. R. Rice, racking aneuvering targets with an interacting ultiple odel filter containing exponentially correlated acceleration odels, Proc. wenty-hird Southeastern Syposiu on Systes heory, Colubia, SC, USA, 99, pp [8] Y. He, Z. J. Guo and J. P. Jiang, Model set design of the adaptive interacting ultiple odel tracking algorith, Electronics optics & Control, Vol.9, No., 00, pp [9] M. Lei and C. Z. Han, Expectationaxiization (EM) algorith based on IMM filtering with adaptive noise covariance, Acta Autoatica Sinica, Vol.3, No., 006, pp [0] B. J. Lee, J. B. Park and Y. H. Joo, IMM algorith using intelligent input estiation for aneuvering target tracking, IEICE ransactions on Fundaentals of Electronics, Counications and Coputer Sciences, Vol.88, No.5, 005, pp [] Kouasis, A., and Angelis, C.., Ipleenting adaptive systes using a odified Kalan filter, WSEAS ransactions on Systes, Vol.4, No., Dec. 005, pp [] Y. H. Zhang, W. Zhang, B. Zan, J. Wang and H. Chang, A ethod of reducing angular error in tracking noise jaer with ono-pulse radar, WSEAS ransactions on Counications, Vol.6, No.9, Sep. 007, pp ISSN: Issue 8, Volue 8, August 009

10 WSEAS RANSACIONS on SYSEMS [3] J. Jiang, Sliding-ode variable structure control for the Position racking Servo Syste, WSEAS ransactions on Systes, Vol.6, No., Feb. 007, pp [4] Bar-Shalo Y and Biriwal K. Variable diension filter for aneuvering target tracking, IEEE rans. on AES, Sep. 98, Vol.8, No.5, pp [5] J. C. Fang, A odified Kalan filter for tracking aneuvering targets, Acta Electronica Sinica, Vol., No.6, 983, pp [6] H. Zhou and K. S. P. Kuar, A current statistical odel and adaptive algorith for estiating aneuvering targets, AIAA Journal of Guidance, Vol.7, No.5, 984, pp [7] Z. Q. Zhu, A new adaptive Kalan filtering algorith for aneuvering target tracking, Acta Aeronautica et Astronautica Sinica, Vol.3, No.4, 99, pp [8] X. X. Liu, F. Ding, Z.. Hu and L. Zhou, An adaptive filtering algorith of noise variancesbased on odified current statistical odel, Chinese Journal of Electronics, Vol.5, No., 006, pp [9] B. Ekstrand. racking filters and odels for seeker applications, IEEE ransactions on Aerospace and Electronic Systes, Vol.37, No.3, Jan., 00, pp ISSN: Issue 8, Volue 8, August 009

An Improved Particle Filter with Applications in Ballistic Target Tracking

An Improved Particle Filter with Applications in Ballistic Target Tracking Sensors & ransducers Vol. 72 Issue 6 June 204 pp. 96-20 Sensors & ransducers 204 by IFSA Publishing S. L. http://www.sensorsportal.co An Iproved Particle Filter with Applications in Ballistic arget racing

More information

Warning System of Dangerous Chemical Gas in Factory Based on Wireless Sensor Network

Warning System of Dangerous Chemical Gas in Factory Based on Wireless Sensor Network 565 A publication of CHEMICAL ENGINEERING TRANSACTIONS VOL. 59, 07 Guest Editors: Zhuo Yang, Junie Ba, Jing Pan Copyright 07, AIDIC Servizi S.r.l. ISBN 978-88-95608-49-5; ISSN 83-96 The Italian Association

More information

Non-Parametric Non-Line-of-Sight Identification 1

Non-Parametric Non-Line-of-Sight Identification 1 Non-Paraetric Non-Line-of-Sight Identification Sinan Gezici, Hisashi Kobayashi and H. Vincent Poor Departent of Electrical Engineering School of Engineering and Applied Science Princeton University, Princeton,

More information

Department of Electronic and Optical Engineering, Ordnance Engineering College, Shijiazhuang, , China

Department of Electronic and Optical Engineering, Ordnance Engineering College, Shijiazhuang, , China 6th International Conference on Machinery, Materials, Environent, Biotechnology and Coputer (MMEBC 06) Solving Multi-Sensor Multi-Target Assignent Proble Based on Copositive Cobat Efficiency and QPSO Algorith

More information

Filtering and Fusion based Reconstruction of Angle of Attack

Filtering and Fusion based Reconstruction of Angle of Attack Filtering and Fusion based Reconstruction of Angle of Attack N Shantha Kuar Scientist, FMC Division NAL, Bangalore 7 E-ail: nskuar@css.nal.res.in Girija G Scientist, FMC Division NAL, Bangalore 7 E-ail:

More information

Effective joint probabilistic data association using maximum a posteriori estimates of target states

Effective joint probabilistic data association using maximum a posteriori estimates of target states Effective joint probabilistic data association using axiu a posteriori estiates of target states 1 Viji Paul Panakkal, 2 Rajbabu Velurugan 1 Central Research Laboratory, Bharat Electronics Ltd., Bangalore,

More information

Smoothing Framework for Automatic Track Initiation in Clutter

Smoothing Framework for Automatic Track Initiation in Clutter Soothing Fraework for Autoatic Track Initiation in Clutter Rajib Chakravorty Networked Sensor Technology (NeST) Laboratory Faculty of Engineering University of Technology, Sydney Broadway -27,Sydney, NSW

More information

Ufuk Demirci* and Feza Kerestecioglu**

Ufuk Demirci* and Feza Kerestecioglu** 1 INDIRECT ADAPTIVE CONTROL OF MISSILES Ufuk Deirci* and Feza Kerestecioglu** *Turkish Navy Guided Missile Test Station, Beykoz, Istanbul, TURKEY **Departent of Electrical and Electronics Engineering,

More information

Identical Maximum Likelihood State Estimation Based on Incremental Finite Mixture Model in PHD Filter

Identical Maximum Likelihood State Estimation Based on Incremental Finite Mixture Model in PHD Filter Identical Maxiu Lielihood State Estiation Based on Increental Finite Mixture Model in PHD Filter Gang Wu Eail: xjtuwugang@gail.co Jing Liu Eail: elelj20080730@ail.xjtu.edu.cn Chongzhao Han Eail: czhan@ail.xjtu.edu.cn

More information

Proc. of the IEEE/OES Seventh Working Conference on Current Measurement Technology UNCERTAINTIES IN SEASONDE CURRENT VELOCITIES

Proc. of the IEEE/OES Seventh Working Conference on Current Measurement Technology UNCERTAINTIES IN SEASONDE CURRENT VELOCITIES Proc. of the IEEE/OES Seventh Working Conference on Current Measureent Technology UNCERTAINTIES IN SEASONDE CURRENT VELOCITIES Belinda Lipa Codar Ocean Sensors 15 La Sandra Way, Portola Valley, CA 98 blipa@pogo.co

More information

SPECTRUM sensing is a core concept of cognitive radio

SPECTRUM sensing is a core concept of cognitive radio World Acadey of Science, Engineering and Technology International Journal of Electronics and Counication Engineering Vol:6, o:2, 202 Efficient Detection Using Sequential Probability Ratio Test in Mobile

More information

Keywords: Estimator, Bias, Mean-squared error, normality, generalized Pareto distribution

Keywords: Estimator, Bias, Mean-squared error, normality, generalized Pareto distribution Testing approxiate norality of an estiator using the estiated MSE and bias with an application to the shape paraeter of the generalized Pareto distribution J. Martin van Zyl Abstract In this work the norality

More information

ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FOR EARTH-TO-MOON TRANSFER

ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FOR EARTH-TO-MOON TRANSFER IEPC 003-0034 ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FOR EARTH-TO-MOON TRANSFER A. Bober, M. Guelan Asher Space Research Institute, Technion-Israel Institute of Technology, 3000 Haifa, Israel

More information

International Journal of Scientific & Engineering Research, Volume 4, Issue 9, September ISSN

International Journal of Scientific & Engineering Research, Volume 4, Issue 9, September ISSN International Journal of Scientific & Engineering Research, Volue 4, Issue 9, Septeber-3 44 ISSN 9-558 he unscented Kalan Filter for the Estiation the States of he Boiler-urbin Model Halieh Noorohaadi,

More information

Department of Physics, Sri Venkateswara University, Tirupati Range Operations, Satish Dhawan Space Centre SHAR, ISRO, Sriharikota

Department of Physics, Sri Venkateswara University, Tirupati Range Operations, Satish Dhawan Space Centre SHAR, ISRO, Sriharikota Trajectory Estiation of a Satellite Launch Vehicle Using Unscented Kalan Filter fro Noisy Radar Measureents R.Varaprasad S.V. Bhaskara Rao D.Narayana Rao V. Seshagiri Rao Range Operations, Satish Dhawan

More information

TEST OF HOMOGENEITY OF PARALLEL SAMPLES FROM LOGNORMAL POPULATIONS WITH UNEQUAL VARIANCES

TEST OF HOMOGENEITY OF PARALLEL SAMPLES FROM LOGNORMAL POPULATIONS WITH UNEQUAL VARIANCES TEST OF HOMOGENEITY OF PARALLEL SAMPLES FROM LOGNORMAL POPULATIONS WITH UNEQUAL VARIANCES S. E. Ahed, R. J. Tokins and A. I. Volodin Departent of Matheatics and Statistics University of Regina Regina,

More information

IN modern society that various systems have become more

IN modern society that various systems have become more Developent of Reliability Function in -Coponent Standby Redundant Syste with Priority Based on Maxiu Entropy Principle Ryosuke Hirata, Ikuo Arizono, Ryosuke Toohiro, Satoshi Oigawa, and Yasuhiko Takeoto

More information

A Simplified Analytical Approach for Efficiency Evaluation of the Weaving Machines with Automatic Filling Repair

A Simplified Analytical Approach for Efficiency Evaluation of the Weaving Machines with Automatic Filling Repair Proceedings of the 6th SEAS International Conference on Siulation, Modelling and Optiization, Lisbon, Portugal, Septeber -4, 006 0 A Siplified Analytical Approach for Efficiency Evaluation of the eaving

More information

Analyzing Simulation Results

Analyzing Simulation Results Analyzing Siulation Results Dr. John Mellor-Cruey Departent of Coputer Science Rice University johnc@cs.rice.edu COMP 528 Lecture 20 31 March 2005 Topics for Today Model verification Model validation Transient

More information

Experimental Design For Model Discrimination And Precise Parameter Estimation In WDS Analysis

Experimental Design For Model Discrimination And Precise Parameter Estimation In WDS Analysis City University of New York (CUNY) CUNY Acadeic Works International Conference on Hydroinforatics 8-1-2014 Experiental Design For Model Discriination And Precise Paraeter Estiation In WDS Analysis Giovanna

More information

2 Q 10. Likewise, in case of multiple particles, the corresponding density in 2 must be averaged over all

2 Q 10. Likewise, in case of multiple particles, the corresponding density in 2 must be averaged over all Lecture 6 Introduction to kinetic theory of plasa waves Introduction to kinetic theory So far we have been odeling plasa dynaics using fluid equations. The assuption has been that the pressure can be either

More information

Testing equality of variances for multiple univariate normal populations

Testing equality of variances for multiple univariate normal populations University of Wollongong Research Online Centre for Statistical & Survey Methodology Working Paper Series Faculty of Engineering and Inforation Sciences 0 esting equality of variances for ultiple univariate

More information

A Simple Regression Problem

A Simple Regression Problem A Siple Regression Proble R. M. Castro March 23, 2 In this brief note a siple regression proble will be introduced, illustrating clearly the bias-variance tradeoff. Let Y i f(x i ) + W i, i,..., n, where

More information

COS 424: Interacting with Data. Written Exercises

COS 424: Interacting with Data. Written Exercises COS 424: Interacting with Data Hoework #4 Spring 2007 Regression Due: Wednesday, April 18 Written Exercises See the course website for iportant inforation about collaboration and late policies, as well

More information

This model assumes that the probability of a gap has size i is proportional to 1/i. i.e., i log m e. j=1. E[gap size] = i P r(i) = N f t.

This model assumes that the probability of a gap has size i is proportional to 1/i. i.e., i log m e. j=1. E[gap size] = i P r(i) = N f t. CS 493: Algoriths for Massive Data Sets Feb 2, 2002 Local Models, Bloo Filter Scribe: Qin Lv Local Models In global odels, every inverted file entry is copressed with the sae odel. This work wells when

More information

Fast Montgomery-like Square Root Computation over GF(2 m ) for All Trinomials

Fast Montgomery-like Square Root Computation over GF(2 m ) for All Trinomials Fast Montgoery-like Square Root Coputation over GF( ) for All Trinoials Yin Li a, Yu Zhang a, a Departent of Coputer Science and Technology, Xinyang Noral University, Henan, P.R.China Abstract This letter

More information

A remark on a success rate model for DPA and CPA

A remark on a success rate model for DPA and CPA A reark on a success rate odel for DPA and CPA A. Wieers, BSI Version 0.5 andreas.wieers@bsi.bund.de Septeber 5, 2018 Abstract The success rate is the ost coon evaluation etric for easuring the perforance

More information

PULSE-TRAIN BASED TIME-DELAY ESTIMATION IMPROVES RESILIENCY TO NOISE

PULSE-TRAIN BASED TIME-DELAY ESTIMATION IMPROVES RESILIENCY TO NOISE PULSE-TRAIN BASED TIME-DELAY ESTIMATION IMPROVES RESILIENCY TO NOISE 1 Nicola Neretti, 1 Nathan Intrator and 1,2 Leon N Cooper 1 Institute for Brain and Neural Systes, Brown University, Providence RI 02912.

More information

Multiscale Entropy Analysis: A New Method to Detect Determinism in a Time. Series. A. Sarkar and P. Barat. Variable Energy Cyclotron Centre

Multiscale Entropy Analysis: A New Method to Detect Determinism in a Time. Series. A. Sarkar and P. Barat. Variable Energy Cyclotron Centre Multiscale Entropy Analysis: A New Method to Detect Deterinis in a Tie Series A. Sarkar and P. Barat Variable Energy Cyclotron Centre /AF Bidhan Nagar, Kolkata 700064, India PACS nubers: 05.45.Tp, 89.75.-k,

More information

Tracking using CONDENSATION: Conditional Density Propagation

Tracking using CONDENSATION: Conditional Density Propagation Tracking using CONDENSATION: Conditional Density Propagation Goal Model-based visual tracking in dense clutter at near video frae rates M. Isard and A. Blake, CONDENSATION Conditional density propagation

More information

Handwriting Detection Model Based on Four-Dimensional Vector Space Model

Handwriting Detection Model Based on Four-Dimensional Vector Space Model Journal of Matheatics Research; Vol. 10, No. 4; August 2018 ISSN 1916-9795 E-ISSN 1916-9809 Published by Canadian Center of Science and Education Handwriting Detection Model Based on Four-Diensional Vector

More information

Sharp Time Data Tradeoffs for Linear Inverse Problems

Sharp Time Data Tradeoffs for Linear Inverse Problems Sharp Tie Data Tradeoffs for Linear Inverse Probles Saet Oyak Benjain Recht Mahdi Soltanolkotabi January 016 Abstract In this paper we characterize sharp tie-data tradeoffs for optiization probles used

More information

Uncertainty Propagation and Nonlinear Filtering for Space Navigation using Differential Algebra

Uncertainty Propagation and Nonlinear Filtering for Space Navigation using Differential Algebra Uncertainty Propagation and Nonlinear Filtering for Space Navigation using Differential Algebra M. Valli, R. Arellin, P. Di Lizia and M. R. Lavagna Departent of Aerospace Engineering, Politecnico di Milano

More information

ANALYSIS ON RESPONSE OF DYNAMIC SYSTEMS TO PULSE SEQUENCES EXCITATION

ANALYSIS ON RESPONSE OF DYNAMIC SYSTEMS TO PULSE SEQUENCES EXCITATION The 4 th World Conference on Earthquake Engineering October -7, 8, Beijing, China ANALYSIS ON RESPONSE OF DYNAMIC SYSTEMS TO PULSE SEQUENCES EXCITATION S. Li C.H. Zhai L.L. Xie Ph. D. Student, School of

More information

Ph 20.3 Numerical Solution of Ordinary Differential Equations

Ph 20.3 Numerical Solution of Ordinary Differential Equations Ph 20.3 Nuerical Solution of Ordinary Differential Equations Due: Week 5 -v20170314- This Assignent So far, your assignents have tried to failiarize you with the hardware and software in the Physics Coputing

More information

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines Intelligent Systes: Reasoning and Recognition Jaes L. Crowley osig 1 Winter Seester 2018 Lesson 6 27 February 2018 Outline Perceptrons and Support Vector achines Notation...2 Linear odels...3 Lines, Planes

More information

ESTIMATING AND FORMING CONFIDENCE INTERVALS FOR EXTREMA OF RANDOM POLYNOMIALS. A Thesis. Presented to. The Faculty of the Department of Mathematics

ESTIMATING AND FORMING CONFIDENCE INTERVALS FOR EXTREMA OF RANDOM POLYNOMIALS. A Thesis. Presented to. The Faculty of the Department of Mathematics ESTIMATING AND FORMING CONFIDENCE INTERVALS FOR EXTREMA OF RANDOM POLYNOMIALS A Thesis Presented to The Faculty of the Departent of Matheatics San Jose State University In Partial Fulfillent of the Requireents

More information

Extension of CSRSM for the Parametric Study of the Face Stability of Pressurized Tunnels

Extension of CSRSM for the Parametric Study of the Face Stability of Pressurized Tunnels Extension of CSRSM for the Paraetric Study of the Face Stability of Pressurized Tunnels Guilhe Mollon 1, Daniel Dias 2, and Abdul-Haid Soubra 3, M.ASCE 1 LGCIE, INSA Lyon, Université de Lyon, Doaine scientifique

More information

GREY FORECASTING AND NEURAL NETWORK MODEL OF SPORT PERFORMANCE

GREY FORECASTING AND NEURAL NETWORK MODEL OF SPORT PERFORMANCE Journal of heoretical and Applied Inforation echnology 3 st March 03 Vol 49 No3 005-03 JAI & LLS All rights reserved ISSN: 99-8645 wwwatitorg E-ISSN: 87-395 GREY FORECASING AND NEURAL NEWORK MODEL OF SPOR

More information

Block designs and statistics

Block designs and statistics Bloc designs and statistics Notes for Math 447 May 3, 2011 The ain paraeters of a bloc design are nuber of varieties v, bloc size, nuber of blocs b. A design is built on a set of v eleents. Each eleent

More information

Using a De-Convolution Window for Operating Modal Analysis

Using a De-Convolution Window for Operating Modal Analysis Using a De-Convolution Window for Operating Modal Analysis Brian Schwarz Vibrant Technology, Inc. Scotts Valley, CA Mark Richardson Vibrant Technology, Inc. Scotts Valley, CA Abstract Operating Modal Analysis

More information

Ch 12: Variations on Backpropagation

Ch 12: Variations on Backpropagation Ch 2: Variations on Backpropagation The basic backpropagation algorith is too slow for ost practical applications. It ay take days or weeks of coputer tie. We deonstrate why the backpropagation algorith

More information

Feature Extraction Techniques

Feature Extraction Techniques Feature Extraction Techniques Unsupervised Learning II Feature Extraction Unsupervised ethods can also be used to find features which can be useful for categorization. There are unsupervised ethods that

More information

Estimation of ADC Nonlinearities from the Measurement in Input Voltage Intervals

Estimation of ADC Nonlinearities from the Measurement in Input Voltage Intervals Estiation of ADC Nonlinearities fro the Measureent in Input Voltage Intervals M. Godla, L. Michaeli, 3 J. Šaliga, 4 R. Palenčár,,3 Deptartent of Electronics and Multiedia Counications, FEI TU of Košice,

More information

Interactive Markov Models of Evolutionary Algorithms

Interactive Markov Models of Evolutionary Algorithms Cleveland State University EngagedScholarship@CSU Electrical Engineering & Coputer Science Faculty Publications Electrical Engineering & Coputer Science Departent 2015 Interactive Markov Models of Evolutionary

More information

Pattern Recognition and Machine Learning. Artificial Neural networks

Pattern Recognition and Machine Learning. Artificial Neural networks Pattern Recognition and Machine Learning Jaes L. Crowley ENSIMAG 3 - MMIS Fall Seester 2017 Lessons 7 20 Dec 2017 Outline Artificial Neural networks Notation...2 Introduction...3 Key Equations... 3 Artificial

More information

ma x = -bv x + F rod.

ma x = -bv x + F rod. Notes on Dynaical Systes Dynaics is the study of change. The priary ingredients of a dynaical syste are its state and its rule of change (also soeties called the dynaic). Dynaical systes can be continuous

More information

An Adaptive UKF Algorithm for the State and Parameter Estimations of a Mobile Robot

An Adaptive UKF Algorithm for the State and Parameter Estimations of a Mobile Robot Vol. 34, No. 1 ACTA AUTOMATICA SINICA January, 2008 An Adaptive UKF Algorith for the State and Paraeter Estiations of a Mobile Robot SONG Qi 1, 2 HAN Jian-Da 1 Abstract For iproving the estiation accuracy

More information

Model Fitting. CURM Background Material, Fall 2014 Dr. Doreen De Leon

Model Fitting. CURM Background Material, Fall 2014 Dr. Doreen De Leon Model Fitting CURM Background Material, Fall 014 Dr. Doreen De Leon 1 Introduction Given a set of data points, we often want to fit a selected odel or type to the data (e.g., we suspect an exponential

More information

Vulnerability of MRD-Code-Based Universal Secure Error-Correcting Network Codes under Time-Varying Jamming Links

Vulnerability of MRD-Code-Based Universal Secure Error-Correcting Network Codes under Time-Varying Jamming Links Vulnerability of MRD-Code-Based Universal Secure Error-Correcting Network Codes under Tie-Varying Jaing Links Jun Kurihara KDDI R&D Laboratories, Inc 2 5 Ohara, Fujiino, Saitaa, 356 8502 Japan Eail: kurihara@kddilabsjp

More information

are equal to zero, where, q = p 1. For each gene j, the pairwise null and alternative hypotheses are,

are equal to zero, where, q = p 1. For each gene j, the pairwise null and alternative hypotheses are, Page of 8 Suppleentary Materials: A ultiple testing procedure for ulti-diensional pairwise coparisons with application to gene expression studies Anjana Grandhi, Wenge Guo, Shyaal D. Peddada S Notations

More information

A note on the multiplication of sparse matrices

A note on the multiplication of sparse matrices Cent. Eur. J. Cop. Sci. 41) 2014 1-11 DOI: 10.2478/s13537-014-0201-x Central European Journal of Coputer Science A note on the ultiplication of sparse atrices Research Article Keivan Borna 12, Sohrab Aboozarkhani

More information

An Extension to the Tactical Planning Model for a Job Shop: Continuous-Time Control

An Extension to the Tactical Planning Model for a Job Shop: Continuous-Time Control An Extension to the Tactical Planning Model for a Job Shop: Continuous-Tie Control Chee Chong. Teo, Rohit Bhatnagar, and Stephen C. Graves Singapore-MIT Alliance, Nanyang Technological Univ., and Massachusetts

More information

Inspection; structural health monitoring; reliability; Bayesian analysis; updating; decision analysis; value of information

Inspection; structural health monitoring; reliability; Bayesian analysis; updating; decision analysis; value of information Cite as: Straub D. (2014). Value of inforation analysis with structural reliability ethods. Structural Safety, 49: 75-86. Value of Inforation Analysis with Structural Reliability Methods Daniel Straub

More information

DISSIMILARITY MEASURES FOR ICA-BASED SOURCE NUMBER ESTIMATION. Seungchul Lee 2 2. University of Michigan. Ann Arbor, MI, USA.

DISSIMILARITY MEASURES FOR ICA-BASED SOURCE NUMBER ESTIMATION. Seungchul Lee 2 2. University of Michigan. Ann Arbor, MI, USA. Proceedings of the ASME International Manufacturing Science and Engineering Conference MSEC June -8,, Notre Dae, Indiana, USA MSEC-7 DISSIMILARIY MEASURES FOR ICA-BASED SOURCE NUMBER ESIMAION Wei Cheng,

More information

Optimal Jamming Over Additive Noise: Vector Source-Channel Case

Optimal Jamming Over Additive Noise: Vector Source-Channel Case Fifty-first Annual Allerton Conference Allerton House, UIUC, Illinois, USA October 2-3, 2013 Optial Jaing Over Additive Noise: Vector Source-Channel Case Erah Akyol and Kenneth Rose Abstract This paper

More information

An Approximate Model for the Theoretical Prediction of the Velocity Increase in the Intermediate Ballistics Period

An Approximate Model for the Theoretical Prediction of the Velocity Increase in the Intermediate Ballistics Period An Approxiate Model for the Theoretical Prediction of the Velocity... 77 Central European Journal of Energetic Materials, 205, 2(), 77-88 ISSN 2353-843 An Approxiate Model for the Theoretical Prediction

More information

Ocean 420 Physical Processes in the Ocean Project 1: Hydrostatic Balance, Advection and Diffusion Answers

Ocean 420 Physical Processes in the Ocean Project 1: Hydrostatic Balance, Advection and Diffusion Answers Ocean 40 Physical Processes in the Ocean Project 1: Hydrostatic Balance, Advection and Diffusion Answers 1. Hydrostatic Balance a) Set all of the levels on one of the coluns to the lowest possible density.

More information

MODIFICATION OF AN ANALYTICAL MODEL FOR CONTAINER LOADING PROBLEMS

MODIFICATION OF AN ANALYTICAL MODEL FOR CONTAINER LOADING PROBLEMS MODIFICATIO OF A AALYTICAL MODEL FOR COTAIER LOADIG PROBLEMS Reception date: DEC.99 otification to authors: 04 MAR. 2001 Cevriye GECER Departent of Industrial Engineering, University of Gazi 06570 Maltepe,

More information

Bayesian Approach for Fatigue Life Prediction from Field Inspection

Bayesian Approach for Fatigue Life Prediction from Field Inspection Bayesian Approach for Fatigue Life Prediction fro Field Inspection Dawn An and Jooho Choi School of Aerospace & Mechanical Engineering, Korea Aerospace University, Goyang, Seoul, Korea Srira Pattabhiraan

More information

MSEC MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL SOLUTION FOR MAINTENANCE AND PERFORMANCE

MSEC MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL SOLUTION FOR MAINTENANCE AND PERFORMANCE Proceeding of the ASME 9 International Manufacturing Science and Engineering Conference MSEC9 October 4-7, 9, West Lafayette, Indiana, USA MSEC9-8466 MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL

More information

THE ROCKET EXPERIMENT 1. «Homogenous» gravitational field

THE ROCKET EXPERIMENT 1. «Homogenous» gravitational field THE OCKET EXPEIENT. «Hoogenous» gravitational field Let s assue, fig., that we have a body of ass Μ and radius. fig. As it is known, the gravitational field of ass Μ (both in ters of geoetry and dynaics)

More information

e-companion ONLY AVAILABLE IN ELECTRONIC FORM

e-companion ONLY AVAILABLE IN ELECTRONIC FORM OPERATIONS RESEARCH doi 10.1287/opre.1070.0427ec pp. ec1 ec5 e-copanion ONLY AVAILABLE IN ELECTRONIC FORM infors 07 INFORMS Electronic Copanion A Learning Approach for Interactive Marketing to a Custoer

More information

Automated Frequency Domain Decomposition for Operational Modal Analysis

Automated Frequency Domain Decomposition for Operational Modal Analysis Autoated Frequency Doain Decoposition for Operational Modal Analysis Rune Brincker Departent of Civil Engineering, University of Aalborg, Sohngaardsholsvej 57, DK-9000 Aalborg, Denark Palle Andersen Structural

More information

Physically Based Modeling CS Notes Spring 1997 Particle Collision and Contact

Physically Based Modeling CS Notes Spring 1997 Particle Collision and Contact Physically Based Modeling CS 15-863 Notes Spring 1997 Particle Collision and Contact 1 Collisions with Springs Suppose we wanted to ipleent a particle siulator with a floor : a solid horizontal plane which

More information

Ensemble Based on Data Envelopment Analysis

Ensemble Based on Data Envelopment Analysis Enseble Based on Data Envelopent Analysis So Young Sohn & Hong Choi Departent of Coputer Science & Industrial Systes Engineering, Yonsei University, Seoul, Korea Tel) 82-2-223-404, Fax) 82-2- 364-7807

More information

Biostatistics Department Technical Report

Biostatistics Department Technical Report Biostatistics Departent Technical Report BST006-00 Estiation of Prevalence by Pool Screening With Equal Sized Pools and a egative Binoial Sapling Model Charles R. Katholi, Ph.D. Eeritus Professor Departent

More information

Support Vector Machine Classification of Uncertain and Imbalanced data using Robust Optimization

Support Vector Machine Classification of Uncertain and Imbalanced data using Robust Optimization Recent Researches in Coputer Science Support Vector Machine Classification of Uncertain and Ibalanced data using Robust Optiization RAGHAV PAT, THEODORE B. TRAFALIS, KASH BARKER School of Industrial Engineering

More information

Forecasting Financial Indices: The Baltic Dry Indices

Forecasting Financial Indices: The Baltic Dry Indices International Journal of Maritie, Trade & Econoic Issues pp. 109-130 Volue I, Issue (1), 2013 Forecasting Financial Indices: The Baltic Dry Indices Eleftherios I. Thalassinos 1, Mike P. Hanias 2, Panayiotis

More information

1 Proof of learning bounds

1 Proof of learning bounds COS 511: Theoretical Machine Learning Lecturer: Rob Schapire Lecture #4 Scribe: Akshay Mittal February 13, 2013 1 Proof of learning bounds For intuition of the following theore, suppose there exists a

More information

Nonmonotonic Networks. a. IRST, I Povo (Trento) Italy, b. Univ. of Trento, Physics Dept., I Povo (Trento) Italy

Nonmonotonic Networks. a. IRST, I Povo (Trento) Italy, b. Univ. of Trento, Physics Dept., I Povo (Trento) Italy Storage Capacity and Dynaics of Nononotonic Networks Bruno Crespi a and Ignazio Lazzizzera b a. IRST, I-38050 Povo (Trento) Italy, b. Univ. of Trento, Physics Dept., I-38050 Povo (Trento) Italy INFN Gruppo

More information

be the original image. Let be the Gaussian white noise with zero mean and variance σ. g ( x, denotes the motion blurred image with noise.

be the original image. Let be the Gaussian white noise with zero mean and variance σ. g ( x, denotes the motion blurred image with noise. 3rd International Conference on Multiedia echnology(icm 3) Estiation Forula for Noise Variance of Motion Blurred Iages He Weiguo Abstract. he ey proble in restoration of otion blurred iages is how to get

More information

State Estimation Problem for the Action Potential Modeling in Purkinje Fibers

State Estimation Problem for the Action Potential Modeling in Purkinje Fibers APCOM & ISCM -4 th Deceber, 203, Singapore State Estiation Proble for the Action Potential Modeling in Purinje Fibers *D. C. Estuano¹, H. R. B.Orlande and M. J.Colaço Federal University of Rio de Janeiro

More information

Qualitative Modelling of Time Series Using Self-Organizing Maps: Application to Animal Science

Qualitative Modelling of Time Series Using Self-Organizing Maps: Application to Animal Science Proceedings of the 6th WSEAS International Conference on Applied Coputer Science, Tenerife, Canary Islands, Spain, Deceber 16-18, 2006 183 Qualitative Modelling of Tie Series Using Self-Organizing Maps:

More information

Detection and Estimation Theory

Detection and Estimation Theory ESE 54 Detection and Estiation Theory Joseph A. O Sullivan Sauel C. Sachs Professor Electronic Systes and Signals Research Laboratory Electrical and Systes Engineering Washington University 11 Urbauer

More information

Symbolic Analysis as Universal Tool for Deriving Properties of Non-linear Algorithms Case study of EM Algorithm

Symbolic Analysis as Universal Tool for Deriving Properties of Non-linear Algorithms Case study of EM Algorithm Acta Polytechnica Hungarica Vol., No., 04 Sybolic Analysis as Universal Tool for Deriving Properties of Non-linear Algoriths Case study of EM Algorith Vladiir Mladenović, Miroslav Lutovac, Dana Porrat

More information

Reducing Vibration and Providing Robustness with Multi-Input Shapers

Reducing Vibration and Providing Robustness with Multi-Input Shapers 29 Aerican Control Conference Hyatt Regency Riverfront, St. Louis, MO, USA June -2, 29 WeA6.4 Reducing Vibration and Providing Robustness with Multi-Input Shapers Joshua Vaughan and Willia Singhose Abstract

More information

Estimating Parameters for a Gaussian pdf

Estimating Parameters for a Gaussian pdf Pattern Recognition and achine Learning Jaes L. Crowley ENSIAG 3 IS First Seester 00/0 Lesson 5 7 Noveber 00 Contents Estiating Paraeters for a Gaussian pdf Notation... The Pattern Recognition Proble...3

More information

Pattern Recognition and Machine Learning. Artificial Neural networks

Pattern Recognition and Machine Learning. Artificial Neural networks Pattern Recognition and Machine Learning Jaes L. Crowley ENSIMAG 3 - MMIS Fall Seester 2016 Lessons 7 14 Dec 2016 Outline Artificial Neural networks Notation...2 1. Introduction...3... 3 The Artificial

More information

A Model for the Selection of Internet Service Providers

A Model for the Selection of Internet Service Providers ISSN 0146-4116, Autoatic Control and Coputer Sciences, 2008, Vol. 42, No. 5, pp. 249 254. Allerton Press, Inc., 2008. Original Russian Text I.M. Aliev, 2008, published in Avtoatika i Vychislitel naya Tekhnika,

More information

On Conditions for Linearity of Optimal Estimation

On Conditions for Linearity of Optimal Estimation On Conditions for Linearity of Optial Estiation Erah Akyol, Kuar Viswanatha and Kenneth Rose {eakyol, kuar, rose}@ece.ucsb.edu Departent of Electrical and Coputer Engineering University of California at

More information

Neural Network-Aided Extended Kalman Filter for SLAM Problem

Neural Network-Aided Extended Kalman Filter for SLAM Problem 7 IEEE International Conference on Robotics and Autoation Roa, Italy, -4 April 7 ThA.5 Neural Network-Aided Extended Kalan Filter for SLAM Proble Minyong Choi, R. Sakthivel, and Wan Kyun Chung Abstract

More information

2.003 Engineering Dynamics Problem Set 2 Solutions

2.003 Engineering Dynamics Problem Set 2 Solutions .003 Engineering Dynaics Proble Set Solutions This proble set is priarily eant to give the student practice in describing otion. This is the subject of kineatics. It is strongly recoended that you study

More information

Spine Fin Efficiency A Three Sided Pyramidal Fin of Equilateral Triangular Cross-Sectional Area

Spine Fin Efficiency A Three Sided Pyramidal Fin of Equilateral Triangular Cross-Sectional Area Proceedings of the 006 WSEAS/IASME International Conference on Heat and Mass Transfer, Miai, Florida, USA, January 18-0, 006 (pp13-18) Spine Fin Efficiency A Three Sided Pyraidal Fin of Equilateral Triangular

More information

THE KALMAN FILTER: A LOOK BEHIND THE SCENE

THE KALMAN FILTER: A LOOK BEHIND THE SCENE HE KALMA FILER: A LOOK BEHID HE SCEE R.E. Deain School of Matheatical and Geospatial Sciences, RMI University eail: rod.deain@rit.edu.au Presented at the Victorian Regional Survey Conference, Mildura,

More information

On Constant Power Water-filling

On Constant Power Water-filling On Constant Power Water-filling Wei Yu and John M. Cioffi Electrical Engineering Departent Stanford University, Stanford, CA94305, U.S.A. eails: {weiyu,cioffi}@stanford.edu Abstract This paper derives

More information

NUMERICAL MODELLING OF THE TYRE/ROAD CONTACT

NUMERICAL MODELLING OF THE TYRE/ROAD CONTACT NUMERICAL MODELLING OF THE TYRE/ROAD CONTACT PACS REFERENCE: 43.5.LJ Krister Larsson Departent of Applied Acoustics Chalers University of Technology SE-412 96 Sweden Tel: +46 ()31 772 22 Fax: +46 ()31

More information

Multi-Scale/Multi-Resolution: Wavelet Transform

Multi-Scale/Multi-Resolution: Wavelet Transform Multi-Scale/Multi-Resolution: Wavelet Transfor Proble with Fourier Fourier analysis -- breaks down a signal into constituent sinusoids of different frequencies. A serious drawback in transforing to the

More information

Machine Learning Basics: Estimators, Bias and Variance

Machine Learning Basics: Estimators, Bias and Variance Machine Learning Basics: Estiators, Bias and Variance Sargur N. srihari@cedar.buffalo.edu This is part of lecture slides on Deep Learning: http://www.cedar.buffalo.edu/~srihari/cse676 1 Topics in Basics

More information

List Scheduling and LPT Oliver Braun (09/05/2017)

List Scheduling and LPT Oliver Braun (09/05/2017) List Scheduling and LPT Oliver Braun (09/05/207) We investigate the classical scheduling proble P ax where a set of n independent jobs has to be processed on 2 parallel and identical processors (achines)

More information

Support recovery in compressed sensing: An estimation theoretic approach

Support recovery in compressed sensing: An estimation theoretic approach Support recovery in copressed sensing: An estiation theoretic approach Ain Karbasi, Ali Horati, Soheil Mohajer, Martin Vetterli School of Coputer and Counication Sciences École Polytechnique Fédérale de

More information

Donald Fussell. October 28, Computer Science Department The University of Texas at Austin. Point Masses and Force Fields.

Donald Fussell. October 28, Computer Science Department The University of Texas at Austin. Point Masses and Force Fields. s Vector Moving s and Coputer Science Departent The University of Texas at Austin October 28, 2014 s Vector Moving s Siple classical dynaics - point asses oved by forces Point asses can odel particles

More information

E0 370 Statistical Learning Theory Lecture 6 (Aug 30, 2011) Margin Analysis

E0 370 Statistical Learning Theory Lecture 6 (Aug 30, 2011) Margin Analysis E0 370 tatistical Learning Theory Lecture 6 (Aug 30, 20) Margin Analysis Lecturer: hivani Agarwal cribe: Narasihan R Introduction In the last few lectures we have seen how to obtain high confidence bounds

More information

A MESHSIZE BOOSTING ALGORITHM IN KERNEL DENSITY ESTIMATION

A MESHSIZE BOOSTING ALGORITHM IN KERNEL DENSITY ESTIMATION A eshsize boosting algorith in kernel density estiation A MESHSIZE BOOSTING ALGORITHM IN KERNEL DENSITY ESTIMATION C.C. Ishiekwene, S.M. Ogbonwan and J.E. Osewenkhae Departent of Matheatics, University

More information

On the theoretical analysis of cross validation in compressive sensing

On the theoretical analysis of cross validation in compressive sensing MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.erl.co On the theoretical analysis of cross validation in copressive sensing Zhang, J.; Chen, L.; Boufounos, P.T.; Gu, Y. TR2014-025 May 2014 Abstract

More information

Randomized Recovery for Boolean Compressed Sensing

Randomized Recovery for Boolean Compressed Sensing Randoized Recovery for Boolean Copressed Sensing Mitra Fatei and Martin Vetterli Laboratory of Audiovisual Counication École Polytechnique Fédéral de Lausanne (EPFL) Eail: {itra.fatei, artin.vetterli}@epfl.ch

More information

3.8 Three Types of Convergence

3.8 Three Types of Convergence 3.8 Three Types of Convergence 3.8 Three Types of Convergence 93 Suppose that we are given a sequence functions {f k } k N on a set X and another function f on X. What does it ean for f k to converge to

More information

Pattern Recognition and Machine Learning. Artificial Neural networks

Pattern Recognition and Machine Learning. Artificial Neural networks Pattern Recognition and Machine Learning Jaes L. Crowley ENSIMAG 3 - MMIS Fall Seester 2016/2017 Lessons 9 11 Jan 2017 Outline Artificial Neural networks Notation...2 Convolutional Neural Networks...3

More information

Intelligent Systems: Reasoning and Recognition. Artificial Neural Networks

Intelligent Systems: Reasoning and Recognition. Artificial Neural Networks Intelligent Systes: Reasoning and Recognition Jaes L. Crowley MOSIG M1 Winter Seester 2018 Lesson 7 1 March 2018 Outline Artificial Neural Networks Notation...2 Introduction...3 Key Equations... 3 Artificial

More information

13.2 Fully Polynomial Randomized Approximation Scheme for Permanent of Random 0-1 Matrices

13.2 Fully Polynomial Randomized Approximation Scheme for Permanent of Random 0-1 Matrices CS71 Randoness & Coputation Spring 018 Instructor: Alistair Sinclair Lecture 13: February 7 Disclaier: These notes have not been subjected to the usual scrutiny accorded to foral publications. They ay

More information