Rain sensor with fractal capacitor(s)

[0001] The present invention relates to a system for detecting the presence of rain on a vehicle window in which one or more sensor capacitors have fractal geometry.

BACKGROUND OF THE INVENTION AND SUMMARY OF EMBODIMENTS [0002] The presence of moisture (e.g. rain or condensation) on the front and / or rear windows of vehicles, if not removed immediately, may create dangerous driving conditions for drivers, passengers as well as pedestrians. Wiper blades are well known, which are a common way to remove this type of substance and reduce the risk of driving in hazardous conditions. Rain sensors have been developed to detect the presence of moisture (e.g. rain or other condensation) on the windshield of the vehicle and to turn on and off the wipers, if necessary, when moisture is detected. Automatic detection of rain, snow with rain, fog and the like and taking appropriate action - for example, turning on / off the wiper blades at the right speed - potentially reduces the amount of distraction for the driver, allowing him to better focus on the road ahead. However, incorrectly switching on / off the wipers or not activating the wipers in the presence of moisture may also pose a threat to driving safety. In addition, these types of systems are also susceptible to interfering pollutants, which, if present on the glass, can cause false readings / wiping.

[0003] Some traditional rain sensors are based on an electro-optical principle of operation. In some such techniques, raindrops are detected only by measuring the change in total internal reflection of the light beam at the glass - air border. Other electro-optical techniques try to analyze the brightness of the "image" of the glass to detect rain droplets or fog on the glass. However, these types of optical techniques have limited detection ranges, are quite expensive, and can cause erroneous readings as a result of using optical imaging as the only detection method.

53 / 51P26341PL00

EP 1 971 507 B1 [0004] US Patent No. 6,373,263 to Netzer describes the use of capacitive rain sensors and the reading of the differential current between two capacitors located on the windshield. Unfortunately, the Netzer system also has significant disadvantages. For example, the Netzer system may be susceptible to some harmful effects due to electromagnetic interference (EMI) as well as interference from other sources. For example, due to the fact that external objects (for example, a human hand, radio waves and other) interfere with the operation of capacitors, the excitation charges and receiver electrodes may change uncontrolled in the Netzer system, leading to false alarms or detection, and possibly cause false wiping and / or detection. The Netzer system is also prone to possible false readings caused by drastic temperature changes with respect to the reference capacitor used in the Netzer system, where the reference capacitor has a different geometry / shape / size than the sensing capacitor.

[0005] US 2003/0080871 describes a rain sensor with a meander-shaped condenser.

[0006] It should be mentioned that there is a need in the art for a rain sensor that will have effective operation and / or detection.

[0007] According to the present invention, capacitors are formed based on a fractal pattern. One or more capacitors are formed based on the Hilbert fractal pattern or Cantor set. These fractal structures maximize or enlarge the periphery and as a result increase the capacity for a given surface area. The use of two-dimensional fractal designs also means that the sensor takes up less physical space on the window, while being electrically larger than its physical size. The concentration of the transverse stream in fractal geometry may also allow the sensor to detect rain / water not necessarily dispersed over the entire physical surface of the sensor in some embodiments of the invention. In addition, in their higher iterations, the fractal capacitors / capacitors have the property of being their own Faraday cage or Faraday quasi-cage, which can reduce the adverse effects of electromagnetic interference or the like.

53 / 51P26341PL00

EP 1 971 507 B1

BRIEF DESCRIPTION OF THE DRAWINGS [0008] These and other properties and advantages will become better and more clearly understood with reference to the attached detailed description of the embodiments together with the drawings in which:

[0009] Fig. 1 (a) is a block diagram of the components of an exemplary rain sensor.

[0010] Fig. 1 (b) is a cross-sectional view of a rain sensor that can take advantage of the features of Fig. 1 (a) and / or one or more of Figs. 2-12.

[0011] Fig. 1 (c) is a cross-sectional view of a rain sensor that can use the properties of Fig. 1 (a) and / or one or more figures

2-12.

[0012] Fig. 1 (d) is a cross-sectional view of a rain sensor that can use the properties of Fig. 1 (a) and / or one or more figures

2-12.

[0013] Fig. 1 (e) is a cross-sectional view of a rain sensor that can use the properties of Fig. 1 (a) and / or one or more figures

2-12.

[0014] Fig. 1 (f) is a cross-sectional view of a rain sensor that can use the properties of Fig. 1 (a) and / or one or more figures

2-12.

[0015] Fig. 2A shows an example of an optimized pattern for a quarter of a capacitive matrix based on Hilbert fractals, such capacitors may be present on the window in the form of a sensor matrix, for example from one or more of Figures 1 (a) - 1 ( f) and 4 - 12.

[0016] Fig. 2B illustrates another example of an optimized pattern for a quarter of a capacitive matrix, such capacitors may be present on the glass as a sensor matrix, for example, from one or more of Figures 1 (a) 1 (f) and 4 - 12.

[0017] Figure 3 is an enlarged image of yet another quarter of the capacitive matrix, where such capacitors can be present on the glass in the form of a sensor matrix, for example, from one or more of figures 1 (a) - 1 (f) and 4 - 12.

53 / 51P26341PL00

[0018] Fig. 4 is an exemplary circuit diagram containing an exemplary circuitry used for a clock recording pulse in readout electronics, for use in, for example, one or more of Figures 1 (a) - 1 (f ) and 5 - 12.

[0019] Fig. 5 is an exemplary circuit diagram containing an exemplary circuitry used for a reset clock pulse in readout electronics, for use in, for example, one or more of Figures 1 (a) - 1 (f), 4 and 6 - 12.

[0020] Fig. 6 is an example time diagram obtained from the reading circuitry of Figs. 4-5.

[0021] Fig. 7 is an exemplary flowchart or state diagram showing how autocorrelation and cross-correlation data can be used to control wipers that can be used in conjunction with one or more of Figures 1-6 and 8-12.

[0022] Fig. 8 is an example flowchart showing how autocorrelation and cross-correlation data can be used to control wipers that can be used in conjunction with one or more of Figures 1-7 and 9-12.

[0023] Fig. 9 is an exemplary stylized view illustrating how a raindrop may move on the windshield.

[0024] Fig. 10 is a graph illustrating exemplary experimentally obtained maximum values ​​of non-normalized autocorrelation for various disorders.

[0025] Fig. 11A is an example of an experimentally obtained autocorrelation snapshot indicating intense rain.

[0026] Fig. 11B is an example of an experimentally obtained autocorrelation snapshot indicative of a light fog.

[0027] Fig. 11C is an example of an experimentally obtained autocorrelation snapshot indicating interference with CB radio.

[0028] Fig. 11D is an example of an experimentally obtained autocorrelation snapshot indicating a live grounded element.

[0029] Fig. 12A is an example correlation matrix indicating light rain.

53 / 51P26341PL00

[0030] Fig. 12B is an example correlation matrix indicating heavy rain.

[0031] Fig. 13 is an example of autocorrelation according to an embodiment of the invention.

[0032] Fig. 14 is a graph showing exemplary cross-correlation data derived from capacitors C1, C2.

[0033] Fig. 15 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using certain signals of Fig.

14.

[0034] Fig. 16 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using some of the signals in Fig.

14.

[0035] Fig. 17 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using certain signals of Fig.

14.

[0036] Fig. 18 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using certain signals of Fig.

14.

[0037] Fig. 19 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using some of the signals of Fig.

14.

[0038] Fig. 20 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using some of the signals in Fig. 14.

[0039] Fig. 21 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using some of the signals in Fig. 14.

53 / 51P26341PL00

[0040] Fig. 22 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using certain signals of Fig. 14.

[0041] Fig. 23 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using some of the signals of Fig. 14.

[0042] Fig. 24 is a cross-correlation graph illustrating cross-correlation values ​​as a function of time delays (time delays are expressed in microseconds in the time domain), using some of the signals in Fig. 14.

[0043] Fig. 25 is a block diagram illustrating an electrical circuit assembly and / or signal processing where a sensing capacitor (for example C1) is present, including sigma-delta modulation.

[0044] Fig. 26 is a block diagram illustrating an electrical circuit assembly and / or signal processing where multiple capacitors (for example C1-C4) are present, including sigma-delta modulation.

[0045] Fig. 27 is a block diagram illustrating sigma-delta modulation, the processing being performed in an electrical circuit assembly, firmware and / or computer program.

[0046] Figures 28 (a) and 28 (b) are block diagrams illustrating the advantages of using floating electrodes for sensing capacitors (for example C1-C4).

DETAILED DESCRIPTION [0047] Referring now in greater detail to the attached drawings, in which the same reference signs indicate the same elements in several views.

[0048] The rain sensor system includes a capacitive detection system that translates a physical input signal (e.g., the presence of a drop of water on the windshield or the like) into a digital electric voltage signal that is received and interpreted by the software (s) or circuit (- y) which decides

53 / 51P26341PL00

EP 1 971 507 B1 if the windshield wipers should be activated and if so, at what speed. Thus, capacitive coupling is used to detect water and / or other substance on the outer surface of the windshield, e.g. a car windshield, sunroof and / or rear windshield. It should be noted that computational methods can be performed by computer hardware or a combination of computer hardware and software in various embodiments of the present invention. No capacitance or reference capacitor is required (i.e. no compensation capacitor is needed).

[0049] The system may use an electric permeability equation that gives a physical quantity that describes how an electric field affects a given medium and that the medium affects the electric field. An example of a basic equation of electrical permeability is:

D = εοΕ + P, where D is an electric flux, ε0 is a vacuum dielectric constant, E is an electric field (for example, the voltage between plates or electrodes divided by distance or V / m), and P is polarization. P polarization can also be described mathematically as:

P = εεοΕ, where ε, is the relative electric permeability (for example, the dielectric constant of water, ice, dirt or anything else that may be on the outer surface of the window, e.g. windshield). In general, a high ε value will correspond to a high polarity. The permeability of glass is approximately 8, and the permeability of water about 85. By substitution and factorization, the equation of electric permittivity can be rewritten as:

D = εο (ε, +1) E,

In this form, it can be seen that D is the response to E stimulation.

53 / 51P26341PL00

[0050] The capacity C is given as C = Q / V, where Q is the load and V is the potential expressed in volts. In addition, C = Φ / ν, where Φ is the electric flux associated with the Q charge. From Gauss's law:

where dA is the surface area of ​​the differential square on the closed surface S. By substitution it becomes clear that the capacity is related to the potential difference:

C = \ DdA / V.

[0051] These equations form the basis of an exemplary technique for measuring the interaction of water on glass using a sensor with a capacitive matrix for sampling above a window (e.g., glass). In particular, data from a sensor containing at least one or two or more capacitors (e.g. C1, C2, C3 etc.) can be used to determine whether moisture is present on the outer surface of the window, e.g. the windshield or rear window of the vehicle (e.g. rain or the like). The above equations show that the presence of water on the window surface can affect the capacity of a properly located sensor capacitor.

[0052] Fig. 1 (a) is a block diagram of exemplary moisture sensor (e.g., rain) components. Power supply 10 is connected to the electronic reading system 12, which may include one or more hardware, firmware and / or software. As will be described in more detail below, the sensor includes one or more capacitors forming a capacitive sensor in some embodiments. Each of the capacitors has a pair of approximately coplanar electrodes arranged in a fractal pattern and used in the sensor of the present invention. The fractal pattern can be divided into a capacitive matrix. Due to the fact that the window can be flat or curved, the capacitor electrodes of the respective sensor capacitor (C1, C2, C3 and / or C4) are substantially coplanar with each other and are on a flat or curved window, even if there may be a slight curvature of the glass .

53 / 51P26341PL00

EP 1 971 507 B1

Data originating from and / or related to the capacitor / sensor capacitors of the capacitive sensor 5 is received and read by the electronic reading system 12, which may consist of one or more hardware, firmware and / or software. The electronic reading circuit 12 collects electrical noise and converts it into digital signal / signals. This digital signal (s) is fed into the calculation module 14 (which may consist of one or more hardware, firmware and / or software) which determines the action that the wipers should take. For example, wipers can perform a single wipe, low-speed wiping, high-speed wiping and the like based on the data being analyzed from and / or associated with a capacitive sensor. The wipers can also be turned off, slow down / accelerate wiping movements and the like, based on the analyzed data originating and / or related to the capacitive sensor. The engine 16 of the wiper control system receives orders from the calculation module 14 and instructs the wipers 18 to take the appropriate action.

[0053] The capacitive sensor 5 connects to the vehicle's LIN (Local Interconnect Bus). The LIN bus (not shown) is usually a serial bus to which slaves are connected to the vehicle. The LIN bus usually performs the reconciliation procedure (s) with the slave devices to ensure that they are connected and functioning, for example. In addition, the LIN bus can provide other information to slaves, such as the current time.

[0054] The capacitive sensor 5 comprises a plurality of capacitors in the form of a suitable matrix.

[0055] Fig. 1 (b) is a cross-sectional view of a vehicle window comprising a moisture sensor. The vehicle windshield contains an inner glass substrate 1 and an outer glass substrate 2 that are laminated together via a polymer intermediate layer 3 of a material such as poly (vinyl butyral) (PVB) or others. An optional low-emission coating 4 may be present on the inner surface of the outer glass substrate 2 (or even on the surface of the substrate 1). The low-emission coating 4 usually contains at least one thin layer reflecting infrared radiation from a material such as silver, gold or the like, sandwiched between at least

53 / 51P26341PL00

EP 1 971 507 B1 to the first and second dielectric layers of such material as silicon nitride, tin oxide, zinc oxide or the like. Examples of low-emission coatings 4, for example and without limitation, are described in US Patent Nos. 6,686,050, 6,723,211, 6,782,718, 6,749,941, 6,730,352, 6,802,943, 4,782,216, 3,682,528, 6,936,347.

[0056] Fig. 1 (b) shows an example of a capacitive sensor capacitor. Although the capacitive sensor of Fig. 1 (a) typically contains a plurality of capacitors in the matrix, Fig. 1 (b) shows, for simplicity, only one capacitor. The remaining capacitors are similar in cross-section to the capacitor shown in Fig. 1 (b) in certain example embodiments of this invention. An example of the capacitor (C1, C2, C3, C4) of the capacitive sensor in Fig. 1 (b) comprises a pair of spaced apart coplanar or substantially coplanar electrodes 7 and 8. The electrodes 7 and 8 are made of a conductive material that can be printed or otherwise formed on the glass. For example, the capacitor electrodes 7 and 8 of the sensing capacitor can be made of or contain silver, ITO (indium zinc oxide), or other suitable conductive material. The capacitor shown in fig. 1 (b) is subjected to the action of a rain droplet on the outer surface of the glass, because the electric field ES of the condenser extends to or outside the outer surface of the glass as shown in Fig. 1 (b) and can react with a droplet of rain or other material on the outer surface of the glass. The signals received from and / or associated with the sensing capacitor (sensing capacitors) and their analysis are described later in this document.

[0057] Fig. 1 (b) shows an opaque insulating layer 9 (e.g. black frit or enamel or other) present on the windshield above electrodes 7 and 8 to cover the electrodes 7,8 from the view of passengers (passengers) sitting inside the vehicle . It should be mentioned that the opaque layer 9 is only present on a small part of the glass, including the area where the capacitive matrix of the rain sensor capacitors is located. The capacitive matrix of the rain sensor and thus also the opaque layer 9 can be located on the windshield of the vehicle in the area near the rearview mirror mounting bracket. The opaque layer 9 (e.g. from black frit or enamel) can directly contact the fractal pattern of the capacitor electrodes 7, 8, because the layer 9 is not conductive. However, even if black frit layer 9 were conductive (which is possible), its dielectric constant is

53 / 51P26341PL00

EP 1 971 507 B1 close to the value for water, so that it will not adversely interfere with the collection of data originating from and / or related to the capacitors C1 - C4, as well as the related analysis:

[0058] Fig. 2a is a top view illustrating an exemplary capacitive sensor matrix comprising four capacitors C1, C2, C3, C4. Each of these capacitors C1, C2, C3, C4 includes coplanar spaced apart first and second capacitor electrodes, 7 and 8, as shown in Fig. 1 (b) (or any of Figs. 1 (c) - 1 (f) ). The capacitor electrodes 7 and 8 of each of the capacitors C1 - C4 may be made of conductive silver frit or similar material as shown in Fig. 2A. In addition, a gap 22 of from about 0.2 to 1.5 mm, more preferably from about 0.3 to about 1.0 mm (e.g. 0.6 mm) may be present between the coplanar capacitor electrodes 7 and 8 of the capacitor (C1, C2, C3 and / or C4) as shown in Fig. 2A. In fig. 2A, capacitors C1 - C4 are covered with an insulating layer 9 of black frit, which is the same as the opaque layer 9 discussed above with reference to Fig. 1 (b). In Fig. 2A, a contact field matrix is ​​present in the center of the sensor matrix that includes four contact fields electrically connected to the respective electrodes of 7 capacitors C1 - C4, as well as four contact fields electrically connected to the respective electrodes of 8 capacitors C1 - C4. An example contact field is indicated by reference number 28 in Fig. 2A. The four white contact fields 28 in Fig. 2A are electrically connected to the respective capacitor electrodes 7 of the C1 - C4 capacitors, and the dark gray contact fields 28 in Fig. 2A are electrically connected to the respective capacitor electrodes 8 of the C1 - C4 capacitors. All C1 - C4 sensing capacitors are sensitive to moisture, e.g. rain, on the outer surface of the glass.

[0059] In the embodiment of the present invention of Fig. 2A, each of the capacitors C1-C4 of the capacitive sensor is formed using fractal geometry. In particular, each of the coplanar electrodes 7 and 8 of each of the capacitors C1 - C4 is shaped in fractal geometry. Fractal patterns allow the realization of high capacity on a small surface area and are therefore desirable in relation to other geometries in rain sensor applications.

[0060] In the embodiment of Fig. 2A, it can be seen that the coplanar electrodes 7 and 8 of each capacitor (where the electrodes 7 and 8 are shown, but

53 / 51P26341PL00

EP 1 971 507 B1 not marked in Fig. 2A due to the black color of the frits 9, but are separated by slits 22) have fractal geometries and are arranged substantially parallel to each other along the meander length of each capacitor. In other words, each electrode 7, 8 of a given capacitor (e.g., C1, C2, C3, C4) has a meandering shape in fractal geometry, but remains substantially parallel to the second electrode (the second of 7, 8) of the capacitor along the length of the capacitor's meander. The total length of each capacitor (e.g. C1) over the length of the fractal meander is from about 25 to 200 mm in some embodiments of the present invention, more preferably from about 30 to 90 mm, for example about 50 mm.

[0061] The fractal pattern of Fig. 2A is a Hilbert fractal pattern. The electrodes 7, 8 of the capacitors C1 - C4 in the embodiment of Fig. 2A form a Hilbert fractal pattern. In particular, the capacitors shown in Fig. 2A are shaped in the form of third order Hilbert fractals. Hilbert fractals are continuous, space-filling fractals with a fractal dimension of two. This means that higher order fractals will be more square. The Hilbert fractal can be created by using the following L system:

Hilbert {

Angle 90 Axiom XX = -YF + XFX + FYY = + XF-YFY-FX +}

where "Angle 90" sets the following revolutions to 90 degrees, X and Y are defined functions, "F" means "draw forwards" "" + "means" rotate counterclockwise "and" - "means" rotate clockwise". In certain example embodiments of this invention, as shown in Figs. 2A, 2B and 3, all sensing capacitors of the sensing matrix may have identical or substantially identical shapes.

[0062] Each of the capacitors C1 - C4 in the sensor matrix can be electrically floating (this can be called in some cases a virtual mass), so as not to have a fixed common mass, for example a constant zero volts and / or spatially separated or similar, which may be useful in relation to correlation functions. In addition, the lack of a common mass means that the capacitive matrix

53 / 51P26341PL00

EP 1 971 507 B1 will not be adversely affected by interference, for example electromagnetic interference, thereby reducing the potential for false wipes, false detection and the like.

[0063] The fractal structure of capacitors C1 - C4 can be used on any of Figs. 1 (a) - 1 (f).

[0064] Fig. 1 (c) is a cross-sectional view of another example that may use the system of Fig. 1 (a) and one or more of Fig. 212. In Fig. 1 (c) a vehicle window (e.g. rear window) is made of only one glass sheet, and the capacitor electrodes 7, 8 are formed directly or indirectly on the inner main surface of the glass sheet

10. The capacitor (e.g. capacitor C1) shown in Fig. 1 (c) is constructed so that it is affected by a droplet of rain (or other substance) on the outer surface of the glass, because the electric field ES of the capacitor extends into or out of the outer surface of the glass, such as shown in Fig. 1 (c) and can therefore interact with a droplet of rain or other substance present on the outer surface of the glass. Each of the C1 - C4 capacitors is formed in a similar manner. It should be noted that the use of the word "na" here means positioning both directly and indirectly on, and is not limited to mere physical contact or touching, unless explicitly stated. An opaque layer 9, similar to the layer shown in the embodiment of Fig. 1 (b) may also be present in the embodiment of Fig. 1 (c) if desired.

[0065] Fig. 1 (d) is a cross-sectional view of another example that can use the system of Fig. 1 (a) and one or more of Fig. 2 12. In Fig. 1 (d) the vehicle window ( e.g. laminated windshield) comprises glass sheets 1 and 2 laminated together via an intermediate polymer layer 3, and optionally includes a low-emission coating 4 either on substrate 1 or substrate 2. Figure 1 (d) differs from Fig. 1 (b) in that the capacitor electrodes 7, 8 are present on the main surface of the glass substrate 1 which is furthest from the interior of the vehicle. In this embodiment, the capacitor electrodes 7, 8, in some cases, may contact the polymer intermediate layer 3. The capacitor (e.g. C1, C2, C3 or C4) shown in Fig. 1 (d) is designed in such a way that a droplet of rain (or other substance) located on the outer surface of the glass affects it because the electric field ES

53 / 51P26341PL00

The capacitor extends into or out of the outer surface of the glass as shown in Fig. 1 (d) and can therefore interact with a droplet of rain or other substance on the outer surface of the glass. Each of the sensor matrix C1 - C4 capacitors is formed in a manner similar to that shown for the capacitor of Fig. 1 (d). In fig. 1 (d) an opaque layer 9 may also be present, if desired, on a portion of the window so as to cover the capacitor electrodes from the eyes of passengers. In Fig. 1 (d), the electrodes 7, 8 can be formed of a conductive silver frit or ITO printed or made directly onto and in contact with the surface of the substrate 1. However, the present invention is not limited to this only, and the electrodes 7 and 8 of one or more sensor capacitors may instead be formed and made of a metallic conductive infrared reflecting layer (e.g. silver based layer) of the low-emission coating 4 which is on glass.

[0066] Fig. 1 (e) is a cross-sectional view of another embodiment that can use the system of Figs. 1 (a) and one or more of Figs. 2-12. In Fig. 1 (e) the glass vehicle (e.g., laminated windshield), comprises glass sheets 1 and 2 laminated together via an intermediate polymer layer 3, and optionally includes a low-emission coating 4 either on substrate 1 or substrate 2. Fig. 1 (e) differs from Fig. 1 (b) in that the capacitor electrodes 7, 8 (e.g. C1, C2, C3, C4) are located on the main surface of the outer glass substrate 2 that is closest to the interior of the vehicle. The capacitor electrodes 7, 8 can contact the polymer intermediate layer 3. The capacitor (e.g. C1, C2, C3, C4) shown in Fig. 1 (e) is designed in such a way that a droplet of rain (or other substance) present on the outer surface of the glass acts on it because the electric field ES of the capacitor penetrates into or out of the outer surface of the glass as shown in Fig. 1 (e) in connection with which may interact with a droplet of rain or other substance on the outer surface of the glass. Each of the sensor matrix capacitors C1 - C4 is formed in a manner similar to that shown for the capacitor in Fig. 1 (e). An opaque layer 9 may also be present in Fig. 1 (e), if desired, on a portion of the glass so as to shield the electrodes from the passenger (s) of the vehicle.

53 / 51P26341PL00

[0067] Fig. 1 (f) is a cross-sectional view of another embodiment that can use the system of Fig. 1 (a) and one or more of Figs. 2-12. 1 (f) the vehicle window (e.g., laminated windshield) comprises glass sheets 1 and 2 laminated together via the polymer intermediate layer 3 and optionally includes a low-emission coating 4 on substrate 1 or substrate 2. Fig. 1 (f) is different from Fig. . 1 (b) in that the electrodes 7, 8 of the capacitor (e.g. C1, C2, C3, C4) are located on the main surface of the inner glass substrate 1, which is located closest to the interior of the vehicle, via a support element 12. The support element 12 is arranged between the glass substrate 1 and the electrodes 7, 8 can be made of glass, silicone or similar material. The capacitor (e.g. C1, C2, C3, C4) shown in Fig. 1 (e) is designed in such a way that it is affected by a droplet of rain (or other substance) on the outer surface of the glass, because the electric field ES of the capacitor extends to or outside the outer surface of the glass, as shown in Figure 1 (f) in connection with what can interact with this droplet of rain or other substance on the outer surface of the glass. Each of the sensor matrix C1 - C4 capacitors is formed similarly to the one shown for the capacitor of Fig. 1 (f). An opaque layer 9 may also be present in Fig. 1 (f), if desired, on part of the glass so as to cover the electrodes 7, 8 from the eyes of the vehicle passengers.

[0068] Fig. 2B is a top view of an exemplary quadrant sensor array pattern of the fractal shape of capacitors C1-C4 for a capacitive sensor according to the present invention. The four capacitors shown in Fig. 2B are similar to the capacitors of Fig. 2A, except for their precise shapes. The capacitors in Fig. 2B can be used in any of Figs. 1 (a) - (f). The superimposed dashed lines show the division into four clear C1 - C4 capacitors. The width of the outer line may be about 2 mm, and the inner line about 1 mm.

[0069] Fig. 3 is an enlarged image of another example quadrant of the sensor matrix of the fractal shape of capacitors C1-C4 for a capacitive sensor according to the present invention. The four capacitors shown in Fig. 3 are similar to the capacitors of Figs. 2A and 2B except for their precise shapes. The fractal capacitors of Fig. 3 can be used on any of Figs. 1 (a) - (f). The superimposed lines show an example of division between

53 / 51P26341PL00

EP 1 971 507 B1 with capacitors C1 - C4 in Fig. 3. It should be noted that some embodiments may include sensor arrays with only two capacitors. However, it is preferred that they contain at least four capacitors to capture and determine the nuances of the disorder.

[0070] The use of fractal geometry for C1 - C4 sensing capacitors is beneficial in reducing false readings due to electromagnetic interference. In particular, high-iteration fractals help reduce the effect of this electromagnetic interference, because a high-iteration cage or quasi-Faraday fractal reduces the coupling of electromagnetic interference, thus reducing the adverse effects of electromagnetic interference. Fractals with a high number of iterations form quad Faraday cages. [0071] The readout electronics monitor the effects of rain and / or other disturbances on the glass. This process can be achieved by sequential charging of capacitors, reading their condition, quantizing data and / or deleting charges.

[0072] Fig. 4 is a schematic of a sensor or readout circuit. The sensor circuit of Fig. 4 can be made of electronic unit 12 and capacitive sensor matrix 5 of Fig. 1. As capacitors C1 C4 of the system of Fig. 4, any of the capacitors of Fig. 1 (b) - 1 (f ), 2A, 2B and / or 3. The electric circuit assembly of Fig. 4 is used for the recording clock pulse in the reading electronics. Transistors Q1, Q2 and Q7 are p-channel MOSFETs, while transistors Q1 and Q2 are mainly responsible for the write phase. Transistors Q5 and Q6 are n-channel MOSFETs.

[0073] Referring to Figure 4, during the write phase of the transistor Q7, a write pulse ClkWr is applied, which transistor works as a resistor or switch, charging one or more capacitors C1 - C4 with sensor capacity Cs. Fig. 6 shows some of the signals used in the arrangement of Fig. 4 in the recording cycle. In the write cycle, the Q1 transistor works in saturation mode, because its gate and drain are shorted, thanks to which the Q1 transistor is turned on. In write mode, transistors Q4, Q5 and Q6 are off, while transistor Q2 is on. Transistors Q3 and Q4 are optional. When the transistor Q7 is turned on by the write pulse, we get a write cycle and Vcc appears on the capacitance Cs via the A line and charges one or more sensor capacitors C1 - C4

53 / 51P26341PL00

EP 1 971 507 B1 capacity Cs. The Vcc voltage can be a DC voltage, for example, 5 V. At the same time, one or more C1 - C4 capacitors can be charged during the write cycle. However, this circuit charges and reads capacitors C1, C2, C3, C4 individually (see, for example, Fig. 6). Due to the above, only one capacitor among the C1, C2, C3, C4 capacitors is charged in one recording cycle.

[0074] The above process described for the left side of the sensor system of Fig. 4 is substantially reflected on the opposite or right side of the system of Fig. 4. As the current flows through the left branch, the current also flows through line B through the right branch and changes in capacity Cs are imitated or essentially imitated in the internal imitating capacity of Cint. When the transistor Q7 is turned on, the current also flows through the transistor Q2 (which is turned on) and charges the capacity Cint with the voltage Vcc. The charging of one of the C1 - C4 capacitors is therefore imitated by charging the Cint capacitor. In other words, the Cint capacitor is charged to the same degree or substantially the same as the capacitor (e.g. C1) charged on the other side of the system of Fig. 4. The output voltage of the system of Fig. 4, Vout (or Vo) depends on the capacity of Cint and is measured on or near the electrode of the Cint capacitor, as shown in Figure 4. An exemplary expression for Vout or Vo voltage is shown at the bottom of Figure 4. It should be noted that the output voltage Vout (or Vo) from the circuit of Figures 4-5 is associated and based on the capacitors C1 - C4 of the sensor Cs. In particular, the output Vout voltage of the system of Figs. 4-5 is related to and indicates the condition of the C1-C4 capacitors and the effect of the moisture present on the outer surface of the glass on these capacitors, although the voltage Vout is not measured directly on the C1-C4 capacitors. In particular, the voltage Vout (or Vo) is read during the write cycle as a result of the write pulse shown in Fig. 4 (see also Fig. 6). In the expression at the bottom of Figure 4 for the voltage Vout, W1 refers to Q1, W2 to Q2, L1 to Q1, L2 to Q2, where W is the transistor channel width and L is the transistor channel length. VT is the threshold voltage for each of the MOSFETs. It should be noted that alternatively the output voltage Vout of the system can be measured directly (instead of indirectly via Cint capacity) from the sensor capacitors C1 - C4.

53 / 51P26341PL00

EP 1 971 507 B1 [0075] Transistors Q3 and Q4 are optional. These transistors may be at low voltages (e.g., off) during the write phase and may be on during the erase phase.

[0076] The output signal Vout (or Vo) of the sensor circuit of Fig. 4 (and Fig. 5) is subjected to sigma-delta modulation. Sigma-delta modulators, which can be used in a sigma-delta-type digital-to-analog converter (DAC), can provide some degree of shaping or filtering of quantization noise that may be present. Examples of sigma-delta modulators that can be used are described in US Patent Nos. 6,975,257, 6,972,704, 6,967,608 and 6,980,144. Over-sampling, noise shaping and / or decimation filtering can be performed in sigma-delta conversion. Examples of advantages of sigma-delta modulation include one or more of the following: (i) the requirements for an analogue anti-aliasing filter are complex so that they can be cheaper than some types of nyquist based systems; (ii) a sampling and holding circuitry can be used due to the high input sampling frequency and low precision A / C conversion; (iii) due to the fact that the degree (s) of digital filtration may be behind the A / C conversion, the noise introduced during the conversion process, e.g. power ripple, reference voltage noise and the noise of the A / C converter itself, can be controlled ;

(iv) since the sigma-delta transducer may be substantially linear, it may not be subject to significant non-linearity, and / or the level (levels) of background noise may be independent of the level of the input signal. Improved signal / noise (S / N) ratios can be realized.

[0077] Fig. 25 is a simplified version of the sigmadelta modulator system for modulating and / or converting the output signal Vout (or Vo) of the circuit of Fig. 4 (and Fig. 5). In Fig. 25, the write pulse (see pulse at the bottom of Fig. 25) is used to charge the sensor capacitor (C1, C2, C3, C4) as discussed above with reference to Fig. 5. Rectangular excitation is used (e.g. for write and / or delete cycles) on the sensing capacitor to charge and discharge it. This process is reflected or emulated for Cint capacity as described. The output signal Vout (or Vo) of the system of Fig. 4 is subjected to sigma-delta modulation by a 60 sigma-delta modulator. The modulator 60 may be in the form of an electrical circuit assembly, firmware and / or software. pulses

53 / 51P26341PL00

Clock 62 from the clock are fed into the modulator 60 and trigger the latch of the modulator 60 quantizer. After sigma-delta modulation of the output signal Vout (or Vo) in the modulator 60, the modulated signals 64 are fed to the optional digital filter 66 (on example of a low-pass filter or similar). Digital filter 66 processes the digital output signal 64 of the sigma-delta modulator, which is a stream of ones and zeros. These data are then scaled accordingly using the calibration factor (s). The filtered data 68 is then read via a serial interface 69 or similar and sent to a computer that performs correlation calculations for fragments of data packets. Data from interface 69 is then correlated (e.g., autocorrelated and / or cross-correlated) as explained here. Figure 26 is similar to Figure 25, except that in Figure 26 the matrix of sensor capacitors C1 - C4 is illustrated, which are multiplexed via a multiplexer.

[0078] Fig. 27 is a block diagram illustrating an example of sigma-delta modulation that can be performed in the modulator 60 of Figs. 25-26. Again, this modulation can be performed in an electrical circuit assembly, firmware and / or software. The analog output signal Vout (or Vo) of the circuit of Fig. 4 (and Fig. 5) is received by the combiner 70 of the 60 sigma-delta modulator. The combiner 70 receives the analog Vout (or Vo) signal as well as the feedback signal from the feedback loop 71 of the modulator 60. The signal from the output of the combiner 70 is received by the integrator 72, in turn the output signal which is received by the quantizer 74, e.g. a one-bit quantizer. Digital output 64 is then filtered 66, as explained above, and a little further. Sigma-delta modulation is advantageous in that it provides oversampling and allows the processing of noise, such as electromagnetic interference type and reduction of its adverse effects. In particular, the noise is distributed by sigma-delta modulation on the frequency band, so that the signal / noise (S / N) ratio can be improved.

[0079] Referring again to Figure 4, each capacitor (C1, C2, C3, C4) is discharged before the next one is charged. The discharge process of each capacitor is described with reference to the reset pulse, with reference to Figs. 5-6.

53 / 51P26341PL00

[0080] Fig. 5 is a schematic diagram of the sensor circuit of Fig. 4 with respect to the erase cycle. During the erase cycle, the previously charged capacitor (C1, C2, C3 or C4) is discharged before the next write cycle. Figure 6 contains examples of signals used during the erase cycle (s). No read is made during the erase phase. During the cycle or erase phase, the transistor Q7 is turned off (the ClkWr write pulse is absent), and the transistors Q5 and Q6 are turned on via the ClkEr erase pulse (see also Fig. 6). Thus, the capacitor (C1, C2, C3 and / or C4) discharges to the mass level (e.g. V = 0) or virtual mass (VG), similar to Cint. Again, the Cint capacity mimics the capacity of the Cs sensor. After connecting the capacities Cs and Cint with the mass and discharge, the impulse and the delete cycle are completed. Then another capacitor (C1, C2, C3 or C4) can be prepared, charged and read.

[0081] Referring to Fig. 4 - 6 it should be noted that the rain sensor comprises: a sensing circuit comprising at least a first and a second sensing capacitor (e.g. C1 and C2) which are sensitive to moisture present on the outer surface of the glass, as well as at least one mimicking capacitor (Cint), which mimics at least the charging and / or discharging processes of at least one sensing capacitor, first and second; wherein the write pulse (ClkWr) charges at least the first sensing capacitor (e.g. C1), and the erase pulse (ClkEr) substantially discharges each of the first sensing capacitor (e.g. C1) and the mimic capacitor (Cint); wherein the presence of rain on the outer surface of the glass in the sensing field of the first sensing capacitor (e.g. C1) causes voltage fluctuations (see Vo or Vout) on the output electrode of the mimicking capacitor (Cint) in proportion to the voltage fluctuation on the output electrode (8) of the first sensing capacitor ( for example C1), even though there is no rain in the field of the mimic capacitor (Cint), and rain is also detected on the basis of the output signal (Vo or Vout) from the output electrode of the mimic capacitor (Cint), the output signal being read at least between the end of the recording pulse (ClkWr) and the beginning of the erasing pulse (ClkEr) (see "reading area "In Fig. 6).

[0082] Referring still to Fig. 5, during the erase cycle the erase pulse ClkEr discharges the capacitor (C1, C2, C3 and / or C4) and thus also

53 / 51P26341PL00

EP 1 971 507 B1 of a Cint-mimicking capacitor, to ground (for example a fixed potential of V = 0) (see traditional mass symbol in Fig. 5). However, it has been noted that the set mass can lead to some problems. Thus, during the erase cycle, the ClkEr pulse discharges the capacitor (C1, C2, C3 and / or C4) and thus also the Cint-imitating capacitor to the level of the virtual mass VG that is floating (see VG and the mass symbol in Fig. 5). In other words, the electrode of each of the C1 - C4 capacitors is floating. It may be at floating potential / voltage or reference potential / voltage. It has been noted that a floating or virtual mass can be very beneficial (for example, a floating mass and / or a capacitor electrode (s) can lead to a significant reduction of problems associated with electromagnetic interference). For example, this type of floating or virtual mass can help reduce the risk of sensor system confusion due to electromagnetic interference. In this regard, reference is made to Figures 28 (a) and 28 (b) (together with Figure 5).

[0083] In Figs. 28 (a) - (b), reference numbers 7 and 8 designate capacitor electrodes (e.g., C1, C2, C3, C4). In these figures, the symbol "q" refers to charge, and φ refers to potential (φ1 is different from φ2). In Fig. 28 (a), the capacitor (e.g. C1) is grounded at a fixed potential, e.g. zero volt (the charge at grounded electrode 7 is set at + q). In this regard, when the charge of the grounded electrode 7 is set to + q, when some external EB object approaches the sensing capacitor area (e.g. a human finger with a higher dielectric constant) (e.g. touching the front surface of the glass above the capacitor), then this external object induces a charge change -Aq, and the state of the second electrode 8, which is not fixed, changes from the charge -q to the charge -q + ​​Aq, trying to balance the charge. So if the capacitor is grounded at a fixed potential, for example zero volt, then when reading the output voltage from the capacitor, the change caused by the change in charge Aq will be read, which is unnecessary and can lead to false readings. By comparing Figures 28 (a) and 28 (b), in Fig. 28 (b) illustrates the benefit of making the sensing capacitor 7 electrode (e.g., any of the C1-C4 capacitors) floating (e.g., floating or virtual ground). In Fig. 28 (b), the charge q at the electrode 7 is not fixed. For example, the charge on electrode 7 changes from + q 'to + q' 'when the external object

53 / 51P26341PL00

EP 1 971 507 B1 comes into contact with the windshield in the sensor area of ​​the capacitor, thereby indicating the floating nature of the electrode. In Fig. 28 (b), when an external object (e.g. a human finger) is applied to the glass in the capacitor sensing area, the free charges on both electrodes 7 and 8 of this capacitor change. Thus, the adverse effect of the charge change Δq is eliminated or reduced by the use of a floating or virtual mass (electrode 7 is floating). In particular, when the electrode 7 is floating, as in Fig. 28 (b), the external object (EB) does not adversely affect the charge summation, since the addition of charges (+ q '' and -q '') of the electrodes 7 and 8 in the presence of this external object gives zero or essentially zero. Also, false readings due to electromagnetic interference can be reduced by using this floating mass. The floating nature of the electrodes may therefore allow the absolute values ​​of the charges q on the capacitor electrodes 7 and 8 to be the same or substantially the same, even if an external object is present, because the electrode 7 is floating and is not fixed on the mass potential. This is one example of why it can be beneficial to make electrodes 7 of capacitors C1 - C4 fluid or set them on the virtual mass VG, as shown in Fig. 5. Referring to Figs. 5 and 28, the sensor capacitors C1 C4 are floating and both electrodes are isolated from the ground. As said, the rain sensor comprises at least one sensing capacitor (C1, C2, C3 and / or C4) which is sensitive to moisture on the outer surface of the pane, the sensing capacitor comprising a first capacitor electrode (8) which receives a charging signal and a second capacitor electrode (7) remote from the first capacitor electrode (8); wherein the second capacitor electrode (7) of the capacitor is floating, so that the sensing capacitor is isolated from ground.

[0084] Fig. 6 is an exemplary time chart of signals applied or read from the system of Figs. 4-5 during write / delete modes / cycles. As noted above, capacitors (C1 - C4) are sequentially charged, read, quantized and erased. Fig. 6 shows the clock pulses for recording (ClkWr) and erasing (ClkEr) for each capacitor C1-C4 in sequence. Then the voltages are quantized and output. Variable output voltage Vo1 - Vo4 corresponds to the capacitors C1 - C4, and therefore Cint. It should be noted that the Vo1-Vo4 output signals in Fig. 6 are taken at the Vout level

53 / 51P26341PL00

EP 1 971 507 B1 (or Vo) in Figs. 4-5. Furthermore, in Fig. 6, the output Vo signals are read or analyzed (for example, for autocorrelation and / or cross-correlation) in the peak reading areas (see "Read" in Fig. 6. output signals where the output signals are substantially stabilized and / or the capacitor is saturated. In particular, the output signal Vout (or Vo) in Fig. 6 for a given capacitor (C1) it is read in the "reading area" after the write pulse (ClkWr) for this capacitor has ended, and before and / or until the start of the delete pulse (ClkEr) for this capacitor.

[0085] Still referring to Fig. 6, for example a raindrop on the outer surface of the windshield will affect the size of the output signal (s) Vout (or Vo). For example, a drop of water in the area of ​​a given capacitor (for example C1) will cause the level of output signal (s) Vout (or Vo) for this capacitor in the area of ​​"reading" the signal will be higher compared to the situation of the absence of this drop of water. The exact size or level depends on the size of this drop of water. With increasing amounts of water, the signal size in the "reading" area increases because the dielectric constant of water is higher than that of glass and / or air, which increases the capacity. Similarly, if there is no water droplet on the windshield above the area of ​​a given capacitor (for example C1), then this will cause a decrease in the output level of the signal (s) Vout (or Vo) for this capacitor in the "reading" area compared to the situation if such a drop was present.

[0086] Signals from the capacitors can be converted from analog to digital via a sigma-delta modulation scheme or the like that can be implemented at the program level or in any other suitable manner, e.g. in hardware. The principle behind the sigma-delta architecture is to make rough estimates of the signal, measure the error, integrate it, and then compensate for that error. Data can be oversampled at a given frequency, for example, at least 32 kHz, more preferably 64 kHz, however it should be noted that other sampling frequencies can be used. The quantization run can be reproduced using a sigma-delta modulation scheme to produce a simple binary output 0 or 1 corresponding to the on and off state, respectively. The sigma-delta modulation scheme can therefore be used to reduce noise (na

53 / 51P26341PL00

EP 1 971 507 B1 for example in the tail of the signal) and producing a digital output stream (e.g. ones and zeros).

[0087] Before thoroughly discussing the work and mathematical model on which the algorithm of an exemplary sensor is based, an overview of the states in which the sensor and / or wipers may be located in relation to Fig. 7 is provided, which is an exemplary state diagram showing how it can be use data related to autocorrelation and cross-correlation to control vehicle wipers. The system starts operating in the S702 Start / Initialization state. In this state all buffers are cleared. Based on the inputs of the C1, C2, ... Cn capacitors, analog-to-digital conversion of signals from the respective inputs is performed via sigma-delta modulation. The data are read for numerous channels in a certain period of time T. Status S704 Mode Selection works as a switch for choosing between manual or automatic wipers operating mode. If the status of the S704 Mode Selection indicates that the manual mode is selected, then in the S706 state of the Manual Mode, the automatic mode can be turned off and the pre-existing manual mode activated. Then the system returns to the S702 Start / Initialization state. However, if the Mode S704 of the Mode Selection indicates that the automatic mode is selected, then the automatic mode is turned on in the state S708 of the Automatic Mode.

[0088] At least three calculations are performed in state S710 of the Auto Correlation Engine. First, normalized autocorrelation is calculated for each sensor input signal input. Secondly, the autocorrelation gradient is calculated. Thirdly, the difference between the signal input and the non-disturbed reference signal (Δ1) can be calculated. This information is forwarded to the state S712 Is it raining ?, in which at least three conditions are checked to see if it is possible that it is raining, there is moisture on the windshield and the like. The likely rain indications are that the autocorrelation gradient is greater than 1, all autocorrelation values ​​are positive and / or the Δ1 value is greater than some predetermined threshold t1. If these conditions are not met, the system goes to state S714 Park Wipers / Stop Engine, in which the wipers are parked (if they are moving) or are not started, and the engine is stopped (if it is just turned on) and the system returns to the state S702 Start / Initialization.

53 / 51P26341PL00

[0089] On the other hand, if all the conditions are met (for example, it is likely that there is water, moisture or other disturbance on the glass, etc.), the system goes into the S716 Smallest Speed ​​state in which the engine the wiper is operated at the lowest speed available. In the S718 condition of the Cross-Correlation Motor, the cross-correlation between input signals from capacitors is calculated. The shape of the cross-correlation curve is determined and the symmetry of both sides of the cross-correlation curve is checked. As will be described below, these checks help, for example, to determine the type of disturbance (e.g. light rain, heavy rain, fog, snow and the like) encountering a windshield (e.g. windshield). In state S720 Estimation of precipitation intensity, the "precipitation level" is determined (for example, intensive, light and so on). Based on this determination, the S722 Speed ​​Selection state starts the engine at the appropriate speed. Eventually the system returns to the S702 Start / Initialization state to determine if there have been any changes in the conditions outside the car.

[0090] The steps performed by the rain sensor will be described in more detail with reference to Fig. 8, which is an example flowchart showing how to use autocorrelation and cross-correlation data to control wipers. In Fig. 8, the buffers are cleaned in step S800 and the data generated in the circuit of Figs. 4-5 (for example, from Cint or from capacitors C1 - C4) are subjected to sigma-delta modulation and are read in step S802. [0091] An algorithm for determining whether wipers should be turned on and, if so, at what speed, is started by autocorrelating data in step S804 subjected to sigma-delta modulation. Auto-correlation can be used to analyze a function or series of values, for example signals in the time domain. Autocorrelation is a cross-correlation of a signal with itself. Autocorrelation is used to find repetitive or substantially repetitive patterns in a signal, such as, for example, determining the presence of a periodic signal hidden in noise, identifying the fundamental frequency of a signal that does not actually contain this frequency component, but is suggested by the content of multiple harmonic frequencies and the like. Cross-correlation is a measure of the similarity of two signals and is used to find properties in an unknown signal by comparing it with a signal

53 / 51P26341PL00

EP 1 971 507 B1 known. In other words, it can be used to perform "fingerprinting" patterns in the signal. Cross-correlation is a function of the relative time between signals. Digital signals from any two capacitors (e.g. C1 and C2) are cross-correlated in close spatial proximity, and the system looks for any degree of correlation with non-zero time delays. This type of space-time cross-correlation allows the system to extract patterns like electrically, falling rain, projecting itself onto the sensor matrix. For example, this system may take the case of raindrops moving over one capacitor C1 at the time t0 and the same droplet "hitting" another capacitor C4 (spaced at a distance L from the capacitor C1). If this drop moves at an average speed of Vi, time (t0 + T), where T = L / Vi, the cross-correlation function will have another extreme or bend. The normalized amplitude for this extreme value may allow the system to determine the degree of intensity of rain falling on the sensor.

[0092] Each capacitor C1 - C4 has an autocorrelation function related to the digitized voltage Vout resulting from its reading (or the corresponding Cint reading). The autocorrelation function depends on the time difference rather than the actual time. Calculation of autocorrelation is advantageous because it allows, for example, to deduce the basic frequency regardless of phase. Auto-correlations have an advantage over other methods, such as Fourier transforms (which can also be used), which only provide information about harmonic content. The use of autocorrelation for readings from C1 - C4 capacitors (which, as explained above, includes appropriate readings from a Cint mimetic capacitor) can be used to detect and distinguish between water, dirt, dust, droplets, droppings and the like.

[0093] It should be noted that in this document, the data from the Cint capacitor is considered to be data from the C1 - C4 capacitors, since the capacity Cint mimics or substantially mimics the capacities of C1 - C4, as explained above. So, when we talk about receiving data from capacitors (for example C1 - C4) this refers to and includes receiving data from Cint capacity. In other words, for the output of the system from Fig. 4 - output from C1 - C4 capacitors is considered to be even if this signal is not taken directly from them.

53 / 51P26341PL00

B (r, t) = <

[0094] Rain as a function of time can be represented by the following formula:

the rain casts electrically ₀ otherwise

Basically b takes a binary value indicating whether it is raining (1) or not (0). Note that b contains at least two bits and that 24 bits can be used for sigma-delta modulation. It should also be noted that a scale can be introduced, potentially to capture more voltage related data on capacitors C1 - C4 (or Cint).

[0095] At the end of the sampling cycle L, for example, the output of the system of Figs. 4-5, for example from a matrix of four capacitors C1-C4 (or via Cint) is in the range 0000 to 1111 using digital binary data. A single bit enabled can initiate wiper operation. In the event that all bits are turned off (0000) or all bits are turned on (1111), the wipers cannot be initiated because there is probably nothing on the windshield, the car is completely covered with water and the like, because all the capacitors in the matrix read the same that does not correspond to rainfall on the glass. So the most likely events where wipers will be needed are those in the range from 0001 to 1110 (that is, when the outputs of all capacitors in the matrix are not the same). When the data has a value in this range or even not in this range, correlation functions (autocorrelation and / or cross-correlation) can be performed using the following integral. It should be noted that the integral below may be written in a different form, for example in the form of a sum. Correlations between two drops over a long period of time can be calculated according to the following formula:

₁L ^Rb ((vol ^t2 ) = 7 J ^t1 + ^tt2 + ^{t) h}^L₀

Rb ((, t; r₂, t2) = Rb (Ar, At)

53 / 51P26341PL00

Where Rb is a binary event correlation, given as a function of resistance ri in given time moments ti; and L is the long sampling period during which the data sequence is captured. The sampling period L can be from about 10 to 100 ms, and more preferably from about 20-30 ms, which corresponds approximately to the average frequency that can be distinguished by the human eye. Rb is also equal to the function of correlation of changes in resistances on Ar capacitors and time change. When Ar = 0, the autocorrelation value is determined because data from the same capacitor is analyzed, and if Ar a 0, then cross-correlations are calculated on data from different capacitors.

[0096] These functions are subject to several examples of limitations and assumptions. First:

Ar = Vi At

This limitation generally means that a drop of water or similar disorder travels over a given time scale. Secondly:

b (r + Vi At, t + At) =) (r, t).

This limitation mimics or essentially mimics what happens when water droplets or the like move from one capacitor to another. Therefore, correlation functions can be considered as discrete stages p in space and T in time. This property can be mathematically represented by the following equation:

R_b (mp, nT) ξ R (Vi At, At)

Basically, the left side of this equation establishes a theoretical grid in space and time over which a drop of water travels. For example, Fig. 9 is an exemplary stylized view of how a rain droplet may move. Fig. 9 shows a rain drop moving on the windshield in the XZ plane during the initial period of time (t = 0) and at some later period of time (t = T). The assumption that the distribution of drops is homogeneous in space and time allows the creation of a binary field created by rain, which is broadly stationary. This system also assumes that the time correlation between preferred pixels in the same neighborhood is high in the direction of rain. Finally, the degree of autocorrelation and cross-correlation over time quantizes rainfall and other disorders.

53 / 51P26341PL00

[0097] It should be noted that computation time can be saved due to the nature of the correlation matrix and the nature of rainfall. For example, correlation matrices can be symmetrical. In addition, as another example, due to the fact that the rain tends to fall from the sky and move up the windshield, it may be sufficient to compare only capacitors that are arranged vertically with respect to each other in cross-correlation, while ignoring adjacent capacitors level.

[0098] It should be noted that although binary data is used, grayscale data can also be used for the outputs of the system of Figs. 4-5 or similar or other suitable system (s).

[0099] After autocorrelation in step S804 (for example, using the equations discussed above or other useful correlation equations), one or more checks may be performed to increase the system accuracy. Examples of this type of check (for example, if the Rxx autocorrelation data are negative, if the gradient is greater than one and / or the shape of the Rxx curve is different or substantially different from the non-disturbed normalized autocorrelation data stored in memory) are listed at the bottom of the frame of step S804 on Figure 8. One, two or all three of these checks can be performed.

[0100] For example, one check of the autocorrelation data in step S806 may be to determine whether the autocorrelated data from one or more capacitors (C1, C2, C3 and / or C4; or via a Cint mimetic capacitor) contain negative values. For example, when the autocorrelated data contains negative values, then the system or method may indicate that there is no rainfall, park the wipers and / or do not operate the wipers (see step S808). This test is intended to determine, for example, whether the detected disorder is actually rainfall. To this end, Fig. 10 is a graph of sample experimental data for maximum values ​​of non-normalized autocorrelation for various disorders. In fig. 10 water signals are shown to be greater than undisturbed signals and are positive, as well as external interference, such as electromagnetic waves from the CB radio and human contact with the glass are generally below undisturbed levels and may be negative. Thus, to eliminate or reduce false detections under the influence of external disturbances, such as human touch the glass, radio signal interference and so on

53 / 51P26341PL00

Similarly, any signal with negative autocorrelation values ​​is considered "non-rainy" events. Negative autocorrelation values ​​may be considered, or other measures may be taken to eliminate or reduce false detections due to external interference by, for example, comparing gradients (for example, any curve lower or smaller than the undisturbed curve / graph of Fig. 10, can be considered as a "non-rainy" event), shielding of capacitors and other activities.

[0101] A second example test of autocorrelation data is to check if the autocorrelation curve gradient associated with autocorrelated data is greater than one, and if this is not the case, then the system or method may indicate that there is no rainfall, park the wipers and / or do not start vehicle wipers (see step S808). For this test, the normalized autocorrelation gradient of the disorder is checked. The gradient of the normalized autocorrelation of the undisturbed signal is close to one. Gradient measurement is beneficial because it is not affected by temperature change. The rain sensor can thus be substantially immune to false readings due to temperature changes in certain example embodiments of this invention. Gradients with a value less than 1 (or some other predetermined value) can be considered as non-rainy events.

[0102] A third example test of autocorrelation data is to determine whether there is a match or substantial match between the autocorrelation curve associated with the autocorrelated data and one or more predefined autocorrelation curves present in the database and / or memory. When the shape of the autocorrelation curve associated with the autocorrelation data of the system of Fig. 4 - 5 is different or substantially different from the autocorrelation curve associated with normalized undisturbed autocorrelation data, this can be considered a non-rain event and it can be indicated that there is no rainfall, the wipers may be parked and / or the wipers may not be activated (see step S808 ). However, when there is a match or substantial match between the autocorrelation curve associated with the autocorrelation data of the system of Fig. 4 - a predetermined autocorrelation curve related to the presence of moisture, for example rain, then it can be said that it is actually raining, the wipers can be activated or kept in motion.

53 / 51P26341PL00

[0103] In this regard, the shape of the autocorrelation curve can be used to reduce false wipes and / or false detections. In particular, normalized autocorrelation of a non-disturbed signal is used as a reference. Then, the normalized autocorrelation of each signal measured in the circuit of Figs. 4-5 is compared with a reference to identify the closest "fingerprint" pattern in some example cases. In general, the more water is present in the measurement area, the greater the difference between the reference signal and the observed signal. Thanks to this, you can compare correlation snapshots with reference snapshots of well-known events, such as the presence of rain, dirt, lack of disturbance, ice and so on. In general, correlation snapshots can be normalized, although the invention is not limited to this. Correlation snapshots favorably plot r values ​​as a function of time quanta over a discrete time interval.

[0104] When there is a match or substantial match between the autocorrelation curve associated with the autocorrelated data derived from the Figs. 4-5 system and the predetermined autocorrelation curve associated with a non-moisture substance such as dirt, then this can be considered an event non-rainy and it can be said that there is no rainfall, the wipers can be parked and / or not activated (see step S808).

[0105] It should therefore be noted that the shape of the autocorrelation curve resulting from the data derived from the circuit of Figs. 4-5 (from capacitors C1-C4 or via Cint) can be used to reduce false wipes as a third condition. For example, a normalized autocorrelation curve of a non-disturbed signal can be used as a reference. Then normalized autocorrelation of each signal taken from the circuit of Fig. 4 - 5, is compared with a reference to identify the nearest "fingerprint" pattern. Generally, the more water is present in the sensor area, the greater the difference between the reference signal and the observed / measured signal. This allows correlation snapshots to be compared with reference snapshots of well-known events. In general, correlation snapshots are preferably normalized. Correlation snapshots favorably plot r values ​​as a function of time quanta over a discrete time interval.

[0106] A potential problem with capacitive rain sensors is that rapid temperature changes (e.g. as a result of using absorbing

53 / 51P26341PL00

The radiation of the black frit used to cover the sensor pattern cosmetically) causes a change in the "constant" dielectric (permeability) of the glass. This is then recorded as a change in capacity and may be misinterpreted as a rain signal. However, the normalized autocorrelation function is unchanged or essentially unchanged for different temperatures even though there may be differences for non-normalized autocorrelation functions for different temperatures. Thus, the sensor system is unaffected or essentially unaffected by temperature changes.

[0107] In addition, extremely small water accumulation, such as ultra fine mist, for example, can slowly rise to a level that will turn on sensors based on Nyquist frequency converters. During observation, which is associated with human vision (e.g. 30-60 Hz), the autocorrelation function is able to distinguish between ultra slow mist or condensation accumulation and normal fog and rain.

[0108] Figures 11A-11D illustrate example experimentally obtained correlation snapshots. These correlation snapshots or event "fingerprint" patterns can be stored as reference "fingerprint" patterns or correlation curves. Observed / measured correlation snapshots (e.g., autocorrelation curves) can be compared to these reference standards to determine the type of event occurring. For example in fig. 11A is an experimental autocorrelation snapshot indicating intense rain. Figure 11B shows an experimentally obtained autocorrelation snapshot indicative of a light fog. Fig. 11C is an experimental autocorrelation snapshot indicating interference with a CB radio signal. Fig. 11D shows an experimentally obtained autocorrelation snapshot indicating a grounded object with voltage. It should be noted that these patterns are given as examples and reflect experimental data. Actual events can differ in many ways. Thus, if it is determined that there is a match or substantial match between the autocorrelation curve associated with the autocorrelation data from the circuit of Figs. 4-5, a predetermined non-moisture related autocorrelation curve, such as Fig. 11C or Fig. 11D, then it can be considered a non-rain event and it can be concluded that there is no rainfall, the wipers can be parked and / or not activated (see step S808). However, if it is determined that there is a fit or substantial fit between the curve

53 / 51P26341PL00

Autocorrelated data related to the autocorrelated data from the circuit of Figs. 4-5 a predetermined autocorrelation curve related to the presence of moisture, such as Fig. 11A or Fig. 11B, then this can be considered a rainy event, and it is raining, the wipers can be turned on and / or kept in motion. In addition to the predetermined autocorrelation curves of Fig. 11A-11D, other "fingerprint" reference standards may be stored and / or compared with observed correlation snapshots in other example embodiments of this invention.

[0109] Returning to Fig. 8, it is determined in step S808 that each of the three conditions shown at the bottom of frame S804 are met. In particular, it is determined in step S806 that each of the following conditions is met: (a) the autocorrelated data does not contain negative values; (b) the autocorrelation curve gradient associated with these autocorrelated data is greater than a predetermined value, for example one; and (c) whether the shape of the autocorrelation curve associated with the autocorrelated data of the perimeter of FIG. 4 - 5 is different from the predetermined autocorrelation cams associated with undisturbed autocorrelation data. If all of these conditions are not met, then this suggests a non-rain event and the process proceeds to step S808, in which the vehicle wipers are parked (if moving) or held off and the S800 initialization begins again. However, if all these requirements are met at step S806, then the process proceeds to step S810 and the wipers of the vehicle (e.g., windshield wiper) are operated at the lowest speed.

[0110] Fig. 13 shows an example of autocorrelation. In Fig. 13, the values ​​from the sensor capacitor C1 (or associated therewith) occurring in subsequent moments -t2, -t1, t0, t1, t2, and t3 have the values ​​0, 0, 1, 1, 0, and 0, respectively. Autocorrelation for the moment 0 (aco) is determined by multiplying the values ​​associated with the capacitor C1 without offset, and then adding or adding up the results. In Fig. 13 it can be seen that in this case aco is equal to 2. Accordingly, in the autocorrelation plot at the bottom of Fig. 13, the entry for moment 0 is made for an autocorrelation value of 2. It should be noted that the autocorrelation plot at the bottom of Fig. 13 is similar but simpler than the autocorrelation plot in Fig. 10 and the autocorrelation values can be obtained for Fig. 10 in a similar manner. Then, still referring to Fig. 13, it is performed

53 / 51P26341PL00

Autocorrelation using the capacitance values ​​associated with the C1 capacitor for the next point in time to obtain the ac1 autocorrelation value. This next autocorrelation value (ac1) is obtained by shifting the value sequence for capacitor C1 from the bottom row relative to the top row, as shown in Fig. 13, and then multiplying the values ​​in those lines that were in one line and adding up the results. In fig. 13 it has been shown that the value of ac1, for the moment 1, is 1. Therefore, this autocorrelation value of 1 for the moment t1 can be entered on the graph at the bottom of Fig. 13 and a line can be drawn between the two entered data points to illustrate an example and understanding . Then, for the next time value (or time delay), the bottom row is again shifted by another segment relative to the top row and this process is repeated and so on. It can be seen that the autocorrelation plots in Fig. 10 can be obtained in a similar manner. In Fig. 13, it can be seen that cross-correlation can be performed by replacing the values ​​associated with the C1 capacitor in the bottom row with values ​​derived from or associated with another capacitor, for example C2 (or C3 or C4).

[0111] Autocorrelation and / or cross-correlation testing can also help distinguish between, for example, light rain and heavy rain. For example, if only the autocorrelation in time is high (while the cross-correlation is low), then there is probably only light rain. Fig. 12A is an example correlation matrix showing light rain. In fig. 12A it is important that the correlations between C1 and C1, C2 and C2, C3 and C3 and C4 and C4 (these are autocorrelation) in a given period are large, and the rest of the correlation (cross-correlations) are small. As a result of hypothesis and confirmed experimental data, this type of matrix will indicate light rain.

[0112] On the other hand, if both autocorrelation and cross-correlation in time between the capacitor signals are high, there is probably heavy rain. Fig. 12B is an example correlation matrix showing heavy rain. In fig. 12B not only the auto-correlations of individual capacitors are high (i.e., the auto-correlations are correlations between capacitors C1 and C1, C2 and C2, C3 and C3 and C4 and C4), but also the correlations between different capacitors are high (correlations in Fig. 12B located on the diagonal from the upper left corner to the lower right corner are autocorrelations, while the others are cross-correlations). As a result of hypothesis and confirmation

53 / 51P26341PL00

With experimental data, a matrix of this kind will indicate heavy rain. The degree of cross-correlation can be quantized to determine the relative rainfall. These data can in turn be used to turn on different wiper speeds appropriate to the intensity of the rain. For example, the more cross-correlations that are large, the higher the wiper speed used.

[0113] More systematically, in step S812, cross-correlations are calculated (correlations between data associated with different capacitors), and both sides of the cross-correlation curve are used to determine the level of symmetry L. If the level of symmetry is lower than the predetermined threshold tmin , step S814 directs the system to step S816, where the wipers are run at the lowest speed and the system returns to initialization at step S800. If the level of symmetry is greater than tmin but less than the arbitrary value of t, step S818 directs the system to step S820, where the wipers are operated at a higher or medium speed and the system returns to the initialization stage S800. It should be noted that any arbitrary values ​​of ti can be determined, and a level of symmetry between ti and ti + 1 will activate the corresponding corresponding wiper speed and then return the system to the S800 initialization stage. Finally in step S822, if the level of symmetry is greater than the predetermined tmax, step S822 directs the system to step S824, where the wipers are operated at the highest speed and the system returns to the step of initialization S800. Correlations of the output data of the circuit of Figs. 4-5 can therefore be used to adjust the wiper speed. The more high the cross-correlation is, the higher the wiper speed is used due to the greater likelihood of heavy rain.

[0114] Figs. 14-24 are examples of cross-correlations performed. Fig. 14 shows cross-correlated data, while Figs. 15-24 are cross-correlation plots of some of Fig. 14 data where rain has been detected. In Figs. 15-24, each interval on the horizontal axis is one microsecond (1 ps) for the purposes of the example, and sampling was performed every one microsecond. As explained above with reference to Fig. 13, in Fig. 15 - 24 at the time = 0 (delay 0) there is no time shift of correlated values ​​from different capacitors. Fig. 14 illustrates that when rain was present (see signals S1 - S5 and W1 - W5), the delta signals regarding autocorrelation were large. On

53 / 51P26341PL00

Figs. 15-24 are cross-correlation plots associated with these signals. It is helpful to find the symmetry between the graphs on the left and right of each of Figs. 15-24 (one side of zero is compared to the other side of zero). Generally speaking, if there is symmetry around the zero delay axis, there is not much cross-correlation, indicating that the detected rain is not very intense. However, if there is asymmetry around the zero delay axis, this means more cross-correlation and indicates the presence of heavy or more intense rain. For example, you can see the asymmetry in Figs. 18, 19 and 23 around the zero axis as a result of humps or valleys on one or both sides. More cross-correlation indicates that raindrops are moving from the sensing area of ​​one capacitor to the sensing area of ​​another capacitor. In this respect, each interaction of a raindrop and windshield surface has its own time-domain correlation signature. A high cross-correlation value indicates that the same droplet is detected by different capacitors at different times (for example, see also Figure 9). It should be noted that the lower case letter "t" in Fig. 9 is the same as the delay axis in Figs. 15-24.

[0115] It should be mentioned that a moisture sensor (e.g. a rain sensor) can detect rain on the vehicle window without the need for a reference capacitor. Spatial-time correlation can be used. All capacitors or multiple capacitors in the sensor matrix may be identical or substantially identical in shape in some embodiments. For example purposes, at a given point in time (e.g. t1), the system may compare the values ​​associated with the C1 capacitor with the values ​​associated with the C2 capacitor and / or the values ​​associated with another capacitor. At the moment t1, the system can also compare values ​​related to the C1 capacitor with itself (autocorrelation) and can also compare the autocorrelation for the C1 capacitor with the autocorrelation for the C2 capacitor and / or other sensing capacitor (other sensing capacitors).

53 / 51P26341PL00

EP 1 971 507 B1