Processing math: 100%
A journal of IEEE and CAA , publishes high-quality papers in English on original theoretical/experimental research and development in all areas of automation
Weiqiang Dong and Moshe Kam, "Dependent Randomization in Parallel Binary Decision Fusion," IEEE/CAA J. Autom. Sinica, vol. 8, no. 2, pp. 361-376, Feb. 2021. doi: 10.1109/JAS.2021.1003823
Citation: Weiqiang Dong and Moshe Kam, "Dependent Randomization in Parallel Binary Decision Fusion," IEEE/CAA J. Autom. Sinica, vol. 8, no. 2, pp. 361-376, Feb. 2021. doi: 10.1109/JAS.2021.1003823

Dependent Randomization in Parallel Binary Decision Fusion

doi: 10.1109/JAS.2021.1003823
More Information
  • We consider a parallel decentralized detection system employing a bank of local detectors (LDs) to access a commonly-observed phenomenon. The system makes a binary decision about the phenomenon, accepting one of two hypotheses (H0 (“absent”) or H1 (“present”)). The kth LD uses a local decision rule to compress its local observations yk into a binary local decision uk; uk=0 if the kth LD accepts H0 and uk=1 if it accepts H1. The kth LD sends its decision uk over a noiseless dedicated channel to a Data Fusion Center (DFC). The DFC combines the local decisions it receives from n LDs (u1,u2,,un) into a single binary global decision u0 (u0=0 for accepting H0 or u0=1 for accepting H1). If each LD uses a single deterministic local decision rule (calculating uk from the local observations yk) and the DFC uses a single deterministic global decision rule (calculating u0 from the n local decisions), the team receiver operating characteristic (ROC) curve is in general non-concave. The system’s performance under a Neyman-Pearson criterion may then be suboptimal in the sense that a mixed strategy may yield a higher probability of detection when the probability of false alarm is constrained not to exceed a certain value, α>0. Specifically, a “dependent randomization” detection scheme can be applied in certain circumstances to improve the system’s performance by making the ROC curve concave. This scheme requires a coordinated and synchronized action between the DFC and the LDs. In this study, we specify when dependent randomization is needed, and discuss the proper response of the detection system if synchronization between the LDs and the DFC is temporarily lost.

     

  • WE study the parallel decentralized binary detection architecture shown in Fig. 1. The system uses n local detectors (LDs) to observe a binary phenomenon (“target/no target”). The objective is to decide if a target is present (hypothesis H1) or absent (hypothesis H0). We use P1 to denote the a priori probability that a target is present (hypothesis H1) and use P0=1P1 to denote the a priori probability that a target is absent (hypothesis H0). The local observations collected by the kth LD are denoted yk. All the local observations are assumed to be statistically independent, conditioned on the hypothesis, therefore, Pr(y1,,yn|Hj)=kPr(yk|Hj),k{1,,n},j{0,1}. Each LD compresses its local observations into a local decision; the local decision of the kth LD is uk=γk(yk), uk{0,1} and U={u1,u2,,un}. Here uk=0 means that the kth LD prefers hypothesis H0, and uk=1 means that the kth LD prefers hypothesis H1. A Data Fusion Center (DFC) combines all the local decisions to generate a global decision u0=γ0(U), u0{0,1}, where u0=0 indicates preference for hypothesis H0, and u0=1 indicates preference for hypothesis H1. The probability of false alarm of the DFC is Pf=Pr(u0=1|H0). The probability of detection of the DFC is Pd=Pr(u0=1|H1). The tuple (Pf,Pd) is considered the operating point of the detection system. Similarly, the local operating point of the kth LD is (Pfk,Pdk), where Pfk=Pr(uk=1|H0) and Pdk=Pr(uk=1|H1). We will sometimes refer to (PAf,PAd) as “point A”.

    Figure  1.  Parallel decentralized detection network.

    The architecture of Fig. 1 was investigated extensively (e.g., [1]–[10]). Its design in this study aims to satisfy a Neyman-Pearson criterion, namely to maximize the global probability of detection Pd while keeping the global probability of false alarm Pf not larger than a specified value α (0<α<1).

    To achieve this objective, the implementation tasks are to determine the local decision rules (mapping γk from yk to uk, k=1,,n) and the global decision rule (mapping γ0 from U to u0). The local decision rule of the kth LD, γk(.), determines how the kth LD compresses its local observations yk into its local decision uk as uk=γk(yk). The global decision rule, γ0, represents how the DFC integrates all the local decisions into the global decision u0 as u0=γ0(U). The combination of all the local decision rules is γLD={γ1,,γn}. The combination of the global decision rule and all the local decision rules is the detection strategy, γ={γ0,γLD}={γ0,,γn}.

    Among other approaches, the existing literature provides analysis of three detection strategies for the system in Fig. 1.

    1) A Deterministic Strategy: Each LD uses a deterministic local decision rule and the DFC uses a deterministic global decision rule [1].

    2) A Strategy With Randomization at the DFC Only: Each LD uses a deterministic local decision rule and the DFC uses a randomized global decision rule [2], [7].

    3) Dependent Randomization: all the LDs and the DFC use randomized decision rules. The randomization between the LDs and the DFC is coordinated and synchronized [5].

    The need for dependent randomization (or “scheduled test”) was introduced in [4] and [5]. In [4], it was indicated that dependent randomization is certainly feasible but “may perform poorly if synchronization is lost.” In [5] scenarios were studied where dependent randomization can improve performance over a deterministic strategy and independent randomization. In this study, we find conditions under which dependent randomization is beneficial, quantify the effect of synchronization loss, and demonstrate how to recover (partially) from synchronization loss. Specifically, this paper analyzes the following cases.

    4) Dependent Randomization (DFC Unsynchronized): All the LDs and the DFC use randomized decision rules. The LDs are coordinated and synchronized with each other. The DFC uses a randomized global decision rule but is not synchronized with the LDs.

    5) Dependent Randomization (LDs Unsynchronized): All the LDs and the DFC use randomized decision rules. The randomization between some LDs and the DFC is coordinated and synchronized. Other LDs use randomized local decision rules independently of the other LDs and the DFC.

    In Section II we present the three detection strategies (a)–(c) and summarize their characteristics. In Section III we present two examples of parallel decentralized detection system which we use to exemplify different operating conditions and performance. Section IV discusses strategies needed when the DFC loses synchronization with the LDs (d). Section V discusses strategies needed when some LDs lose synchronization with other LDs and the DFC (e).

    A detection strategy is deterministic if each LD uses a single deterministic local decision rule and the DFC uses a single deterministic global decision rule. The operating point of the system A=(PAf,PAd) is determined by the deterministic strategy γA={γA0,,γAn}. The corresponding operating point of the kth LD in the system, determined by γAk, is (PAfk,PAdk).

    It is shown in [5] that under the assumption that the local observations y1,,yn are conditionally independent given the hypothesis H0 or H1, for satisfying a Neyman-Pearson criterion, both γ0 and γk are likelihood ratio tests of the form

    u0=γ0(U)={0Λ(U)<η01Λ(U)η0 (1)
    uk=γk(yk)={0Λ(yk)<ηk1Λ(yk)ηk  (2)

    where

    Λ(x)=P(x|H1)P(x|H0). (3)

    Here, η0 is the threshold of the global decision rule and ηk is the threshold of the local decision rule of the kth LD. These thresholds are often part of the design, and they determine the trade-off between the probability of false alarm and the probability of detection. Under a Neyman-Pearson criterion, η0,η1,,ηn are designed to maximize the global probability of detection while keeping the global probability of false alarm not greater than α(0,1). The value of α is determined by the designer, based on the level of tolerance of false alarm in the specific application.

    For a binary parallel decentralized detection system with n LDs, the local decision vector U has 2n possible values, which translates into 22n possible global decision rules. However, not every global decision rules is eligible for consideration as a potentially optimal decision rule. According to (1), the global decision rule is a likelihood ratio test and u0 is a non-decreasing function of Λ(U). Thomopoulos et al. [2] showed that the optimal deterministic global decision rule that satisfies the Neyman-Pearson criterion must be a monotonic fusion rule (per Lemma 1 of [2], function d). A fusion rule is monotonic if, for every combination of local decisions U={u1,,un}, switching one of the local decision from 0 to 1 can only cause the global decision u0 to switch from u0=0 to u0=1 and not from u0=1 to u0=0. An algorithm that calculates all the monotonic fusion rules of a system with n LDs is provided in [11]. Since some monotonic fusion rules dominate others (would always result in better performance than others), the eligible optimal deterministic global decision rules would be a subset of all the monotonic fusion rules.

    The probability of false alarm and the probability of detection of the architecture shown in Fig. 1 are

    {Pf=Pr(u0=1|H0)=Λ(U)η0P(Λ(U)|H0)Pd=Pr(u0=1|H1)=Λ(U)η0P(Λ(U)|H1). (4)

    The probability of false alarm and the probability of detection at the kth LD are

    {Pfk=Pr(uk=1|H0)=yk|Λ(yk)ηkP(yk|H0)dykPdk=Pr(uk=1|H0)=yk|Λ(yk)ηkP(yk|H1)dyk (5)

    where Λ(yk) is the likelihood ratio of yk.

    When all the local operating points (Pfk,Pdk),k=1,,n are known, then (4) can be written as ([3, pp. 567–568])

    {Pf=1u1=01un=0nk=1Pukfk(1Pfk)(1uk)×U1(nk=1(PdkPfk)uk(1Pdk1Pfk)(1uk)η0)Pd=1u1=01un=0nk=1Pukdk(1Pdk)(1uk)×U1(nk=1(PdkPfk)uk(1Pdk1Pfk)(1uk)η0) (6)

    where U1(.) is the unit step function

    U1(x)={0x<01x0. (7)

    In (6), the value of the unit step function provides the global decision u0 for a given local decision set U={u1,,un}.

    If the local operating points are identical, (Pfk,Pdk)=(pf,pd),k=1,,n, the global decision rule (1) becomes a “k out of n” rule, which means if k or more LDs in the system decide ‘1’, then u0=1; otherwise, u0=0. In this circumstance, the global probability of false alarm and the global probability of detection are

    {Pf=nk(nk)pfk(1pf)(nk)Pd=nk(nk)pdk(1pd)(nk). (8)

    We consider two cases of finding a deterministic strategy, depending on the local observations.

    1) Local Observations Contain No Point Masses of Probability: Hoballah and Varshney [1] studied this case, using a Person-by-Person optimization (PBPO) approach to synthesize γ0 and γk in (1) and (2). Acharya et al. [8] proposed a method for solving for the optimal γ0 and γk simultaneously when the LDs are identical.

    2) Local Observations Contain Point Masses of Probability: If the local observations are discrete and finite, the probability distribution of the local observations contain point masses of probability. In this case, the number of achievable local operating points {(Pf1,Pd1),,(Pfn,Pdn)}, corresponding to the finite set of local decision rules {γ1,,γn}, is finite. For each combination of a monotonic global fusion rule and local operating points, we can calculate the operating point of the system (Pf,Pd) by using (6). We can then calculate all the operating points of the system by running a search on all the combinations of a monotonic global fusion rule and local operating points and find the optimal deterministic strategy satisfying the Neyman-Pearson criterion.

    Since the vector U is finite-dimensional (and binary), Λ(U) in (4) has a finite number of values, each with a corresponding probability of false alarm. The value of U that corresponds to the highest probability of false alarm β that satisfies βα may have a significant gap αβ compared to α.

    Fig. 2 shows the receiver operating characteristic (ROC) curve [12] of a system with discrete local observations (this curve comes from the system we will later introduce in Section III-A). All possible operating points of the system are shown as blue circles. The probability of false alarm constraint α is shown as the dash line. In this circumstance, the best operating point is ω3 and Pω3f<α.

    Figure  2.  Deterministic strategy with isolated operating points.

    Performance of the architecture of Fig. 1 under the circumstance such as the one described in Fig. 2 can benefit from randomization. By randomization we mean that one or more of the decision makers in the system (an LD or the DFC) is selecting its decision rule (γk or γ0) at each time instant by selecting one rule from a finite set of decision rules. The kth LD selects a rule from among {γ1k,,γik,,γNk}, for some positive integer N. The rule γik is selected with probability pik and Ni=1pik=1. The DFC selects a rule from among {γ10,,γi0,,γM0}, where γi0 is selected with probability pi0 and Mi=1pi0=1. Randomization will be considered when the specified probability of false alarm constraint α is not achievable by a deterministic strategy (such as in Fig. 2) or when the deterministic strategy achieves the values of false alarm constraint α but the system’s ROC curve is not concave.

    It is possible that only the DFC employs randomization (e.g., [2], [4], [7], [9]) or that a subset of the of LDs and the DFC employ randomization (independently or dependently). The term “dependent randomization” is used when both the LDs and the DFC employ randomization, and when, in addition, their switching between decision rules is coordinated and synchronized ([4]–[6]).

    One possible design has the system of Fig. 1 operate at one of two operating points, A and B. At each time step, one of the two is selected (A with probabilities p and B with probability 1p). “Operation at point A” means that the DFC selects γA0 and simultaneously each LD (k=1,2,,n) selects γAk. “Operation at point B” means that the DFC selects γB0 and simultaneously each LD (k=1,2,,n) selects γBk. By changing the value of p, the system can effectively operate anywhere along the line segment that connects A and B (every combinations of (Pf,Pd) along this line segment is realizable). We denote the operating point generated by the randomized strategy as C=(PCf,PCd), where

    PCf=pPAf+(1p)PBf (9)
    PCd=pPAd+(1p)PBd. (10)

    In order to satisfy the constraint on the probability of false alarm, we require PCfα. When PCf=α, the probability of selecting point A, p, is

    p=PBfαPBfPAf. (11)

    If γALD=γBLD (γAk=γBk for all k), the randomization occurs only at the DFC. If γALDγBLD and the selection of operating at point A and point B is coordinated and synchronized between the LDs and the DFC, then the scheme is known as dependent randomization.

    Fig. 3 shows how randomization connects the isolated operating points shown in the ROC curve of Fig. 2. For example, randomization allows the system to operate at point C, rather than at ω3, thereby achieving a higher probability of detection while not violating the constraint on probability of false alarm. The red curve is the ROC curve of the system employing randomization, which consists of straight line segments connecting all the previously-isolated operating points (the blue circles). In this example, to achieve the highest probability of detection subject to Pfα, we select A=ω3 and B=ω4. The operating point C achieved by randomization is shown as the black circle. PCf=α=pPω3f+(1p)Pω4f while p is calculated by (11).

    Figure  3.  Randomization can improve detection performance by “connecting” the isolated operating points.

    1) Randomization at the DFC Only: The authors of [2], [4], [7], [9], [13] studied strategies requiring that the DFC implement randomization, when the local decision rules are deterministic (γAk=γBk=γk,k). Each local decision rule is of the form (2). The DFC selects either γA0 or γB0 at each time step.

    Thomopoulos et al. [2] showed that under a Neyman-Pearson criterion, a desired value of global false alarm α, can always be achieved by a strategy with randomization at the DFC [2]. In [7] and [13] examples were presented to show that a strategy with randomization at the DFC is able to achieve higher probability of detection than the one achieved by a deterministic detection strategy.

    2) Dependent Randomization: In dependent randomization (or “a scheduled test” as it is called in [4]), both the DFC and the LDs participate in the randomization. At each time step, the system makes a selection between two deterministic strategies, γA={γA0,,γAn} and γB={γB0,,γBn} ([4]–[6]). The system can operate on the line segment connecting any two operating points realizable by the deterministic strategy. The ROC curve of the system with dependent randomization is the upper boundary of the convex hull of all the operating points achieved by the deterministic strategy. In other words, dependent randomization can make ROC curve of the system concave.

    Dependent randomization requires a coordinated action between the DFC and the LDs. The DFC and the LDs would switch simultaneously together, back and forth, between γA0 (for the DFC) and γALD={γA1,,γAn} (for the LDs); and γB0 (for the DFC) and γBLD={γB1,,γBn} (for the LDs). This synchronization challenge is discussed in [5, p. 301], [6], [9]. Among the means to achieve synchronization between the DFC and the LDs is the use of identical pseudo-code generators (or stored sequences of identical pseudo-code) at the DFC and the LDs.

    Strategy with randomization at the DFC only (Section II-B-1)) can be considered as a special case of dependent randomization, with γALD=γBLD. Randomization at the DFC only does not require synchronization between the DFC and the LDs but it does not necessarily result in a concave team ROC curve.

    Table I summarizes the input and output of three different designs of a parallel decentralized detection system of Fig. 1.

    Table  I.  Input and Output Of Three Different Designs of a Parallel Decentralized Detection System of Fig. 1 Under a Neyman-Pearson Criterion
    Input for the design
    1. The number of local detectors, n
    2. The probability of false alarm constraint, α
    3. Conditional probability distributions of the local observations, P(yk|H0) and P(yk|H1), k=1,,n
    Output of a design
    Deterministic strategy1. One global operating point (Pf,Pd)
    2. The corresponding local operating points, (Pfk,Pdk),k=1,,n
    Randomization at the DFC only1. Two global operating points A=(PAf,PAd) and B=(PBf,PBd)
    2. The corresponding local operating points of A and B: (PAfk,PAdk) and (PBfk,PBdk),k=1,,n
    3. The probability of selecting A, p, calculated by (11)
    4. The operating point C=(PCf,PCd), calculated by (9) and (10)
    The local operating points at A and B are identical (PAfk,PAdk)=(PBfk,PBdk),k=1,,n
    Dependent randomizationThe local operating points at A and B are different (PAfk,PAdk)(PBfk,PBdk),k=1,,n
     | Show Table
    DownLoad: CSV

    In the remaining of the paper, we will use the following two examples to illustrate performance under different strategies.

    We consider a system with two LDs (n=2). The local observations are identical logistic random variables (as done in [4]). The conditional probability distribution of the local observations are

    {P(yk|H0)=14sech2(yk2)P(yk|H1)=14sech2(yk2.52). (12)

    The operating point (Pfk,Pdk) of the kth LD (k=1,2) can be calculated as

    {Pfk=τk14ssech2(yk2)dyk=1212tanh(τk2)Pdk=τk14ssech2(yk2.52)dyk=1212tanh(τk2.52) (13)

    where τk is a function of yk

    τk=yk|Λ(yk)=ηk. (14)

    We want to design the system under a Neyman-Pearson criterion with α=0.2009 (this is an arbitrary choice; any value between 0 and 1 can be used). The nontrivial (monotonic) global decision rules are the AND rule (u0=u1&u2) and the OR rule (u0=u1|u2) [4]. For each global decision rule, the operating points of the two LDs are identical, (Pf1,Pd1)=(Pf2,Pd2). The system operating points can be shown to be: (Pf,Pd)=((Pf1)2,(Pd1)2) under the AND rule; and (Pf,Pd)=((Pf1)2+2Pf1(1Pf1),(Pd1)2+2Pd1(1Pd1)) under the OR rule.

    In Fig. 4, the ROC curves of this system, using the AND rule and the OR rule, are shown in red and blue, respectively. The ROC curve of the AND rule is given by

    Figure  4.  The ROC curves of the 2-LD system when the DFC uses an AND rule (red curve) and when it uses an OR rule (blue curve). The upper boundary of the two curves is the ROC curve of the system with deterministic strategy.
    Pd=(1212tanh βAND2.52)2,whereβAND=lnPfPfPf. (15)

    The ROC curve of the OR rule is given by

    Pd=1(12+12tanh βOR2.52)2,whereβOR=ln1+1PfPfPf. (16)

    Although the two individual ROC curves are both concave, the team ROC curve of the system (which is the upper boundary of the two curves) is not. The point of intersection of the two ROC curves is O=(POf,POd)=(0.1859,0.8141).

    Referring to Fig. 4, if the 2-LD system uses a deterministic strategy to achieve the highest possible Pd, then the DFC employs an AND rule when the desired α is less than or equal to POf and employs an OR rule when the the desired α is greater than POf.

    Referring to Fig. 5 (which shows the ROC curve for Pf[0.14,0.28]), let A be the point where the common tangent of the two ROC curves touches the “AND rule ROC curve.” Let B be the point where the common tangent of the two ROC curves touches the “OR rule ROC curve.” If the maximum allowable value of the probability of false alarm, α, satisfies PAf<α<PBf, then dependent randomization would be useful. The ROC curve of the system with dependent randomization as the black curve. In Fig. 5, for the 2-LD system, the points of tangency are A=(PAf,PAd)=(0.1581,0.7870) on the “AND rule ROC curve” and B=(PBf,PBd)=(0.2437,0.8652) on the “OR rule ROC curve.” The system operates at A when both LDs operate at (PAf1,PAd1)=(PAf2,PAd2)=(0.3976,0.8871) and simultaneously the DFC uses the AND rule. The system operates at B when both LDs operate at (PBf1,PBd1)=(PBf2,PBd2)=(0.1304,0.6328) and simultaneously the DFC uses the OR rule. To make the team ROC curve concave, we apply dependent randomization whenever the desired probability of false alarm α satisfies 0.1581=PAf<α<PBf=0.2437. In this range, at each time step the system operates at A with probability p and at B with probability 1p. The equivalent operating point C on the line segment AB is provided by (9) and (10). Otherwise, if α<PAf we use the AND rule, and if PBf<α we use the OR rule.

    Figure  5.  The operating points of the 2-LD system employing different detection strategies under the Neyman-Pearson criterion with α=0.2009: (a) deterministic strategy (and randomization at the DFC) (blue circle); (b) dependent randomization (black circle).

    Fig. 5 shows the operating points of the 2-LD system employing different detection strategies. The blue circle G=(PGf,PGd)=(0.2009,0.8217) is the operating point of the system employing deterministic strategy. The system operates at G when both LDs operate at (PGf1,PGd1)=(PGf2,PGd2)=(0.4482,0.9065) and the DFC uses the AND fusion rule.

    In this example the operating point achieved by the system employing Randomization at the DFC only is also G.

    The value of α (α=0.2009) in our case satisfies 0.1581=PAf<α<PBf=0.2437. Therefore, dependent randomization can improve the probability of detection of the system at that value of α. The black circle C=(PCf,PCd)=(0.2009,0.8261) is the operating point of the system employing dependent randomization. C is generated by operating at A=(PAf,PAd)=(0.1581,0.7870) with probability p=0.5 and at B=(PBf,PBd)=(0.2437,0.8652) with probability 1p=0.5. p=0.24370.20090.24370.1581, calculated by (11).

    Under the Neyman-Pearson criterion with α=0.2009, the input and output of three different designs of the 2-LD system (corresponding to Table I) are shown in Table II. As shown in Table II, dependent randomization is able to increase the probability of detection from Pd=0.8217 to Pd=0.8261 with the same probability of false alarm Pf=0.2009.

    Table  II.  Input and Output of Three Different Designs of a 2-LD System
    Input for the design
    1. The number of local detectors, n=2
    2. The probability of false alarm constraint, α=0.2009
    3. Conditional probability distributions of the local observations, P(yk|H0) and P(yk|H1), k=1,2, shown in (12)
    Output of a design
    Deterministic
    strategy
    1. System operating point G=(PGf,PGd)=(0.2009,0.8217)
    2. The system operates at G when both LDs operate at (PGf1,PGd1)=(PGf2,PGd2)=(0.4482,0.9065) and the DFC uses the AND fusion rule
    Randomization at
    the DFC
    Same as deterministic strategy (Randomization at the DFC does not improve the system performance since the local observations are continuous)
    Dependent
    randomization
    1. Two operating points A=(PAf,PAd)=(0.1581,0.7870) and B=(PBf,PBd)=(0.2437,0.8652)
    2. The system operates at A when both LDs operate at (PAf1,PAd1)=(PAf2,PAd2)=(0.3976,0.8871) and the DFC uses the AND fusion rule The system operates at B when both LDs operate at (PBf1,PBd1)=(PBf2,PBd2)=(0.1304,0.6328) and the DFC uses the OR fusion rule
    3. The probability of selecting A is p=0.5
    4. The resulting operating point is C=(PCf,PCd)=(0.2009,0.8261)
     | Show Table
    DownLoad: CSV

    We consider the 3-LD implementation of the network shown in Fig. 1 (n=3). The local observations of the three LDs in the system have identical discrete probability distributions, as shown in Fig. 6, where the conditional probabilities P(yk|Hi) are given for k=1,2,3, and i=0,1.

    Figure  6.  The conditional probability distributions of the local observations.

    We assume that the local observations are statistically independent, conditioned on the hypothesis. From Fig. 6, each LD has 4 distinct local decision rules, corresponding to 4 distinct local observation thresholds, τ1 (anywhere in the range τ1<0), τ2 (0<τ2<1), τ3 (1<τ3<2),τ4 (τ4>2). At the kth LD, k=1,2,3, if τ1 is used, Pfk=1,Pdk=1; if τ2 is used, Pfk=0.2,Pdk=0.7; if τ3 is used, Pfk=0.1,Pdk=0.6; if τ4 is used, Pfk=0,Pdk=0. Each LD therefore has 4 possible local operating points (Pfk,Pdk), which are (0,0), (0.1,0.6), (0.2,0.7), and (1,1).

    Under a Neyman-Pearson criterion with the probability of false alarm constraint α=0.1708, the input and output of three designs of this 3-LD system are given in Table III. The table shows that in order to satisfy the probability of false alarm constraint Pfα, the deterministic strategy provides a design with Pd=0.7960; randomization at the DFC improves the probability of detection to Pd=0.8208; dependent randomization improves the probability of detection further to Pd=0.8448.

    Table  III.  Input and Output of Three Different Designs of a 3-LD System
    Input for the design
    1. The number of local detectors, n=3
    2. The probability of false alarm constraint, α=0.1708
    3. Conditional probability distributions of the local observations, P(yk|H0) and P(yk|H1), k=1,2,3, shown in Fig. 6
    Output of a design
    Deterministic strategy1. System operating point G=(PGf,PGd)=(0.1360,0.7960)
    2. One way to achieve G is that the LDs operate at {(0.1,0.6),(0.2,0.7),(0.2,0.7)} while the DFC uses the fusion rule u0=u1|(u2&u3)
    Randomization at the DFC1. Two operating points A=(PAf,PAd)=(0.1180,0.7680) and B=(PBf,PBd)=(0.1900,0.8400)
    2. When the LDs operate at (0.2,0.7),(0.2,0.7),(0.1,0.6), the DFC can use the fusion rule u0=u2|(u1&u3) to achieve point A=(0.1180,0.7680) and use the fusion rule u0=u1|u2 to achieve point B=(0.1900,0.8400)
    3. The probability of selecting A is p=0.2667
    4. The resulting operating point is E=(PEf,PEd)=(0.1708,0.8208)
    Dependent randomization1. Two operating points A=(PAf,PAd)=(0.104,0.784) and B=(PBf,PBd)=(0.271,0.936)
    2. A is achieved when (Pfk,Pdk)=(0.2,0.7),k=1,2,3, and the DFC uses a “2 out of 3 rule”, B is achieved when (Pfk,Pdk)=(0.1,0.6),k=1,2,3, and the DFC uses a “1 out of 3 rule”
    3. The probability of selecting A is p=0.6
    4. The resulting operating point is C=(PCf,PCd)=(0.1708,0.8448)
     | Show Table
    DownLoad: CSV

    Fig. 7 shows the ROC curves of the 3-LD system with three different designs for 0.075<Pf<0.3. The operating points of the 3-LD system employing (a) deterministic strategy is G=(0.1360,0.7960), shown by the blue circle; (b) randomization at the DFC is E=(0.1708,0.8208), shown by the red circle; and (c) deterministic strategy is C=(0.1708,0.8448), shown by the black circle.

    Figure  7.  The operating points of the 3-LD system employing different detection strategies under the Neyman-Pearson criterion with probability of false alarm constraint α=0.1708: (a) deterministic strategy (blue circle); (b) randomization at the DFC (red circle); (c) dependent randomization (black circle). The ROC curves of the 3-LD system employing different detection strategies: (a) deterministic strategy (blue); (b) randomization at the DFC (red); (c) dependent randomization (black).

    More details of the numerical examples are available in https://github.com/moshekam/Dependent-Randomization.

    Dependent randomization assumes synchronization between all the LDs and the DFC. When the synchronization is lost, unless a corrective action is taken, the system may exceed the allowed probability of false alarm α for which it was designed under a Neyman-Pearson criterion.

    We will demonstrate the approach to corrective action on the 2-LD example in Section III-A (also see [9]). The ROC curves for the AND rule and the OR rule are shown in Figs. 8 and 9 (also see Fig. 4). Recall that A and B are, respectively, the points of tangency of the original AND rule and OR rule ROC curves. We assume that the 2 LDs are still synchronized with each other but the synchronization between the DFC and the group of LDs was lost. Under these circumstances, the DFC selects γA0 with probability p and γB0 with probability 1p. The group of LDs select γA1,,γAn with probability p and γB1,,γBn with probability 1p. However the DFC selection is not coordinated with the selection of the group of LDs. There are four possible detection strategies: (γA0,γALD), (γB0,γBLD), (γB0,γALD), and (γA0,γBLD). They correspond, respectively, to the four operating points A, B, M1, and M2, shown by green circles in Fig. 8. Operating point A is selected with probability p2, operating point B is selected with probability (1p)2, operating point M1 is selected with probability (1p)p, operating point M2 is selected with probability p(1p).

    Figure  8.  A, B, M1, and M2, shown by green circles, are the possible operating points of the 2-LD system (Section III-A) when the synchronization between the LDs and the DFC is lost. The black circle, C, shows the operating point of the synchronized system. The cyan circle, W, shows the equivalent operating point of the system when it lost synchronization. C is the equivalent operating point after the corrective action is taken.
    Figure  9.  Zooming in on the ROC curve of the 2-LD system employing dependent randomization.

    Let W represent an equivalent operating point which results from this combination. It is a weighted average of the four points A, B, M1, and M2. For this operating point W, the probability of false alarm PWf and the probability of detection PWd are

    PWf=p2PAf+(1p)2PBf+(1p)pPM1f+p(1p)PM2f (17)
    PWd=p2PAd+(1p)2PBd+(1p)pPM1d+p(1p)PM2d. (18)

    In Fig. 8, the operating point W is shown as a cyan circle. It is possible that PWf>α, where α was our upper bound for the system’s probability of false alarm (we wanted Pfα).

    A special case occurs when the global fusion rules at point A and point B are the same, γA0=γB0. In this case, the DFC does not participate in the randomization. Points A and M1 would be the same; points B and M2 would be the same. The operating point W=(PWf,PWd), calculated by (17) and (18), would be exactly the operating point of the system employing dependent randomization (without losing synchronization), C=(PCf,PCd), calculated by (9) and (10). No corrective action is needed in this situation. In all other cases, a corrective action by the DFC may help.

    If the DFC realizes that the synchronization with the group of LDs was lost, it may have the opportunity to take a corrective action to try to satisfy the probability of false alarm constraint Pfα. The DFC can do so by changing the probability of selecting point A from p to a certain q,0q1. Let the DFC choose γA0 with probability q (which may be different from p that was used for (17) and (18)) and γB0 with probability 1q. A new operating point C=(PCf,PCd) is created with

    PCf=pqPAf+(1p)(1q)PBf+p(1q)PM1f+(1p)qPM2f (19)
    PCd=pqPAd+(1p)(1q)PBd+p(1q)PM1d+(1p)qPM2d. (20)

    The role of p in (19) and (20) is due to the continued use of the probability p (calculated before the loss of synchronization) to select the local decision rules at the operating point A by the LDs. The role of q is due to the selection of the global decision rule at the operating point A by the DFC with probability q (calculated after the loss of synchronization).

    Let Q0 be the operating point when q=0 and Q1 be the operating point when q=1

    PQ0f(q=0)=(1p)PBf+pPM1f (21)
    PQ0d(q=0)=(1p)PBd+pPM1d (22)
    PQ1f(q=1)=pPAf+(1p)PM2f (23)
    PQ1d(q=1)=pPAd+(1p)PM2d. (24)

    Q0 is located on the line segment connecting B and M1; Q1 is located on the line segment connecting A and M2. Both PCf and PCd are affine functions of q (see (19) and (20)). By changing the value of q from 0 to 1, the system operating point will move from Q0 to Q1. Both W and C are located on the line segment connecting Q0 and Q1. By selecting an appropriate value of q to determine C, the probability of false alarm constraint Pfα may still be satisfied by point C (shown by cyan square in Figs. 8 and 9), but with a lower probability of detection (PCd) compared to PCd (that was calculated for the synchronized system, see (9) and (10)).

    The value of q, the new probability that determines how the DFC hops between γA0 and γB0, can be derived from (19) and (20)

    q=PQ0fαPQ0fPQ1f=[PBfp(PBfPM1f)]α[PBfp(PBfPM1f)][PM2f+p(PAfPM2f)]. (25)

    It is usable only if 0q1. Otherwise, a corrective action is not possible.

    Recall that before we lost synchronization, dependent randomization would be useful if the probability of false alarm constraint α(PAf,PBf) (assuming PBf>PAf). Satisfying the probability of false alarm constraint after losing synchronization (when the DFC selects γA0 with probability q) requires that αPQ0f if PQ0f<PQ1f; or αPQ1f if PQ1fPQ0f.

    Using (25), the conditions on the new probability of randomization at the DFC q which satisfies PCf<α after the loss of synchronization are summarized in Table IV.

    Table  IV.  Conditions on q (Probability That the DFC Selects γa0) to Satisfy the Neyman-Pearson Constraint After the DFC Loses Synchronization With the LDs Group
    Value of αExistence of qValue of PCf
    PQ0f<PQ1fα(,PQ0f)q does not exist
    α[PQ0f,PQ1f]q[0,1]PCf=α
    α(PQ1f,)q=1PCf=PQ1f
    PQ0fPQ1fα(,PQ1f)q does not exist
    α[PQ1f,PQ0f]q[0,1]PCf=α
    α(PQ0f,)q=0PCf=PQ0f
     | Show Table
    DownLoad: CSV

    So far we have referred to the 2-LD example in Section III-A. In the general case, points A and B reside on two different ROC curves (each one corresponds to a different decision rule of the DFC). A represents the deterministic operating point of the system with one of the possible deterministic strategy γA (the one that corresponds to (PAf,PAd)). B represents the deterministic operating point of the system with one of the possible deterministic strategy γB (the one that corresponds to (PBf,PBd)). A and B reside on the joint tangent to the two ROC curves, and represent the end points of a line segment on which an operating point resides that satisfies a probability of false alarm constraint α (PAf<α<PBf) while maximizing the probability of the detection using dependent randomization. When the synchronization between the DFC and the LDs is lost, we can satisfy the probability of false alarm constraint α by assigning a new probability of randomization for the DFC, q (calculated by (25)).

    1) 2-LD System: The design input and output of the 2-LD system employing dependent randomization with α=0.2009 was shown in Table II. When the DFC lost synchronization with the LDs group in the 2-LD system employing dependent randomization, the input and output of the redesigned algorithm are shown in Table V and Fig. 8. Before the loss of synchronization the system operated at C=(0.2009,0.8261) (Table II). After the loss of synchronization the system operates at C=(0.2009,0.7005).

    Table  V.  The Output of the 2-LD System Employing Dependent Randomization When the DFC Lost Synchronization With the LDs Group Before and After a Corrective Action is Taken
    Output of the non-synchronized 2-LD system before the corrective action is taken (α=0.2009)
    1. The probability of false alarm constraint, α=0.2009
    2. The probability of selecting A, p=0.5, calculated by (11)
    3. Four possible operating points when the DFC lost synchronization with the LDs group, A=(0.1581,0.7870), B=(0.2437,0.8652), M1=(0.6371,0.9873), and M2=(0.0170,0.4004)
    4. The operating point of the non-synchronized system, W=(0.2640,0.7600), calculated by (17) and (18)
    Output of the non-synchronized 2-LD system after the corrective action is taken (α=0.2009)
    1. The new probability for the DFC selecting γA0, q=0.6787, calculated by (25)
    2. The fulfillment of the prerequisite of the correction action, 0<q<1
    3. The operating point of the non-synchronized system after the corrective action is taken, C=(0.2009,0.7005), calculated by (19) and (20)
     | Show Table
    DownLoad: CSV

    2) 3-LD System: The design input and output of the 3-LD system employing dependent randomization with α=0.1708 was shown in Table III. When the DFC lost synchronization with the LDs group in the 3-LD system employing dependent randomization, the input and output of the redesigned algorithm are shown in Table VI and Fig. 10. Before the loss of synchronization the system operated at C=(0.1708,0.8448) (Table II). After the loss of synchronization the system operates at C=(0.1708,0.7974).

    Table  VI.  THe Output of the 3-LD System Employing Dependent Randomization When the DFC Lost Synchronization With the LDs Group Before and After a Corrective Action is Taken
    Output of the non-synchronized 3-LD system before the corrective action is taken (α=0.1708)
    1. The probability of false alarm constraint, α=0.1708
    2. The probability of selecting A, p=0.6, calculated by (11)
    3. Four possible operating points when the DFC lost synchronization with the LDs group, A=(0.1040,0.7840), B=(0.2710,0.9360), M1=(0.4880,0.9730), and M2=(0.0280,0.6480)
    4. The operating point of the non-synchronized system, W=(0.2046,0.8210), calculated by (17) and (18)
    Output of the non-synchronized 3-LD system after the corrective action is taken (α=0.1708)
    1. The new probability for the DFC selecting γA0, q=0.7033, calculated by (25)
    2. The fulfillment of the prerequisite of the correction action, 0<q<1
    3. The operating point of the non-synchronized system after the corrective action is taken, C=(0.1708,0.7974), calculated by (19) and (20)
     | Show Table
    DownLoad: CSV
    Figure  10.  A, B, M1, and M2, shown by green circles, are the possible operating points of the 3-LD system (Section III-A) when the synchronization between the LDs and the DFC is lost. The black circle, C, shows the operating point of the synchronized system. The cyan circle, W, shows the equivalent operating point of the system when it lost synchronization. C is the equivalent operating point after the corrective action is taken.

    More details of the numerical examples are available in https://github.com/moshekam/Dependent-Randomization.

    We assume that a decision fusion architecture was designed per Section II-B-2) (Dependent randomization) to maximize the probability of detection under a probability of false alarm constraint. This approach means that the LDs and the DFC are designed to operate at any given time at one of two operating points (say A and B). They operate at operating point A (corresponding to deterministic strategy γA={γA0,γA1,,γAn}) with probability p and at operating point B (corresponding to deterministic strategy γB={γB0,γB1,,γBn}) with probability 1p. A and B are on the upper boundary of the convex hull of all the operating points which are achievable by deterministic strategies of the system. If a value of p, 0<p<1, exists that would keep the probability of false alarm of the system at the maximal allowable level, α, the resulting operating point of the system C=(PCf=α,PCd) would be optimal.

    In the previous section (IV: Loss of Synchronization Between the DFC and the LDs Group), we assumed that in spite of loss of synchronization between the DFC and the LDs group, all LDs were still synchronized with each other. Under this circumstance, the DFC can in some cases change the probability p of hopping between A and B to satisfy the probability of false alarm constraint, but generally at the expense of reaching a lower probability of detection than PCd.

    In this section, we assume a synchronization failure of the following characteristics:

    a) Only m (1mn1) LDs are synchronized with the DFC and with each other. We call them group Y.

    b) The remaining nm LDs are not synchronized with the DFC, nor are they synchronized with each other. We call them group ¯Y.

    c) The DFC is aware of a) and b) and of the identity of members in Y and ¯Y. The LDs are not.

    Each LD of the system (say LD k) flips a coin, and, based on the outcome, it follows the local decision rule γAk or the other local decision rule, γBk (γAk is used with probability p and γBk is used with probability 1p). If LD k belongs to Y it uses the “joint coin” flipped simultaneously and synchronously by all the LDs in Y and the DFC. If LD k belongs to ¯Y then it flips its own coin, which is not synchronized with either the “joint coin” used by the LDs in Y and the DFC, or the “separate” nm1 coins of the other members of ¯Y.

    Since nm LDs are now unsynchronized, the resulting operating point of the system, W=(PWf,PWd), is a combinations of 2nm+1 possible operating points. If no correction is made, PWf is highly likely to exceed the level α which was satisfied (per the constraint Pfα) before synchronization was lost. In this section, we redesign the global fusion rules at the DFC to satisfy Pfα, and show the resulting performance cost (the reduction in the probability of detection).

    The input of the redesigned algorithm is shown in Table VII.

    Table  VII.  Input of the Redesigned Algorithm When Each LD in ¯Y Lost Synchronization the DFC and Each Other
    Input of the redesigned algorithm
    The design input of dependent randomization (Table I)1. The number of local detectors, n
    2. The probability of false alarm constraint, α
    The design output of dependent randomization (Table I)3. The local operating points of A and B: (PAfk,PAdk) and (PBfk,PBdk),k=1,,n
    4. The probability of selecting A, p, calculated by (11)
    Information of synchronized LDs5. Numbers of synchronized LDs, m (we assume Y={LD1,,LDm} are synchronized)
    6. Identity of all synchronized LDs, Y={LD1,,LDm}
     | Show Table
    DownLoad: CSV

    The local operating points of the m LDs in Y are {(Pif1,Pid1),,(Pifm,Pidm)}, i{A,B}. For the jth LD in ¯Y (j=m+1,,n), the expected value of the probability of false alarm is pPAfj+(1p)PBfj and the expected value of the probability of detection is pPAdj+(1p)PBdj. The local operating points of the nm LDs in ¯Y are {(pPAfm+1+(1p) PBfm+1,pPAdm+1+(1p)PBdm+1),,(pPAfn+(1p)PBfn,pPAdn+(1p) PBdn)}. The equivalent local operating points of the system are Φi={(Pif1,Pid1),,(Pifm,Pidm),(pPAfm+1+(1p)PBfm+1,pPAdm+1+(1p)PBdm+1),,(pPAfn+(1p)PBfn,pPAdn+(1p)PBdn)}. Therefore, at each time step, we use system A, shown in Fig. 11, with probability p, and system B, shown in Fig. 12, with probability 1p. The local operating points of {LD m+1,,LD n} in system A and system B are the same.

    Figure  11.  System A is used with probability p.
    Figure  12.  System B is used with probability 1p.

    In order to satisfy the Neyman-Pearson criterion, under the new condition we can try to redesign the global fusion rules at the DFC. At each time step, the DFC will now use γA0 (was γA0) with probability p and γB0 (was γB0) with probability 1p. Namely, γA0 will be used by system A and γB0 will be used by system B. Unlike γA0 and γB0, which are deterministic fusion rules, γA0 or γB0 could be a randomized fusion rule. We will select γA0 and γB0 from among the monotonic fusion rules [2] for both systems. For each one of γA0 and γB0, we need in general two monotonic fusion rules and a randomization scheme to hop between them.

    We use (6) to calculate the operating points, corresponding to all the monotonic fusion rules, for both system A and system B. The calculated operating points are isolated in the PfPd plane. We denote ROC curve A (ROC curve B) as the ROC curve of system A (system B). The ROC curve of a system with isolated operating points in the PfPd plane is the upper boundary of the convex hull of those isolated operating points, which is a concave piecewise-linear curve. Therefore both ROC curve A and ROC curve B are concave piecewise-linear curves. Finding the fusion rule γA0 for system A is equivalent to finding an operating point of system A on ROC curve A. Similarly, finding a fusion rule γB0 for system B is equivalent to finding an operating point of system B on ROC curve B.

    ROC curve A can be drawn as a sequence of straight line segments ωA1ωA2,ωA2ωA3,,ωAmA1ωAmA, where ΩA={ωA1=(0,0),ωA2,,ωAmA1,ωAmA=(1,1)} are points in the PfPd plane. Similarly, we can find ΩB={ωB1=(0,0),ωB2,,ωBmB1,ωBmB=(1,1)} for the ROC curve B. Each of points in ΩA and ΩB is realizable by a deterministic (monotonic) fusion rule. Meanwhile, each one of the other operating points on ROC curve A (ROC curve B), those that are not in ΩA(ΩB), can be realized by hopping between two points in ΩA (ΩB) by using randomization at the DFC.

    In Fig. 13, the ROC curve of the 2-LD system (Section III-A) with dependent randomization is shown as the black curve. Recall (Figs. 4 and 5) that to create this curve, we have integrated two different ROC curves (one corresponding to an AND global decision rule and one corresponding to an OR global decision rule). We calculated the points of tangency A and B of the ROC curves for the AND rule and the OR rule respectively, and connected them with a straight line. If α(PAf,PBf) then the highest achievable probability of detection, corresponding to α, was on the straight-line segment that joins points A and B. For α shown in Fig. 5, the highest probability of detection is denoted PCd, achieved at point C, which is the the midpoint of line segment connecting A and B (in this example p=0.5).

    Figure  13.  The ROC curve of the system with dependent randomization is shown by the black curve. For α(PAf,PBf), C is the desired operating point of the system (black circle). A and B (green circles) are the operating points used to generate C through a randomization procedure. When the second LD loses synchronization (Y={LD1}, ¯Y={LD2}), if A is selected the ROC curve A is effective (shown in red); if B is selected the ROC curve B is effective (shown in blue).

    Suppose that LD1 and the DFC continue to be synchronized with each other (Y={LD1}) but LD2 lost synchronization with LD1 and with the DFC (¯Y={LD2}). In Fig. 13 we show two ROC curves for this new situation. ROC curve A (red) is obtained when the members of Y (just LD1 in our case) select γA1 (when this happen LD2, oblivious to LD1 and the DFC, selects γA2 with probability p and γB2 with probability 1p). ROC curve B (blue) is obtained when the members of Y (LD1) select γB1 (LD2 still selects γA2 with probability p and γB2 with probability 1p). In Fig. 13, ΩA are shown by the red circles and ΩB are shown by the blue circles. Our next task is to select the points A (on ROC curve A) and B (on ROC curve B) such that we can hop between them and meet the following objectives: a) satisfy the probability of false alarm constraint Pfα; b) maximize the probability of detection Pd.

    Let a=(Paf,Pad) be an operating point of system A on ROC curve A and b=(Pbf,Pbd) be an operating point of system B on ROC curve B. Points a and b are selected such that upper limit of the probability of false alarm of the system, α, satisfies PafαPbf or PbfαPaf. In order to meet the probability of false alarm constraint (Pf=α), points a and b need to be found such that

    α=pPaf+(1p)Pbf. (26)

    This design would yield the probability of detection

    Pd=pPad+(1p)Pbd. (27)

    We now proceed to locate the specific point on ROC curve A, denoted A (so a=A), and the specific point on ROC curve B, denoted as B (so b=B) that allow the system to maximize the probability of detection while satisfying the probability of false alarm constraint. The optimal resulting system operating point C=(PCf,PCd) is on the line segment connecting A and B, where

    PCf=pPAf+(1p)PBf (28)
    PCd=pPAd+(1p)PBd. (29)

    It can be shown that for PCd in (29) to be the maximum probability of detection either i) AΩA or ii) BΩB or both (AΩA and BΩB). The proof can be found in https://github.com/moshekam/Dependent-Randomization. Here we provide an intuitive explanation: the resulting operating point of the system would be on the line segment connecting a point on ROC curve A and a point on ROC curve B. We can “lift” the line segment to improve the probability of detection of the resulting operating point. During the process of “lifting”, the length of the line segment is adjustable to keep its endpoints on different ROC curves. The resulting probability of detection would stop growing when the line segment is about to “leave” the ROC curves. The line segment can only “leave” the ROC curve A at a point in ΩA and the ROC curve B at a point in ΩB.

    We can examine every point in ΩA to find A and every point in ΩB to find B by using the following steps:

    1) For each operating point a=(Paf,Pad)ΩA, we calculate first the probability of the false alarm of the corresponding operating point b=(Pbf,Pbd) on ROC curve B by using (26) (Pbf=αpPaf1p). Since each probability of detection is paired with exactly one probability of false alarm on ROC curve B, Pbf can be used to locate b on ROC curve B and define Pbd. The resulting probability of detection of the system can be calculated by using (27). We store the resulting probability of detection and the corresponding a and b for each aΩA.

    2) For each operating point b=(Pbf,Pbd)ΩB, we calculate first the probability of the false alarm of the corresponding operating point on ROC curve A (Paf=α(1p)Pbfp) by using (26). Paf can be used to locate a on ROC curve A and define Pad. The resulting probability of detection of the system can be calculated by using (27). We store the resulting probability of detection and the corresponding a and b for each bΩB.

    3) Let PCd be the highest probability of detection found for all the pairs examined in steps 1) and 2). A and B are the pair of values of a and b which corresponded to the highest PCd for the final design.

    Computational complexity: in steps 1) and 2) we examine mA+mB points (mA points in ΩA and mB points in ΩB) in order to find A if AΩA and B if BΩB. An algorithm which requires the examination of at most log2(mA+mB) points in ΩA and ΩB is available (https://github.com/moshekam/Dependent-Randomization).

    In most cases, when PCd is maximized, one of the points A and B is realized by randomization at the DFC and the other is realized by a deterministic strategy (since it is an element of ΩA or ΩB). We first discuss the situation that B is a randomized operating point, AΩA, and BΩB. In this circumstance PBf=αpPAf1p (from (26)). B is on a line segment connecting two operating points in ΩB, denoted as ωBa and ωBb.

    Let the probabilities needed to redesign B by hopping between using ωBa and ωBb be q and 1q, respectively. Then PBf can be expressed as

    PBf=qPωBaf+(1q)PωBbf. (30)

    Therefore, PCf is a weighted sum of PAf, PωBaf, and PωBbf

    PCf=pPAf+(1p)[qPωBaf+(1q)PωBbf]. (31)

    Since PCf=α, q can be calculated as

    q=αpPAf(1p)PωBbf(1p)(PωBafPωBbf). (32)

    The probability of detection PCd is

    PCd=pPAd+p[qPωBad+(1q)PωBbd]. (33)

    Similarly, when A is a randomized operating point (AΩA and BΩB), PCd becomes

    PCd=(1p)PBd+p[qPωAad+(1q)PωAbd] (34)

    ωAa and ωAb are the two end points of the line segment on the ROC curve A which A locates on. Let q and 1q be the probabilities of using ωAa and ωAb respectively. q satisfies

    α=(1p)PBf+p[qPωAaf+(1q)PωAbf]. (35)

    q is

    q=α(1p)PBfpPωAbfp(PωAafPωAbf). (36)

    In the case AΩA and BΩB, both A and B are realized by deterministic strategies. The probability of detection can be calculated by (29).

    1) 2-LD System: In Fig. 14, the black curve is the ROC curve of the 2-LD example in Section III-A when using dependent randomization. The design input and output of the 2-LD system employing dependent randomization with α=0.2009 was shown in Table II. The operating point of the system is C=(0.2009,0.8261), shown by the black circle. The operating point C is generated by operating at A=(0.1581,0.7870) with probability p=0.5 and at B=(0.2437,0.8652) with probability 1p.

    Figure  14.  C is the desired operating point of the system with dependent randomization (black circle). When Y={LD1} and ¯Y={LD2}, A, B, M1, and M2 are the four possible operating points (green circles). W is the equivalent operating point (purple circle). The ROC curve A is shown as the red curve. The ROC curve B is shown as the blue curve. C is the operating point with maximized probability of detection given α=0.2009, shown by the purple square.

    If Y={LD1}, ¯Y={LD2}, there are four possible operating points: A, B, M1=(0.0518,0.5614), and M2=(0.4761,0.9586), determined by (γA0,γA1,γA2), (γB0,γB1,γB2), (γA0,γA1,γB2), and (γB0,γB1,γA2). A, B, M1, and M2 are shown by the green circles. The resulting operating point W=(0.2324,0.7930), shown by the purple circle is calculated as

    PWf=p2PAf+(1p)2PBf+p(1p)PM1f+(1p)pPM2f (37)
    PWd=p2PAd+(1p)2PBd+p(1p)PM1d+(1p)pPM2d. (38)

    In this case, ΦA={(PAf1,PAd1),(pPAf2+(1p)PBf2,pPAd2+(1p)PBd2)}. The ROC curve A is shown as the red curve; ΦB={(PBf1,PBd1),(pPAf2+(1p)PBf2,pPAd2+(1p)PBd2)}. The ROC curve B is shown as the blue curve. ΩA and ΩB then can be found, shown in Table VIII. By using the proposed algorithm, we find that when A=(0.1049,0.6742) and B=(0.2968,0.8410), the probability of detection is maximized. AΩA and BΩB. A is shown as the cyan circle. It is achieved when the DFC uses the AND fusion rule (γA0(u1,u2)=u1&u2). PBf can be calculated by (26) and B is shown by the purple triangle. B is generated by the randomized fusion rule γB0, which requires the system hopping between ωBa=(0.1304,0.6328) and ωBb=(0.3599,0.9199). ωBa (achieved by the fusion rule such that u0=u1) is used with probability q and ωBb (achieved by the OR fusion rule) is used with probability 1q, where q=0.2748, calculated by (32). ωBa and ωBb are shown as the cyan triangles.

    Table  VIII.  the Output the 2-LD System Employing Dependent Randomization Before and After a Corrective Action is Taken When Y={LD1} and ¯Y={LD2}
    Input of the redesigned algorithm for the 2-LD system when LD2 lost synchronization
    The design input of dependent randomization (Table II)1. The number of local detectors, n=2
    2. The probability of false alarm constraint, α=0.2009
    The design output of dependent randomization (Table II)3. The local operating points of A: (PAf1,PAd1)=(PAf2,PAd2)=(0.3976,0.8871),
     The local operating points of B: (PBf1,PBd1)=(PBf2,PBd2)=(0.1304,0.6328)
    4. The probability of selecting A, p=0.5
    Information at the synchronized LDs5. Numbers of synchronized LDs, m=1
    6. Identity of all synchronized LDs, Y={LD1}
    Output of the redesigned algorithm for the 2-LD system when LD2 lost synchronization
    1. ΩA={(0,0),(0.1049,0.6742),(0.3976,0.8871),(0.5566,0.9729),(1,1)}, ΩB={(0,0),(0.0344,0.4809),(0.1304,0.6328),(0.3599,0.9119),(1,1)}
    2. Two operating points A=(0.1049,0.6742)ΩA and B=(0.2968,0.8410)ΩB, which allow PCf from (28) satisfying the probability of false alarm constraint α and achieving the highest probability of detection
    3. The deterministic fusion rule γA0(AND fusion rule) used to achieve A
    4. The randomized fusion rule γB0 used to achieve B, which requires the system operating at ωBa=(0.1304,0.6328)(achieved by the fusion rule such that u0=u1) with probability q and ωBb=(0.3599,0.9199)(achieved by the OR fusion rule) with probability 1q, where q=0.2748, calculated by (32)
    5. The operating point of the non-synchronized system after the corrective action is taken, C=(0.2009,0.7547), calculated by (31) and (33), which is achieved by operating at A with probability p and B with probability 1p
     | Show Table
    DownLoad: CSV

    In this example, when the DFC realizes that the LD2 loses synchronization, if γA1 is selected at the LD1 (p=0.5), the system operates at point A; if γB1 is selected at the LD1 (1p=0.5), the system operates at point ωBa with probability q=0.2748 and operates at point ωBb with probability 1q=0.7252. The maximized probability of detection is PCd=0.7547, calculated from (29). C=(0.2009,0.7547) is shown by the purple square. Due to the loss of synchronization we drop from PCd=0.8261 to PCd=0.7547.

    Table VIII provides a summary of the input and output of the redesigned algorithm for the 2-LD system employing dependent randomization when LD2 lost synchronization.

    Fig. 15 and Table IX compare the operating points of the 2-LD system under the Neyman-Pearson criterion with α=0.2009 for the following detection strategies:

    Figure  15.  The operating points of the 2-LD system employing different detection strategies under the Neyman-Pearson criterion with α=0.2009.
    Table  IX.  The Operating Points of the 2-LD System Employing Different Detection Strategies Under the Neyman-Pearson Criterion With α=0.2009 (corresponding to Fig. 15)
    Detection strategyα=0.2009
    1Deterministic strategyG=(0.2009,0.8217)
    2Randomization at the DFCG=(0.2009,0.8217)
    3Dependent randomization (synchronized)C=(0.2009,0.8261)
    4Dependent randomization (the DFC is unsynchronized with the LDs group)C=(0.2009,0.7005)
    5Dependent randomization (the 2nd LD is unsynchronized with the DFC and other LDs)C=(0.2009,0.7547)
    6Dependent randomization (all LDs and the DFC are unsynchronized)C=(0.2009,0.7008)
     | Show Table
    DownLoad: CSV

    a) Deterministic strategy and randomization at the DFC (G, blue circle);

    b) Dependent randomization (C, black circle);

    c) Dependent randomization when the LDs group lost synchronization with the DFC before the redesigned algorithm is applied (W=(0.2640,0.7600), cyan circle) and after the redesigned algorithm is applied (C, cyan square);

    d) Dependent randomization when the LD2 lost synchronization with the LD1 and the DFC before the redesigned algorithm is applied (W=(0.2324,0.7930), purple circle) and after the redesigned algorithm is applied (C, purple square); and

    e) Dependent randomization when both two LDs lost synchronization with each other and the DFC before the redesigned algorithm is applied (W=(0.2044,0.6401), yellow circle) and after the redesigned algorithm is applied (C, yellow square).

    2) 3-LD System: Returning to the 3-LD system (Section III-B), we show in Fig. 16 what happen when Y={LD1,LD2} and ¯Y={LD3} and when the probability of false alarm constraint is Pfα=0.1708. The notation is the same as in Fig. 14. Table X summarizes the input and output of the redesigned algorithm for the 3-LD system employing dependent randomization when LD3 lost synchronization.

    Figure  16.  C is the desired operating point of the system with dependent randomization (black circle). When Y={LD1,LD2} and ¯Y={LD3}, A, B, M1, and M2 are the four possible operating points (green circles). W is the equivalent operating point (purple circle). ROC curve A and ROC curve B are shown as the red curve and the blue curve, respectively. C is the operating point maximizing the probability of detection given α=0.1708, shown by the purple square.
    Table  X.  The Output the 3-LD System Employing Dependent Randomization Before and After a Corrective Action is Taken When Y={LD1, LD2} and ¯Y={LD3}
    Input of the redesigned algorithm for the 3-LD system when LD3 lost synchronization
    The design input of dependent randomization (Table III)1. The number of local detectors, n=3
    2. The probability of false alarm constraint, α=0.1708
    The design output of dependent randomization (Table III)3. The local operating points of A: (Pfk,Pdk)=(0.2,0.7),k=1,2,3,
     The local operating points of B: (Pfk,Pdk)=(0.1,0.6),k=1,2,3
    4. The probability of selecting A, p=0.6
    Information of synchronized LDs5. Numbers of synchronized LDs, m=2
    6. Identity of all synchronized LDs, Y={LD1,LD2}
    Output of the redesigned algorithm for the 3-LD system when LD3 lost synchronization
    1. ΩA={(0,0),(0.0064,0.3234),(0.0576,0.6006),(0.0912,0.7672),(0.1936,0.8266),(0.4624,0.9694),(1,1)}, ΩB={(0,0),(0.0016,0.2376),(0.0100,0.3600),(0.0388,0.6768),(0.1900,0.8400),(0.3196,0.9456),(1,1)}
    2. Two operating points A=(0.0912,0.7672)ΩA and B=(0.2902,0.9216)ΩB, which allow PCf from (28) satisfying the probability of false alarm constraint α and achieving the highest probability of detection
    3. The deterministic fusion rule γA0(2 out of 3 rule) used to achieve A
    4. The randomized fusion rule γB0 used to achieve B, which requires the system operating at ωBa=(0.1900,0.8400)(achieved by the fusion rule such that u0=u1|u2) with probability q and ωBb=(0.3196,0.9456)(achieved by the 1 out of 3 rule) with probability 1q, where q=0.2748, calculated by (32)
    5. The operating point of the non-synchronized system after the corrective action is taken, C=(0.1708,0.8290), calculated by (31) and (33), which is achieved by operating at A with probability p and B with probability 1p
     | Show Table
    DownLoad: CSV

    Fig. 17 compares the operating points of the 3-LD system under the Neyman-Pearson criterion with α=0.1708 for the following detection strategies:

    Figure  17.  The operating points of the 3-LD system employing different detection strategies under the Neyman-Pearson criterion with α=0.1708.

    a) Deterministic strategy (G, blue circle);

    b) Randomization at the DFC (E, red circle);

    c) Dependent randomization (C, black circle);

    d) Dependent randomization when the LDs group lost synchronization with the DFC before the redesigned algorithm is applied (W=(0.2046,0.8210), cyan circle) and after the redesigned algorithm is applied (C, cyan square);

    e) Dependent randomization when the LD2 lost synchronization with the LD1 and the DFC before the redesigned algorithm is applied (W=(0.1826,0.8386), purple circle) and after the redesigned algorithm is applied (C, purple square); and

    f) Dependent randomization when both two LDs lost synchronization with each other and the DFC before the redesigned algorithm is applied (W=(0.2223,0.8380), yellow circle) and after the redesigned algorithm is applied (C, yellow square).

    Table XI shows the operating points of the 3-LD system employing different detection strategies under the Neyman-Pearson criterion with a) α=0.1708 and b) α=0.05. Lines 3 and 4 at column “α=0.05” are identical. The reason is that when α=0.05, dependent randomization requires the system hopping between two operating points with the same global fusion rule. In this circumstance, dependent randomization is “randomization at the LDs only” and the DFC does not participate in the randomization. Therefore, the loss of synchronization between the LDs group and the DFC has no influence to the system performance.

    Table  XI.  The Operating Points of the 3-LD System Employing Different Detection Strategies Under the Neyman-Pearson Criterion With (a) α=0.1708 (Corresponding to Fig. 17) and (b) α=0.05
    Detection strategyα=0.1708α=0.05
    1Deterministic strategyG=(0.1360,0.7960)(0.0460,0.6960)
    2Randomization at the DFCE=(0.1708,0.8208)(0.0500,0.7000)
    3Dependent randomization (synchronized)C=(0.1708,0.8448)(0.0500,0.7031)
    4Dependent randomization (the DFC is unsynchronized with the LDs group)C=(0.1708,0.7974)(0.0500,0.7031)
    5Dependent randomization (the 3rd LD is unsynchronized with the DFC and other LDs)C=(0.1708,0.8290)(0.0500,0.5704)
    6Dependent randomization (all LDs and the DFC are unsynchronized)C=(0.1708,0.8009)(0.0500,0.5534)
     | Show Table
    DownLoad: CSV

    An example of a larger system (with seven LDs) is available in https://github.com/moshekam/Dependent-Randomization.

    In parallel binary decentralized detection networks, dependent randomization can sometimes make the team’s Receiver Operating Characteristic curve concave (if it was non-concave under other detection schemes). This effect improves the system’s performance under a Neyman-Pearson criterion by realizing a higher probability of detection for the same upper bound on the probability of false alarm. Dependent randomization requires that the DFC and the LDs be synchronized, guided by a coordinated randomization scheme. The DFC and the LDs switch simultaneously together, back and forth, between γA0 (for the DFC) and γALD={γA1,,γAn} (for the LDs); and γB0 (for the DFC) and γBLD={γB1,,γBn} (for the LDs). However, if the synchronization is lost, the system may exceed the permitted probability of false alarm. We study the consequences of synchronization loss in the following two sets of circumstances: a) the DFC is not synchronized with the LDs group; and b) some LDs are not synchronized with other LDs and with the DFC. Corrective action is devised in order to restore the detection system to compliance with the probability of false alarm constraint, at a cost of reduced probability of detection. The implication of these results is that operation of systems in “dependent randomization” mode is actually feasible. It can improve performance of parallel binary decision fusion systems, and its failure mode (loss of synchronization) can be quantified, as well as mitigated. Future work includes applications to pertinent networks, such as those studied in [14]–[18].

  • [1]
    I. Y. Hoballah and P. K. Varshney, “Neyman-Pearson detection with distributed sensors,” in Proc. 25th IEEE Conf. Decision and Control, vol. 25, pp. 237–241, 1986.
    [2]
    S. Thomopoulos, R. Viswanathan, and D. Bougoulias, “Optimal distributed decision fusion,” IEEE Trans. Aerospace and Alectronic Aystems, vol. 25, no. 5, pp. 761–765, 1989.
    [3]
    M. Kam, W. Chang, and Q. Zhu, “Hardware complexity of binary distributed detection systems with isolated local Bayesian detectors,” IEEE Trans. Systems,Man and Cybernetics, vol. 21, no. 3, pp. 565–571, 1991. doi: 10.1109/21.97477
    [4]
    P. Willett and D. Warren, “The suboptimality of randomized tests in distributed and quantized detection systems,” IEEE Trans. Information Theory, vol. 38, no. 2, pp. 355–361, 1992. doi: 10.1109/18.119692
    [5]
    J. N. Tsitsiklis, “Decentralized detection,” Advances in Statistical Signal Processing, vol. 2, no. 2, pp. 297–344, 1993.
    [6]
    J. D. Papastavrou and M. Athans, “The team ROC curve in a binary hypothesis testing environment,” IEEE Trans. Aerospace and Electronic Systems, vol. 31, no. 1, pp. 96–105, 1995. doi: 10.1109/7.366296
    [7]
    Y. I. Han, “Randomized fusion rules can be optimal in distributed Neyman-Pearson detectors,” IEEE Trans. Information Theory, vol. 43, no. 4, pp. 1281–1288, 1997. doi: 10.1109/18.605596
    [8]
    S. Acharya, J. Wang, and M. Kam, “Distributed decision fusion using the Neyman-Pearson criterion,” in Proc. 17th IEEE Int. Conf. Information Fusion, pp. 1–7, 2014.
    [9]
    W. Dong and M. Kam, “Parallel decentralized detection with dependent randomization,” in Proc. 52nd IEEE Annu. Conf. Information Sciences and Systems, pp. 1–6, 2018.
    [10]
    P. K. Varshney, E. Masazade, P. Ray, and R. Niu, “Distributed detection in wireless sensor networks,” in Distributed Data Fusion for Network-Centric Operations, D. Hall, C.-Y. Chong, J. Llinas, and M. Liggins II, Eds., ch. 4, CRC Press, 2013.
    [11]
    A. T. Zijlstra, Calculating the 8th Dedekind Number. Ph.D. Thesis, University of Groningen, 2013.
    [12]
    H. L. Van Trees, Detection, Estimation, and Modulation Theory, Section 2.2.2. Wiley, 1968.
    [13]
    Q. Yan and R. S. Blum, “On some unresolved issues in finding optimum distributed detection schemes,” IEEE Trans. Signal Processing, vol. 48, no. 12, pp. 3280–3288, 2000.
    [14]
    J. Luo, J. Ni, and Z. Liu, “Distributed decision fusion under nonideal communication channels with adaptive topology,” Information Fusion, vol. 45, pp. 190–201, 2019. doi: 10.1016/j.inffus.2018.02.001
    [15]
    M. A. Al-Jarrah, A. Al-Dweik, M. Kalil, and S. S. Ikki, “Decision fusion in distributed cooperative wireless sensor networks,” IEEE Trans. Vehicular Technology, vol. 68, no. 1, pp. 797–811, 2018.
    [16]
    Y. He, S. Li, and Y. Zheng, “Distributed state estimation for leak detection in water supply networks,” IEEE/CAA J. Autom. Sinica, pp. 1–9, 2017. DOI: 10.1109/JAS.2017.7510367
    [17]
    M. Ghahramani, M. Zhou, and C. T. Hon, “Mobile phone data analysis: A spatial exploration toward hotspot detection,” IEEE Trans. Automation Science and Engineering, vol. 16, no. 1, pp. 351–362, 2018.
    [18]
    J. Cheng, M. Chen, M. Zhou, S. Gao, C. Liu, and C. Liu, “Overlapping community change-point detection in an evolving network,” IEEE Trans. Big Data, vol. 6, no. 1, pp. 189–200, 2020. doi: 10.1109/TBDATA.2018.2880780
  • Related Articles

    [1]Siwei Lou, Chunjie Yang, Zhe Liu, Shaoqi Wang, Hanwen Zhang, Ping Wu. Release Power of Mechanism and Data Fusion: A Hierarchical Strategy for Enhanced MIQ-Related Modeling and Fault Detection in BFIP[J]. IEEE/CAA Journal of Automatica Sinica. doi: 10.1109/JAS.2024.124821
    [2]Jinyuan Liu, Xingyuan Li, Zirui Wang, Zhiying Jiang, Wei Zhong, Wei Fan, Bin Xu. PromptFusion: Harmonized Semantic Prompt Learning for Infrared and Visible Image Fusion[J]. IEEE/CAA Journal of Automatica Sinica, 2025, 12(3): 502-515. doi: 10.1109/JAS.2024.124878
    [3]Jiaxing Li, Zidong Wang, Jun Hu, Hongli Dong, Hongjian Liu. Cubature Kalman Fusion Filtering Under Amplify-and-Forward Relays With Randomly Varying Channel Parameters[J]. IEEE/CAA Journal of Automatica Sinica, 2025, 12(2): 356-368. doi: 10.1109/JAS.2024.124590
    [4]Qingyuan Ji, Xiaoyue Wen, Junchen Jin, Yongdong Zhu, Yisheng Lv. Urban Traffic Control Meets Decision Recommendation System: A Survey and Perspective[J]. IEEE/CAA Journal of Automatica Sinica, 2024, 11(10): 2043-2058. doi: 10.1109/JAS.2024.124659
    [5]Xiongbo Wan, Chaoling Zhang, Fan Wei, Chuan-Ke Zhang, Min Wu. Hybrid Dynamic Variables-Dependent Event-Triggered Fuzzy Model Predictive Control[J]. IEEE/CAA Journal of Automatica Sinica, 2024, 11(3): 723-733. doi: 10.1109/JAS.2023.123957
    [6]Yanxia Tan, Zhenkun Huang. Synchronization of Drive-Response Networks With Delays on Time Scales[J]. IEEE/CAA Journal of Automatica Sinica, 2024, 11(4): 1063-1065. doi: 10.1109/JAS.2016.7510043
    [7]Xiang Li, Yixiao Xu, Naipeng Li, Bin Yang, Yaguo Lei. Remaining Useful Life Prediction With Partial Sensor Malfunctions Using Deep Adversarial Networks[J]. IEEE/CAA Journal of Automatica Sinica, 2023, 10(1): 121-134. doi: 10.1109/JAS.2022.105935
    [8]Tangyike Zhang, Junxiang Zhan, Jiamin Shi, Jingmin Xin, Nanning Zheng. Human-Like Decision-Making of Autonomous Vehicles in Dynamic Traffic Scenarios[J]. IEEE/CAA Journal of Automatica Sinica, 2023, 10(10): 1905-1917. doi: 10.1109/JAS.2023.123696
    [9]Yunfan Zhang, Feng Liu, Yifan Su, Yue Chen, Zhaojian Wang, João P. S. Catalão. Two-Stage Robust Optimization Under Decision Dependent Uncertainty[J]. IEEE/CAA Journal of Automatica Sinica, 2022, 9(7): 1295-1306. doi: 10.1109/JAS.2022.105512
    [10]Mykhailo Ivanov, Oleg Sergyienko, Vera Tyrsa, Lars Lindner, Wendy Flores-Fuentes, Julio Cesar Rodríguez-Quiñonez, Wilmar Hernandez, Paolo Mercorelli. Influence of Data Clouds Fusion From 3D Real-Time Vision System on Robotic Group Dead Reckoning in Unknown Terrain[J]. IEEE/CAA Journal of Automatica Sinica, 2020, 7(2): 368-385. doi: 10.1109/JAS.2020.1003027
    [11]Wangli He, Zekun Mo, Qing-Long Han, Feng Qian. Secure Impulsive Synchronization in Lipschitz-Type Multi-Agent Systems Subject to Deception Attacks[J]. IEEE/CAA Journal of Automatica Sinica, 2020, 7(5): 1326-1334. doi: 10.1109/JAS.2020.1003297
    [12]Tongle Zhou, Mou Chen, Jie Zou. Reinforcement Learning Based Data Fusion Method for Multi-Sensors[J]. IEEE/CAA Journal of Automatica Sinica, 2020, 7(6): 1489-1497. doi: 10.1109/JAS.2020.1003180
    [13]Fei-Yue Wang, Jie Zhang, Qinglai Wei, Xinhu Zheng, Li Li. PDP: Parallel Dynamic Programming[J]. IEEE/CAA Journal of Automatica Sinica, 2017, 4(1): 1-5.
    [14]Subir Das, Vijay K Yadav. Stability Analysis, Chaos Control of Fractional Order Vallis and El-Nino Systems and Their Synchronization[J]. IEEE/CAA Journal of Automatica Sinica, 2017, 4(1): 114-124.
    [15]Xu Quan, Zhuang Shengxian, Zeng Yingfeng, Xiao Jian. Decentralized Adaptive Strategies for Synchronization of Fractional-order Complex Networks[J]. IEEE/CAA Journal of Automatica Sinica, 2017, 4(3): 543-550. doi: 10.1109/JAS.2016.7510142
    [16]Zeeshan Alam, Liguo Yuan, Qigui Yang. Chaos and Combination Synchronization of a New Fractional-order System with Two Stable Node-foci[J]. IEEE/CAA Journal of Automatica Sinica, 2016, 3(2): 157-164.
    [17]Ling Wang, Shengyao Wang, Xiaolong Zheng. A Hybrid Estimation of Distribution Algorithm for Unrelated Parallel Machine Scheduling with Sequence-Dependent Setup Times[J]. IEEE/CAA Journal of Automatica Sinica, 2016, 3(3): 235-246.
    [18]Shiming Chen, Xiaoling Chen, Zhengkai Pei, Xingxing Zhang, Huajing Fang. Distributed Filtering Algorithm Based on Tunable Weights Under Untrustworthy Dynamics[J]. IEEE/CAA Journal of Automatica Sinica, 2016, 3(2): 225-232.
    [19]Zhenyuan Guo, Shaofu Yang, Jun Wang. Global Synchronization of Stochastically Disturbed Memristive Neurodynamics via Discontinuous Control Laws[J]. IEEE/CAA Journal of Automatica Sinica, 2016, 3(2): 121-131.
    [20]Long Ma, Haibo Min, Shicheng Wang, Yuan Liu, Shouyi Liao. An Overview of Research in Distributed Attitude Coordination Control[J]. IEEE/CAA Journal of Automatica Sinica, 2015, 2(2): 121-133.

Catalog

    通讯作者: 陈斌, bchen63@163.com
    • 1. 

      沈阳化工大学材料科学与工程学院 沈阳 110142

    1. 本站搜索
    2. 百度学术搜索
    3. 万方数据库搜索
    4. CNKI搜索

    Figures(17)  / Tables(11)

    Article Metrics

    Article views (2348) PDF downloads(57) Cited by()

    Highlights

    • Studying parallel decentralized binary detection under the Neyman-Pearson Criterion
    • Studying dependent randomization in a parallel decentralized binary detection system
    • Quantifying the impact of synchronization loss when dependent randomization is used
    • Offering partial recovery from synchronization loss under dependent randomization

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return