Method for determining detection thresholds for a constant false alarm rate device of a sensor, device and associated method

Abstract

Method for determining detection thresholds of a constant false alarm rate (TFAC) device for a sensor, associated devices and methods. The present invention relates to a method for determining detection thresholds of a TFAC device for a sensor observing an environment, the TFAC device being equipped with an estimator suitable for determining an estimate of the noise-related parameters affecting the sensor as a function of environmental parameters. The method comprises the steps of: - estimating noise-related parameters to obtain a parameter vector, - calculating first detection thresholds so that, for all possible configurations of the parameter vector, the TFAC device operates at a predefined false alarm probability, - characterizing the estimation errors of the estimator, and - correcting each first detection threshold to obtain second detection thresholds.The correction step takes into account the error parameters. Figure for the abbreviation: Figure 2,

Description

Title of the invention: Method for determining detection thresholds of a constant false alarm rate device for a sensor, associated devices and methods

[0001] The present invention relates to a method for determining detection thresholds for a constant false alarm rate device of a sensor. The present invention also relates to an observation method implementing the steps of the determination method. Finally, the invention relates to devices, namely a computer, a detection system, and a vehicle.

[0002] In the field of detection, detection systems usually use signal processing methods. These methods are tasked with extracting from a received signal one or more portions corresponding to useful information about the part of the environment being detected.

[0003] If we consider the case of a radar, the useful information corresponds to waves reflected by objects. These reflected waves or echoes can have a variable level depending on the nature and dimensions of the objects considered.

[0004] In cases where the signal reflected by the object has a low amplitude, the radar signal processing seeks to extract this reflected signal from the thermal noise or ambient clutter received by the radar receiver and accompanying this signal. The term "clutter" in this case refers to the signal reflected by any element not being detected. This could be, for example, the signal reflected by topographic features, buildings, bodies of water, vegetation, or atmospheric phenomena such as clouds.

[0005] Such an extraction operation is, for example, implemented by a constant false alarm rate device. The acronym TF AC is often used to refer to the term "constant false alarm rate," so the device is often called a TF AC device.

[0006] In general, a TFAC device operates by analyzing any signal received by the receiver through a series of operations. For example, the TFAC device subtracts an average ambient signal level from the incident signal. The difference is then compared to a detection threshold. Any received signal whose level exceeds the detection threshold is considered representative of the presence of an object at the distance considered, corresponding to the time the signal was received by the radar receiver.

[0007] Thus, the proper functioning of the TFAC device depends heavily on the accuracy of the detection threshold value. Therefore, in practice, it is not a threshold of detection but a plurality of detection thresholds are used, these thresholds depending on the environmental conditions in which the sensor operates.

[0008] However, TFAC devices exhibit degraded performance in certain circumstances, particularly when the environment includes areas of different nature generating significant variations in the level of clutter.

[0009] There is therefore a need for a method of determining detection thresholds of a constant false alarm rate device of a sensor that is more reliable.

[0010] To this end, the description describes a method for determining detection thresholds for a constant false alarm rate device of a sensor,

[0011] the sensor being suitable for observing an environment, the constant false alarm rate device being provided with an estimator suitable for determining an estimate of the parameters related to the noises affecting the sensor as a function of parameters related to the environment that the sensor observes,

[0012] the determination method being implemented by a computer and comprising:

[0013] - a step of estimating parameters related to the noises affecting the sensor, for to obtain a parameter vector,

[0014] - a step of calculating initial detection thresholds for the rate device constant false alarm, the first thresholds being calculated so that, for all possible configurations of the parameter vector, the constant false alarm rate device operates at a predefined false alarm probability,

[0015] - a step of characterizing the estimation errors of the device estimator of a constant false alarm rate, to obtain parameters relating to estimation errors, and

[0016] - a step of correcting the value of each first detection threshold for to obtain second detection thresholds, the correction step taking into account the parameters relating to estimation errors, the second detection thresholds being the thresholds determined for the constant false alarm rate device.

[0017] It is important to understand here that the steps of calculating the initial thresholds and characterizing the errors in estimating the parameter vector are performed without using real data. They can therefore be carried out offline, which is a significant advantage of the method. The gain in reliability is achieved without requiring prolonged use of the detection system to encounter a representative set of use cases.

[0018] According to particular embodiments, the method for determining detection thresholds has one or more of the following characteristics, taken individually or in all technically possible combinations:

[0019] - the correction step includes, for each possible configuration of the vector of parameters, the operations of:

[0020] - determination of the first detection threshold and the first detection thresholds that The constant false alarm rate device is likely to be used, taking into account the estimation errors of the estimator, and

[0021] - modification of the values ​​of the first detection thresholds that the rate device constant false alarm is likely to be used given estimation errors depending on the different values ​​of the first detection thresholds determined.

[0022] - the correction step further comprises a calculation operation of a better value among the first detection thresholds determined to ensure that the constant false alarm rate device operates at a predefined false alarm probability, to obtain a better obtained value, and

[0023] During the modification operation, each of the values ​​of the first detection thresholds that the constant false alarm rate device is likely to use, taking into account estimation errors, are replaced by the best value obtained.

[0024] - the best value is the maximum of the set of values ​​determined during the determination operation.

[0025] - the parameter vector includes a parameter related to the thermal noise of the sensor and a parameter related to the noise generated by the environment that the sensor observes.

[0026] - a parameter related to the noise generated by the environment that the sensor observes characterizes the sea's chaos.

[0027] The description also relates to a method of observing an environment by a sensor adapted to observing the environment, the sensor being part of a detection system comprising:

[0028] - a sensor suitable for observing an environment,

[0029] - a constant false alarm rate device for the sensor, the rate device false alarm constant being provided with an estimator suitable for determining an estimate of the noise parameters affecting the sensor as a function of environmental parameters observed by the sensor, and

[0030] - a calculator,

[0031] the observation method comprising:

[0032] - a step in implementing a method for determining detection thresholds of constant false alarm rate device,

[0033] - a stage of receiving waves from the environment, and

[0034] - a step of analyzing the waves received by the false alarm rate device constant to determine if the waves contain only noise, the analysis being based on the determined thresholds.

[0035] The description also relates to a calculator suitable for determining detection thresholds for a constant false alarm rate device of a sensor,

[0036] the sensor being suitable for observing an environment, the constant false alarm rate device being provided with an estimator suitable for determining an estimate of the parameters related to the noises affecting the sensor as a function of parameters related to the environment that the sensor observes,

[0037] the calculator being specific to:

[0038] - estimating parameters related to noise affecting the sensor, to obtain a vector parameters,

[0039] - calculate initial detection thresholds for the false alarm rate device constant, the first thresholds being calculated so that, for all possible configurations of the parameter vector, the constant false alarm rate device operates at a predefined false alarm probability,

[0040] - characterize estimation errors of the false rate device estimator constant alarm, to obtain parameters relating to estimation errors, and

[0041] - correct the value of the first detection thresholds to obtain second thresholds detection, the correction step taking into account the parameters relating to estimation errors, the second detection thresholds being the thresholds determined for the constant false alarm rate device.

[0042] The description also relates to a detection system comprising:

[0043] - a sensor being suitable for observing an environment,

[0044] - a constant false alarm rate device for the sensor, the rate device false alarm constant being provided with an estimator suitable for determining an estimate of the noise parameters affecting the sensor as a function of environmental parameters observed by the sensor, and

[0045] - a calculator as previously described.

[0046] The description also relates to a vehicle comprising a detection system as previously described.

[0047] In this description, the expression "specific to" means interchangeably "suitable for", "adapted to" or "configured for".

[0048] Some features and advantages of the invention will become apparent from the following description, given solely by way of non-limiting example, and made with reference to the accompanying drawings, in which:

[0049] - [Fig.1] [Fig.1] is a schematic representation of an example of a vehicle,

[0050] - [Fig.2] [Fig.2] is a schematic representation of a flowchart cor responding to an example of the implementation of a method for determining detection thresholds,

[0051] - [Fig.3] [Fig.3] is a schematic representation of the result of the implementation of a step in the process of determining the [Fig.2],

[0052] - [Fig.4] [Fig.4] is a schematic representation of the result of the implementation at an initial stage of another step in the process of determining the [Fig.2],

[0053] - [Fig.5] [Fig.5] is a schematic representation of the result of the implementation at a second stage of the same step as that of [Fig. 4], and

[0054] - [Fig.6] [Fig.6] is a schematic representation of the result of the implementation at a final stage of the same stage as that of [Fig.4].

[0055] A vehicle 10 is schematically illustrated in [Fig.1].

[0056] Vehicle 10 is any type of vehicle and in particular, can be a land, air or sea vehicle.

[0057] The vehicle 10 includes a detection system 12 suitable for detecting elements in an environment.

[0058] In this sense, the detection system 12 can be used as a monitoring system.

[0059] The detection system 12 includes a sensor 14, a constant false alarm rate device 16 and a calculator 18.

[0060] Sensor 14 is, for example, a radar.

[0061] Several types of radar can be envisaged, such as a tracking radar, a detection radar.

[0062] The radar can be used on the ground, at sea or on a mobile platform, including an aircraft.

[0063] Alternatively, sensor 14 is a sonar.

[0064] According to another variant, sensor 14 is an optronic sensor.

[0065] For example, sensor 14 is an infrared detector or a camera.

[0066] As previously stated, the TF AC device 16 is suitable for determining signals exceeding a detection threshold dependent on the environment in which the sensor 14 operates.

[0067] The computer 18 is an embedded device designed to control all the elements of the detection system 12.

[0068] According to the example described, the calculator 18 is suitable for implementing a method for determining detection thresholds of the TFAC device 16.

[0069] For simplicity, it is assumed here that the computer 18 implements all the steps of the determination process.

[0070] However, in other embodiments, certain steps can be carried out offline depending on the data available.

[0071] The operation of the calculator 18 is now described with reference to [Fig.2] which illustrates an example of the implementation of such a determination method.

[0072] The determination method seeks to obtain detection thresholds for the TFAC device 16 which will improve the performance of the device 16.

[0073] The determination process comprises, according to the example in [Fig. 2], four steps which are an estimation step E20, a calculation step E22, a characterization step E24 and a correction step E26.

[0074] During the estimation step E20, the computer 18 receives a set of observations of the environment by the sensor 14.

[0075] The computer 18 estimates a parameter vector C from these observations.

[0076] The parameters of the parameter vector C make it possible to characterize the phenomena causing parasitic signals at the level of the sensor 14.

[0077] The parameters of the parameter vector C are thus parameters characterizing the probability law of occurrence of these phenomena.

[0078] These parameters can generally be easily estimated, in particular using altitude, observation geometry or thermal noise temperature.

[0079] The parameter vector C is therefore a set of parameters related to the noises affecting the sensor 14.

[0080] As a particular example, in the case of a sensor operating in an environment containing both air and water, the parasitic phenomena or nuisances are of two types, a first phenomenon related to sea echoes and a second phenomenon related to the thermal noise of the detection system 12.

[0081] The first phenomenon exhibits a power often modeled by a law in , v)' with 'a average power and v the form factor of the law, describing the impulsiveness of the sea mess.

[0082] In this expression, the form factor v (impulsivity) is the most determining.

[0083] According to the example described, the behavior of the second phenomenon is given by a parameter R» which corresponds to the average power. The thermal noise is then usually modeled by a power law of the type Exp^-^- ) •

[0084] Alternatively, the second phenomenon can be characterized by an RCB parameter, the acronym RCB referring to the corresponding designation of 'Clutter to Noise ratio' (the term clutter is another designation for clutter). The RCB parameter is given by the power of the sea clutter relative to the thermal noise.

[0085] Thus, it can be envisaged in the described example that the parameter vector C comprises two parameters, namely the impulsivity v of the sea clutter and the RCB parameter of clutter-to-noise ratio.

[0086] Through this example, it appears that the parameter vector C preferably includes a parameter related to the thermal noise of the sensor 14 and a parameter related to the noise generated by the environment that the sensor 14 observes.

[0087] The computer 18 thus has the parameter vector C at the end of the implementation of the estimation step E20.

[0088] The parameter vector C is suitable for taking a set of values ​​cor- corresponding to the set of possible values ​​for each parameter. Each set of possible values ​​is a possible configuration of the parameter vector

[0089] The objective of the described method is to determine the thresholds that the TF AC device 16 will use to obtain the best possible performance for the detection system 12.

[0090] During the calculation step E22, the calculator 18 calculates initial detection thresholds taking into account the parameter vector and a predefined false alarm rate.

[0091] The predefined false alarm rate is called Pfa. In the following instructions description.

[0092] Such a calculation can be performed by any means known to a person skilled in the art.

[0093] As a particular example for the described case of a sensor operating in an environment containing both air and water, the computer 18 can implement the following operations.

[0094] The calculator 18 usually determines a testx statistic maximizing the probability of detection for a false alarm probability P fa = 1 - a. rrj instruction

[0095] For this purpose, the computer 18 can advantageously use Swerling models, the statistical distributions of the different noise sources (clutter and thermal noise for the example described here), and the Neyman-Pearson Lemma.

[0096] Once the expression of the optimal test statistic is known, the calculator 18 evaluates the statistical distribution fn(x, C) of the test statistic A under the assumption that it originates from the noise signal.

[0097] The calculator 18 obtains this distribution f analytically or by implementing a Monte Carlo method.

[0098] The calculator 18 then determines a first detection threshold by Inversion of the following integral:

[0099] Pf^cimsime- J f ■ (x,C)dx j set J j musanC(, \ / ^inf

[0100] With the lower bound of the support of A (if x E R+, Bmf = 0)-

[0101] For a simple law, the calculator 18 can perform such an inversion analytically.

[0102] But in the complex case described, the computer 18 performs the inversion numerically in a set of points (parameterizations of C) by dichotomy.

[0103] According to another example, calculator 18 uses a least squares method combined with gradient descent.

[0104] By proceeding in this way for all possible values ​​of the parameter vector C, a set of first detection thresholds is obtained for each value of the parameters of the parameter vector C.

[0105] In the example described, this therefore corresponds to a table giving the correct detection threshold for a value of the impulsivity v of the sea clutter and RCB parameter of clutter-to-noise ratio.

[0106] Such a table can therefore be represented graphically as a map as seen in the representation in [Fig.3].

[0107] In this figure, the threshold value is represented as gray levels as a function of the values ​​of the impulsivity v of the sea clutter (in the form of logi0(v) here) and the RCB parameter of clutter-to-noise ratio (expressed in dB). Each gray level gives the threshold value for a cell corresponding to a set of values ​​of the two parameters.

[0108] Thus, at the end of the calculation step E22, the computer 18 calculated first detection thresholds for the TF AC device 16 so that, for all possible configurations of the parameter vector, and in the absence of estimation errors on this parameter vector (i.e. the computer 18 knows perfectly the parameterization of the statistical laws of clutter and thermal noise nuisances), the TFAC device 16 operates at a predefined false alarm probability.

[0109] During the characterization step E24, the computer 18 seeks to characterize the performance of the estimator denoted Ç of the parameter vector C used during the step E20.

[0110] The estimator £ is an estimator with which device 16 of TFAC is provided.

[0111] The estimator £ is suitable for determining an estimate of the noise-related parameters affecting the sensor 14 as a function of environmental parameters that the sensor 14 observes.

[0112] As an example of parameters related to the environment, the altitude of the vehicle 10, the nature of the environment (e.g., land or water) or meteorological information (presence of clouds, rain or other).

[0113] This estimator £ therefore corresponds to a functionality of the TFAC device 16, which can take any form known to a person skilled in the art.

[0114] In this case, the aim here is to characterize the estimation errors made by the estimator

[0115] According to one example, performance is expressed in terms of the parameters of a distribution law.

[0116] Such a step can be implemented by an analytical solution when the distributions of the phenomena generating nuisances are known.

[0117] Alternatively, it is possible to use Monte Carlo methods or numerical integration methods.

[0118] According to another example, performance is expressed in terms of moments of a distribution law.

[0119] By noting Co the true parameter vector to be estimated, an implementation of this characterization step can therefore consist of theoretically quantifying the first two moments of the estimator £ for each parameterization Co (possible values ​​of the parameters) of the parameter vector C.

[0120] The first two moments of the estimator £ are defined as follows: [° 121 ] Bias è (C Q ) = E [ C - C o ] 101221 Var t .(C0) = e[ (&-C o ) 2

[0123] With £[.] the mathematical expectation operator.

[0124] The first moment (Biais ( Co) ) corresponds to the mean estimation error while the second moment (Varç:( Co) ) corresponds to the variance of the estimation error.

[0125] It would also be possible to consider higher order moments in the implementation of the characterization step E24.

[0126] To return to the above example concerning a sensor operating in an environment containing both air and water, it is observed that TFAC devices 16 generally have the defect of underestimating the value of the impulsivity r of the sea clutter.

[0127] The values ​​characterizing the performance make it possible to quantify such an estimation defect.

[0128] Following the characterization step E22, the calculator 18 therefore also has values ​​characterizing the estimation errors of the TFAC device 16.

[0129] As a point of reference, it should be understood that the initial detection thresholds only take into account the parameters of the parameter vector C and do not consider the values ​​characterizing the estimation performance of the TFAC device 16. The initial detection thresholds are therefore not perfectly reliable.

[0130] Thus, the calculation steps E22 and characterization steps E24 can be carried out jointly after the implementation of the estimation step E20.

[0131] Furthermore, as previously stated, these steps could also be carried out offline.

[0132] During the correction step E24, the computer 18 seeks to correct the values ​​of the first detection thresholds by taking into account the values ​​characterizing the estimation performance of the TFAC device 16.

[0133] To explain how the calculator 18 performs such a correction, an example is now described with reference to figures 4 to 6, again for the case of hybrid air / water detection.

[0134] The case in Figure 4 corresponds to the case of a possible configuration of the parameter vector C named corresponding to v

[0135] For this configuration Cb, a value has been determined ensuring compliance with the predefined false alarm rate P fa. This value is indicated by the reference sign 30 on [Fig.4].

[0136] For this configuration, it is assumed that the error that the estimator ç is likely to make on the estimation of the RCB parameter is less than the resolution of the graph (the size of a cell in Figure 7 is larger than the error made by the estimator £ on the RCB parameter), whereas for the component relating to impulsivity, the estimator is likely to make an error of up to 4 cells. These values ​​are derived from the calculation performed in the characterization step E24.

[0137] The set of first detection threshold values ​​that the TFAC device 16 is likely to use, taking into account estimation errors, corresponds in this case to a rectangle 32 shown in [Fig.4]. These values ​​are the values ​​to be modified.

[0138] According to the example described, the calculator 18 replaces the values ​​to be modified with the maximum of the first determined thresholds. In this case, this leads to an increase in the values ​​to be modified.

[0139] Figure 5 illustrates the same mechanism for a second configuration C2. Here again, this leads to raising the values ​​of the first thresholds contained in a rectangle 32.

[0140] This correspondence between this rectangular shape 32 and the determined estimation errors can be explained as follows.

[0141] The precise dimensioning of the zone requires knowing the prior probability of occurrence of each configuration of the parameter vector C, so as to dimension X = the zone actually seen by the device 16 of TFAC in the map which can be qualified as such (one-dimensional case):

[0142] kup / ^sup / Ysup \ JJ 1 f / Mc^dC, 'INF %INF T margin ( PftpÇj ) ] L-

[0143] With: * the distribution of test statistics Y given the configuration where i is an index designating the set of configurations • f A C1C ) the distribution of the estimates given by the estimator £ of the

[0144]

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159] parameter vector C for configuration • f ( C, ) the prior distribution of configurations C; of the parameter vector C, * [ h / NF' hup ] 'c suPPort of the parameter vector C (sus configurations likely to be encountered in a real-world environment) • YSUP the upper limit of the support for the Y test statistic Thus formulated, the problem to be solved (sizing the area X) also requires having the margin threshold table T margin, which is margined using the range X. As a first approximation, we can consider the a priori distribution of the C configurations; as being uniform, and the maintenance of the Pfa can be obtained approximately in a simplified way. For the example described above, for a configuration C, of ​​C, the characterization of the estimator £ gives: • A bias vector • A covariance matrix of the error Applying the principle of maximum entropy, the distribution of estimation errors for ç for configuration C corresponds to the following formula: With N being the normal distribution. The following criterion must therefore be met: With c < 1 so that the zone is "sufficiently large" with regard to the predefined false alarm rate Pf a. This corresponds to the fact that the probability that the estimator £ falls outside the zone is negligible with regard to the predefined false alarm rate Pf a. In the case of a one-dimensional parameter vector C, we therefore obtain: And in general, the criterion defines the area to be raised as an ellipsoid with the equation: Lq [ Cr - / Zq J 2 With Cr coordinates of the points to be raised. After discretization, this ellipsoid corresponds to the rectangle visible in figures 4 and 5. All the operations just described are repeated for each configuration.

[0160] This gives the map shown in [Fig.6] which brings together the values ​​of the first modified thresholds, which correspond to the second detection threshold values.

[0161] Compared with the map in [Fig.3], the second detection threshold values ​​are increased compared with the first modified detection thresholds.

[0162] These second detection threshold values ​​allow for better compliance with the predefined false alarm probability.

[0163] This improved compliance appears within the framework of an environmental observation process which then includes a stage of receiving waves from the environment, and a stage of analyzing the waves received by the TFAC device 16 to determine if the waves contain only noise, the analysis being based on the second detection thresholds.

[0164] The use of the second detection thresholds thus raised certainly degrades the probability of detection but makes it possible to better respect the desired probability of false alarm.

[0165] The method just described is therefore a method for determining detection thresholds for a device with a constant false alarm rate of a sensor that is more reliable.

[0166] Other embodiments of the process can be considered while still allowing such an increase in reliability to be achieved.

[0167] In particular, during the correction step, it is possible to use an operation other than the maximum, such as an operation consisting simply of taking the value determined by the calculation step.

[0168] It could also be considered to calculate an additional margin to be added, taking into account estimation errors.

Claims

Claims

1. Method for determining detection thresholds of a constant false alarm rate device (16) of a sensor (14), the sensor (14) being capable of observing an environment, the constant false alarm rate device (16) being provided with an estimator capable of determining an estimation of the parameters related to the noises affecting the sensor (14) as a function of parameters related to the environment that the sensor (14) observes, the determination method being implemented by a computer (18) and comprising: - a step of estimating parameters related to the noises affecting the sensor (14), to obtain a parameter vector, - a step of calculating first detection thresholds for the constant false alarm rate device (16), the first thresholds being calculated so that, for all possible configurations of the parameter vector, the constant false alarm rate device (16) operates at a predefined false alarm probability,- a step of characterizing the estimation errors of the estimator of the constant false alarm rate device (16), to obtain parameters relating to the estimation errors, and - a step of correcting the value of each first detection threshold to obtain second detection thresholds, the correction step taking into account the parameters relating to the estimation errors, the second detection thresholds being the thresholds determined for the constant false alarm rate device (16).,

2. Determination method according to claim 1, in which the correction step comprises, for each possible configuration of the parameter vector, the operations of: - determining the first detection threshold and the first detection thresholds that the constant false alarm rate device (16) is likely to use taking into account the estimation errors of the estimator, and - modifying the values ​​of the first detection thresholds that the constant false alarm rate device (16) is likely to use taking into account the estimation errors as a function of the different values ​​of the first detection thresholds determined.

3. The determination method of claim 2, wherein, step correction further comprises an operation of calculating a best value among the first detection thresholds determined to ensure that the constant false alarm rate device (16) operates at a predefined false alarm probability, to obtain a best value obtained, and during the modification operation, each of the values ​​of the first detection thresholds that the constant false alarm rate device (16) is likely to use taking into account the estimation errors are replaced by the best value obtained.

4. A determination method according to claim 3, wherein the best value is the maximum of the set of values ​​determined during the determination operation.

5. A determination method according to any one of claims 1 to 4, wherein the parameter vector comprises a parameter related to the thermal noise of the sensor (14) and a parameter related to the noise generated by the environment that the sensor (14) observes.

6. A determination method according to any preceding claim, wherein a parameter related to the noise generated by the environment that the sensor (14) observes characterizes the sea clutter.

7. Method for observing an environment by a sensor (14) capable of observing the environment, the sensor (14) being part of a detection system (12) comprising: - a sensor (14) capable of observing an environment, - a constant false alarm rate device (16) of the sensor (14), the constant false alarm rate device (16) being provided with an estimator capable of determining an estimation of the parameters linked to the noises affecting the sensor (14) as a function of parameters linked to the environment that the sensor (14) observes, and - a calculator (18), the observation method comprising: - a step of implementing a method for determining detection thresholds of the constant false alarm rate device (16), - a step of receiving waves coming from the environment, and - a step of analyzing the waves received by the constant false alarm rate device (16) to determine whether the waves comprise only noise,the analysis being based on the determined thresholds.,

8. Calculator (18) capable of determining detection thresholds of a constant false alarm rate device (16) of a sensor (14), the sensor (14) being suitable for observing an environment, the constant false alarm rate device (16) being provided with an estimator suitable for determining an estimate of the parameters linked to the noises affecting the sensor (14) as a function of parameters linked to the environment that the sensor (14) observes, the calculator (18) being suitable for: - estimate parameters linked to the noises affecting the sensor (14), to obtain a vector of parameters, - calculating first detection thresholds for the constant false alarm rate device (16), the first thresholds being calculated so that, for all possible configurations of the parameter vector, the constant false alarm rate device (16) operates at a predefined false alarm probability, - characterize estimation errors of the estimator of the device (16) with a constant false alarm rate, to obtain parameters relating to the estimation errors, and - correcting the value of the first detection thresholds to obtain second detection thresholds, the correction step taking into account the parameters relating to the estimation errors, the second detection thresholds being the thresholds determined for the constant false alarm rate device (16).

9. A detection system (12) comprising: - a sensor (14) suitable for observing an environment, - a constant false alarm rate device (16) of the sensor (14), the constant false alarm rate device (16) being provided with an estimator capable of determining an estimate of the parameters linked to the noises affecting the sensor (14) as a function of parameters linked to the environment that the sensor (14) observes, and - a calculator (18) according to claim 8.

10. Vehicle comprising a detection system according to claim 9.

Images