System and method for automatically activating a windshield wiper of a vehicle

Abstract

A system and method for automatically activating a windshield wiper system of a vehicle having a front windshield with a front wiper and a rear windshield with a rear wiper is provided. The method includes evaluating at least one windshield classification of a road condition based on raw information and capturing a front image of the front windshield, a rear image of the rear windshield, and an ambient image of an environment. The method further includes classifying the images to define a first windshield class. The method further includes determining a front windshield perception, a rear windshield perception, and an ambient perception of the first windshield class to define a first combination of detection sources. The method further includes fusing the front windshield perception, the rear windshield perception, and the ambient perception to define a first front probability of the first windshield class. The method further includes activating the front wiper when the first front probability is greater than a front threshold.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to windshield wiper sensors, and more particularly to a system and method for automatically activating a vehicle's windshield wipers. BACKGROUND

[0002] Currently, rain sensors add an undesirable cost to the vehicle in which it is installed. Further, many vehicles with rain sensors have only a sensor for the front windshield and are configured only for rain detection. Many such sensors have low resolution and low reliability. Further, many current vehicles with rain sensors include windshield wipers with inconsistent speeds. SUMMARY

[0003] Therefore, while current windshield wiper systems achieve their intended purpose, there is a need for a new and improved system and method for automatically activating a vehicle's windshield wiper system.

[0004] According to one aspect of the present disclosure, a method of automatically activating a vehicle's windshield wiper system is provided, the vehicle having a front windshield with a front wiper and a rear windshield with a rear wiper. The method includes providing raw information for condition assessment; and assessing at least one windshield classification of a road condition with a condition assessment module based on the raw information. In this aspect, the method further includes capturing a front image of the front windshield, a rear image of the rear windshield, and an ambient image of an environment. The method then includes classifying the front image, the rear image, and the ambient image into the at least one windshield classification with a conditional windshield perception module defining a first windshield class. The method further includes determining a front windshield perception, a rear windshield perception, and an ambient perception of the first windshield class with the conditional windshield perception module to define a first combination of detection sources.

[0005] In this aspect of the present disclosure, the method includes fusing the front windshield perception, the rear windshield perception, and the ambient perception of the first combination of detection sources with a weighted voting module to define a first front probability of the first windshield class. The method then includes activating the front wiper when one of a first front probability greater than a front threshold and a first front probability being greatest in magnitude of the at least one classification occurs.

[0006] In one example, the raw information includes a vehicle interior temperature, a vehicle exterior temperature, local weather information, a geographic climate, global positioning system (GPS) information of the vehicle for condition assessment. In another example, the at least one windshield classification is a road condition enabling activation of one of the front wiper and the rear wiper of the vehicle.

[0007] In another example of this aspect, the weighted voting module includes a first equation as follows:

[0008] .

[0009] P(X)_front_fusion is a first front probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0010] In yet another example, the method further includes estimating a first front intensity of the first windshield class with a conditional windshield perception module. Further, the method includes determining the first front intensity of the first windshield class. The first front intensity is based on a drop-off rate within a region of the front windshield. Further, the method includes adjusting a speed of the front wiper when the front wiper is activated based on the first front intensity.

[0011] In yet another example, the step of determining the first front intensity includes applying a linear fitting equation as follows:

[0012]

[0013] where In(drop) is a normalized intensity, V_drop is a drop-off rate, N_drop is a number of drops per area, and kl and k2 are calibration coefficients.

[0014] In one example, the method further includes fusing the front windshield perception, the rear windshield perception, and the environment perception of the first combination of detection sources with a weighted voting module to define a first rear probability of the first windshield class. Then, the method includes activating the rear wiper when one of the first rear probability is greater than a rear threshold and the first rear probability is maximum in magnitude over the at least one classification occurs.

[0015] In another example, the weighted voting module includes a second equation as follows:

[0016] .

[0017] P(X)_rear_fusion is a first rear probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0018] In another example, the method further includes estimating, with the conditional windshield perception module, a first rear intensity of the first windshield class; and determining the first rear intensity of the first windshield class. The first rear intensity is based on a drop rate within a region of the rear windshield. The method further includes adjusting a speed of the rear wiper based on the first rear intensity when the rear wiper is activated.

[0019] In one example, the step of determining the first rear intensity includes applying a linear fitting equation as follows:

[0020]

[0021] where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area of the rear windshield, and k1 and k2 are calibration coefficients.

[0022] According to another aspect of the disclosure, there is provided a system for automatically activating a windshield wiper of a vehicle, the vehicle comprising a front windshield having a front wiper and a rear windshield having a rear wiper. In this aspect, the system comprises a cloud server arranged remotely from the vehicle. The cloud server is arranged to provide raw information for a condition assessment. The system further comprises at least one sensor arranged on the vehicle and arranged to capture a front image of the front windshield, a rear image of the rear windshield, and an environment image of an environment. Furthermore, the system comprises an electronic control unit (ECU) arranged in the vehicle. The ECU is in communication with the cloud server and the at least one sensor.

[0023] In this aspect, the ECU comprises a condition assessment module arranged to assess at least one windshield classification of a road condition based on the raw information.

[0024] The ECU further comprises a conditional windshield perception module arranged to classify the front image, the rear image, and the environment image into at least one windshield classification defining a first windshield class. Furthermore, the conditional windshield perception module is arranged to determine a front windshield perception, a rear windshield perception, and an environment perception of the first windshield class defining a first combination of detection sources.

[0025] In this aspect, the ECU further comprises a weighted voting module arranged to fuse the front windshield perception, the rear windshield perception, and the environment perception of the first combination of detection sources, thereby defining a first front probability of the first windshield class.

[0026] Furthermore, the ECU comprises an activation module arranged to activate the front wiper when one of: the first front probability is greater than a front threshold value, and the first front probability is maximum in magnitude of the at least one windshield classification.

[0027] In one embodiment, the raw information includes a vehicle interior temperature, a vehicle exterior temperature, local weather information, a geographical climate, global positioning system (GPS) information of the vehicle for condition assessment.

[0028] In another embodiment, the weighted voting module includes a first equation as follows:

[0029]

[0030] where P(X)_front_fusion is a first front probability of the first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and w1, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0031] In yet another embodiment, the conditional windshield perception module is arranged to estimate a first front intensity of the first windshield class. Further, the ECU is arranged to determine the first front intensity of the first windshield class. The first front intensity is based on a drop-off rate within a region of the front windshield. Further, the ECU is arranged to adjust a speed of the front wiper based on the first front intensity when the front wiper is activated.

[0032] In yet another embodiment, the ECU is arranged to determine the first front intensity by applying a linear fitting equation as follows:

[0033]

[0034] where In(drop) is a normalized intensity, V_drop is a drop-off rate, N_drop is a number of drops per area, and k1 and k2 are calibration coefficients.

[0035] In another embodiment, the weighted voting module is arranged to fuse the front windshield perception, the rear windshield perception, and the environment perception of the first combination of detection sources to define a first rear probability of the first windshield class. Further, the activation module is arranged to activate the rear wiper when one of the first rear probability is greater than a rear threshold and the first rear probability is maximum in magnitude of the at least one classification occurs.

[0036] In yet another embodiment, the weighted voting module includes a second equation as follows:

[0037]

[0038] where P(X)_rear_Fusion is a first rear probability of the first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0039] In yet another embodiment, the conditional windshield perception module is arranged to estimate a first rear intensity of the first windshield class. Further, the ECU is arranged to determine the first rear intensity of the first windshield class, the first rear intensity being based on a drop rate within the area of the rear windshield. In addition, the ECU is arranged to adjust a speed of the rear wiper based on the first rear intensity when the rear wiper is activated.

[0040] In another embodiment, the ECU is arranged to determine the first rear intensity by applying a linear fitting equation as follows:

[0041]

[0042] where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area of the rear windshield, and kl and k2 are calibration coefficients.

[0043] According to another aspect of the present disclosure, there is provided a method of automatically activating a windshield wiper system of a vehicle having a front windshield with a front wiper and a rear windshield with a rear wiper. The method includes providing raw information for a condition assessment; and assessing at least one windshield class of a road condition using a condition assessment module based on the raw information. The method further includes capturing a front image of the front windshield, a rear image of the rear windshield, and an environment image of an environment. In addition, the method further includes classifying the front image, the rear image, and the environment image into the at least one windshield class using a conditional windshield perception module defining a first windshield class. The method further includes determining a front windshield perception, a rear windshield perception, and an environment perception of the first windshield class using the conditional windshield perception module to define a first combination of detection sources. The method further includes fusing the first combination of detection sources with respect to the front windshield perception, the rear windshield perception, and the environment perception using a weighted voting module to define a first front probability of the first windshield class.

[0044] In this aspect, the method further includes determining a front windshield perception, a rear windshield perception, and an environment perception of the first windshield class using the conditional windshield perception module to define a first combination of detection sources. The method further includes fusing the first combination of detection sources with respect to the front windshield perception, the rear windshield perception, and the environment perception using a weighted voting module to define a first front probability of the first windshield class.

[0045] In this aspect, the weighted voting module includes a first equation as follows:

[0046]

[0047] where P(X)_front_fusion is a first front probability of the first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0048] The method further includes fusing the first combination of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception with a weighted voting module to define a first rear probability of the first windshield class. The weighted voting module includes a second equation as follows:

[0049]

[0050] where P(X)_rear_fusion is a first rear probability of the first windshield class (X).

[0051] Additionally, the method further includes activating the front wiper when one of the first front probability is greater than a front threshold and the first front probability is greatest in magnitude of the at least one classification. Moreover, the method further includes activating the rear wiper when one of the first rear probability is greater than a rear threshold and the first rear probability is greatest in magnitude of the at least one classification.

[0052] The present disclosure can also include the following aspects.

[0053] 1. A method of automatically activating a windshield wiper system of a vehicle, the vehicle having a front windshield with a front wiper and a rear windshield with a rear wiper, the method comprising:

[0054] providing raw information for a condition assessment;

[0055] assessing at least one windshield classification of a road condition with a condition assessment module based on the raw information;

[0056] capturing a front image of the front windshield, a rear image of the rear windshield, and an environment image of the environment;

[0057] classifying the front image, the rear image, and the environment image into the at least one windshield classification with a conditional windshield perception module defining a first windshield class;

[0058] determining a front windshield perception, a rear windshield perception, and an environment perception of the first windshield class with a conditional windshield perception module to define a first combination of detection sources;

[0059] fusing, with a weighted voting module, the first combination of front windshield perception, rear windshield perception, and environmental perception of the detection sources to define a first front probability of the first windshield class; and

[0060] activating the front wiper when one of the first front probability is greater than a front threshold and the first front probability is greatest in magnitude of the at least one classification occurs.

[0061] 2. The method of Scheme 1, wherein the raw information comprises vehicle interior temperature, vehicle exterior temperature, local weather information, geographic climate, global positioning system (GPS) information of the vehicle for condition assessment.

[0062] 3. The method of Scheme 1, wherein the at least one windshield classification is a road condition that enables activation of one of a front wiper and a rear wiper of the vehicle.

[0063] 4. The method of Scheme 1, wherein the weighted voting module comprises a first equation as follows:

[0064]

[0065] where P(X)_front_fusion is a first front probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environmental perception, and wl, w2, and w3 are weights of detection sources of the front windshield perception, the rear windshield perception, and the environmental perception, respectively.

[0066] 5. The method of Scheme 1, further comprising:

[0067] estimating a first front intensity of the first windshield class with a conditional windshield perception module;

[0068] determining the first front intensity of the first windshield class, the first front intensity being based on a rate of precipitation within a region of the front windshield; and

[0069] adjusting a speed of the front wiper based on the first front intensity when the front wiper is activated.

[0070] 6. The method of Scheme 5, wherein determining the first front intensity comprises applying a linear fitting equation as follows:

[0071]

[0072] where In(drop) is the normalized intensity, V_drop is the drop rate, N_drop is the number of drops per area, and k1 and k2 are calibration coefficients.

[0073] 7. The method of Scheme 1, further comprising:

[0074] fusing the first combination of the front windshield perception, the rear windshield perception, and the environmental perception from the detection sources using a weighted voting module to define a first posterior probability of the first windshield class; and

[0075] activating the rear wiper when one of the first posterior probability is greater than a posterior threshold and the first posterior probability is greatest in magnitude of the at least one classification.

[0076] 8. The method of Scheme 7, wherein the weighted voting module comprises a second equation as follows:

[0077]

[0078] where P(X)_rear_fusion is the first posterior probability of the first windshield class (X), P(X)_front is the front windshield perception, P(X)_rear is the rear windshield perception, P(X)_road is the environmental perception, and wl, w2, and w3 are weights of the detection sources for the front windshield perception, the rear windshield perception, and the environmental perception, respectively.

[0079] 9. The method of Scheme 7, further comprising:

[0080] estimating a first posterior intensity of the first windshield class using a conditional windshield perception module;

[0081] determining the first posterior intensity of the first windshield class, the first posterior intensity being based on a drop rate within a region of the rear windshield; and

[0082] adjusting a speed of the rear wiper based on the first posterior intensity when the rear wiper is activated.

[0083] 10. The method of Scheme 9, wherein determining the first posterior intensity comprises applying a linear fit equation as follows:

[0084]

[0085] where In(drop) is the normalized intensity, V_drop is the drop rate, N_drop is the number of drops per area of the rear windshield, and k1 and k2 are calibration coefficients.

[0086] 11. A system for automatically activating a windshield wiper of a vehicle, the vehicle comprising a front windshield having a front wiper and a rear windshield having a rear wiper, the system comprising:

[0087] a cloud server disposed remotely from the vehicle, the cloud server arranged to provide raw information for condition assessment;

[0088] at least one sensor disposed on the vehicle and arranged to capture a front image of the front windshield, a rear image of the rear windshield, and an ambient image of an environment;

[0089] an electronic control unit (ECU) arranged in the vehicle, the ECU in communication with the cloud server and the at least one sensor, the ECU comprising:

[0090] a condition assessment module arranged to assess at least one windshield classification of a road condition based on raw information;

[0091] a conditional windshield perception module arranged to classify the front image, the rear image, and the ambient image into the at least one windshield classification defining a first windshield class, the conditional windshield perception module arranged to determine a front windshield perception, a rear windshield perception, and an ambient perception of the first windshield class to define a first combination of detection sources;

[0092] a weighted voting module arranged to fuse the front windshield perception, the rear windshield perception, and the ambient perception of the first combination of detection sources to define a first front probability of the first windshield class; and

[0093] an activation module arranged to activate the front wiper when one of a first front probability greater than a front threshold and the first front probability being maximum in magnitude of the at least one windshield classification occurs.

[0094] 12. The system according to scheme 11, wherein the raw information comprises a vehicle interior temperature, a vehicle exterior temperature, local weather information, a geographical climate, a global positioning system (GPS) information of the vehicle for condition assessment.

[0095] 13. The system according to scheme 11, wherein the weighted voting module comprises a first equation as follows:

[0096]

[0097] where P(X)_front_fusion is a first front probability of the first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0098] 14. The system of Scheme 11, wherein the conditional windshield perception module is arranged to estimate a first front intensity of the first windshield class, wherein the ECU is arranged to determine the first front intensity of the first windshield class, the first front intensity being based on a drop rate within a region of the front windshield, and wherein the ECU is arranged to adjust a speed of the front wiper based on the first front intensity when the front wiper is activated.

[0099] 15. The system of Scheme 14, wherein the ECU is arranged to determine the first front intensity by applying a linear fit equation as follows:

[0100]

[0101] where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area, and kl and k2 are calibration coefficients.

[0102] 16. The system of Scheme 11, wherein the weighted voting module is arranged to fuse the front windshield perception, the rear windshield perception, and the environment perception of the first combination of detection sources to define a first rear probability of the first windshield class, and wherein the weighted voting module is arranged to activate the rear wiper when one of the first rear probability is greater than a rear threshold and the first rear probability is greatest in magnitude of the at least one classification occurs.

[0103] 17. The system of Scheme 16, wherein the weighted voting module comprises a second equation as follows:

[0104]

[0105] where P(X)_rear_fusion is a first rear probability of the first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

[0106] 18. The system of paragraph 16, wherein the conditional windshield perception module is arranged to estimate a first rear intensity of the first windshield class, wherein the ECU is arranged to determine the first rear intensity of the first windshield class, the first rear intensity being based on a drop rate within a region of the rear windshield, and wherein the ECU is arranged to adjust a speed of the rear wiper based on the first rear intensity when the rear wiper is activated.

[0107] 19. The system of paragraph 18, wherein the ECU is arranged to determine the first rear intensity by applying a linear fitting equation as follows:

[0108]

[0109] where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area of the rear windshield, and ki and k2 are calibration coefficients.

[0110] 20. A method of automatically activating a windshield wiper system of a vehicle, the vehicle including a front windshield having a front wiper and a rear windshield having a rear wiper, the method comprising:

[0111] providing raw information for condition assessment;

[0112] based on the raw information, utilizing a condition assessment module to assess at least one windshield classification of a road condition;

[0113] capturing a front image of the front windshield, a rear image of the rear windshield, and an ambient image of the environment;

[0114] classifying the front image, the rear image, and the ambient image to the at least one windshield classification with a conditional windshield perception module defining a first windshield class;

[0115] determining a front windshield perception, a rear windshield perception, and an ambient perception of the first windshield class with a conditional windshield perception module to define a first combination of detection sources;

[0116] fusing the front windshield perception, the rear windshield perception, and the ambient perception of the first combination of detection sources with a weighted voting module to define a first front probability of the first windshield class, the weighted voting module including a first equation as follows:

[0117]

[0118] where P(X)_front_fusion is a first front probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively;

[0119] fusing the front windshield perception, the rear windshield perception, and the environment perception of the first combination of detection sources to define a first rear probability of the first windshield class using a weighted voting module comprising a second equation as follows:

[0120]

[0121] where P(X)_rear_fusion is a first rear probability of a first windshield class (X);

[0122] activating the front wiper when one of the first front probability is greater than a front threshold and the first front probability is greatest in magnitude of the at least one classification occurs; and

[0123] activating the rear wiper when one of the first rear probability is greater than a rear threshold and the first rear probability is greatest in magnitude of the at least one classification occurs.

[0124] Further areas of application will become apparent from the description provided herein. It should be understood that the description and specific examples are intended to be illustrative only and are not intended to limit the scope of the disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0125] The drawings described herein are for purposes of illustration only and are not intended to limit the scope of the present disclosure in any way.

[0126] Figure 1 is a schematic diagram of a system for automatically activating windshield wipers of a vehicle comprising a front windshield having a front wiper and a rear windshield having a rear wiper in accordance with one embodiment of the present disclosure.

[0127] Figure 2 is a flowchart of an overall method of automatically activating windshield wipers of a system in Figure 1

[0128] Figure 3 is a flowchart of a method of automatically activating windshield wipers of a system in Figure 1

[0129] Figure 4 ​​is an automatic activation of a windshield wiper of a system according to another example of the disclosure Figure 1 a flowchart of a method of a windshield wiper of a system according to another example of the disclosure. DETAILED DESCRIPTION

[0130] The following description is merely exemplary in nature and is not intended to limit the disclosure, application or uses.

[0131] Embodiments and examples of the disclosure provide systems and methods for automatic activation of a windshield wiper of a vehicle. The embodiments and examples are more efficient, more accurate and more cost effective systems and methods. Implementation of these embodiments and examples generally does not require new hardware on most existing vehicles and does not require any additional hardware on new vehicles.

[0132] Figure 1 A system 10 for automatic activation of a windshield wiper of a vehicle 12 is shown according to one embodiment of the disclosure, the vehicle including a front windshield 14 having a front wiper 16 and a rear windshield 18 having a rear wiper 20. As shown, the system 10 includes a cloud server 22 disposed remotely from the vehicle 12. The cloud server 22 is arranged to provide raw information 24 for condition assessment and processed information 25. In one embodiment, the raw information 24 includes vehicle interior and exterior temperature 26, local weather information 28, geographic climate 30, global positioning system (GPS) information 32 of the vehicle for condition assessment. The processed information 25 includes vehicle to vehicle (V2V) and / or vehicle to infrastructure (V2X) information 42 and detected road condition information 44 for system arbitration and decision making, as described below. Other available information can be included without departing from the spirit or scope of the disclosure. Figure 2

[0133] The system 10 also includes at least one sensor, here a front sensor 34 and a rear sensor 36, disposed in the vehicle 12 and arranged to capture a front image of the front windshield 14, a rear image of the rear windshield 18 and an ambient image of the environment. In addition, the system 10 includes a temperature sensor 38 disposed on the vehicle for detecting exterior and interior temperatures. As shown, the front sensor 34 is preferably disposed adjacent to the front windshield 14 of the vehicle 12 for front windshield detection and road detection. In addition, the rear sensor 36 is preferably adjacent to the rear windshield 18 of the vehicle 12 for rear windshield detection and road detection. Each of the front sensor 34 and the rear sensor 36 can be any suitable sensor having a camera arranged to capture still and / or motion images without departing from the spirit or scope of the disclosure. Figure 1

[0134] ​​Further, the system 10 includes an electronic control unit (ECU) 40 disposed in the vehicle 12. As shown, the ECU 40 is in communication with the cloud server 22 as well as the front sensor 34 and the rear sensor 36. As discussed in greater detail below, the ECU 40 sends and receives signals to and from each of the sensors and the cloud server 22. It should be appreciated that the ECU 40 includes modules and algorithms to assist in controlling the system.

[0135] In this embodiment, the ECU 40 includes a condition assessment module that is arranged to assess at least one windshield classification of a road condition based on the raw information 24. As will be described in greater detail below, the condition assessment module pre-assesses the windshield condition and determines what type of condition to detect. For example, some types of conditions can include wet, dry, rain, snow, sleet, ice, hail, or any other suitable condition without departing from the spirit or scope of the present disclosure. Further, based on the available raw information 24 (e.g., vehicle interior and vehicle exterior temperature 26, local weather information 28, geographic climate 30, GPS information), the list of possible conditions can be narrowed. Thus, based on the geographic climate 30 and the local weather of the vehicle, the condition assessment module can determine a set of rules (e.g., “if…then…” rules) for narrowing the list of possible conditions to assess, detect, or classify. As a result, some windshield conditions can be isolated, thereby saving computation and improving accuracy.

[0136] As an example, if the exterior temperature is greater than a temperature Tl, only the possibility of rain and non-rain conditions are assessed, so sleet, ice, and snow conditions will not be detected. If the exterior temperature is between temperatures Tl and T2, only the possibility of sleet and rain conditions are assessed, and ice / snow conditions will not be detected. If the exterior temperature is below temperature Tl, all possible conditions are assessed and detected. It should be appreciated that any other rules can be set for any other available raw information 24, such as the geographic climate 30, without departing from the spirit or scope of the present disclosure. By isolating windshield conditions or narrowing the list of possible conditions, the computational workload is reduced, resulting in enhanced system efficiency and accuracy.

[0137] In this embodiment, the ECU 40 also includes a conditional windshield perception (CWP) module that is arranged to classify the front image, the rear image, and the environmental image into at least one windshield classification defining a first windshield class (e.g., rain). Further, the CWP module is arranged to determine a front windshield perception, a rear windshield perception, and an environmental perception of the first windshield class defining a first combination of detection sources.

[0138] The conditional windshield perception (CWP) module preferably comprises two functions, a front windshield perception function based on the camera and a rear windshield perception function based on the camera. In this example, the CWP module generates three results: first, a front windshield perception; and second, a rear windshield perception, which are two elements in the first combination of detection sources. A third result that the CWP module can produce is an environmental perception based on the raw information 24, as described below. Each function comprises a windshield perception algorithm that learns the pattern of a defined windshield condition and identifies the windshield condition in real-time along with a confidence level (probability). The windshield perception algorithm is typically based on machine learning (ML). Thus, the CWP module preferably comprises a ML model to help classify and determine the first windshield category of the first combination of detection sources for the front windshield perception, the rear windshield perception, and the environmental perception.

[0139] Preferably, the ML model is arranged to first learn the pattern of each defined windshield condition by training on a large set of labeled images (e.g., clear windshield images, rain windshield images, dust windshield images, snow-covered windshield images). Once the training process is complete, the ML model learns the pattern and when a new windshield camera image is captured and fed into the ML model, the ML model can be used to identify the windshield condition. The ML model classifies the windshield condition into one of the defined categories (e.g., clear, dust, rain, snow).

[0140] In one example, the windshield perception algorithm comprises a regular image analysis and ML combined approach. In this example, regular image analysis techniques are used to extract salient image features that can represent the pattern of the windshield condition and a ML method (e.g., support vector machine (SVM)) is used to classify the windshield condition.

[0141] In another example, the windshield perception algorithm comprises a convolutional neural network (CNN) based approach. The CNN is used to automatically extract salient image features and classify the windshield condition by the machine. It should be appreciated that other methods can be used to classify the windshield condition without departing from the spirit or scope of the present disclosure.

[0142] In this embodiment, the first combination of detection sources comprises an environmental perception of the first windshield category (as described above). Preferably, the environmental perception is determined by the windshield perception algorithm based on the raw information 24, thereby defining the third function / result produced by the CWP module.

[0143] Further, the CWP module is arranged to estimate a first front intensity of the first windshield class. Further, the CWP module is arranged to estimate a first rear intensity of the first windshield class. Each of the first front intensity and the first rear intensity can be determined by the algorithmic model based on the rate of descent within the region of the windshield.

[0144] Further, the ECU 40 also includes a weighted voting module arranged to fuse the first combination of front windshield perception, rear windshield perception, and environmental perception of the detection sources. The fusion of the front windshield perception, the rear windshield perception, and the environmental perception defines a first front probability of the first windshield class.

[0145] In one example of the fusion to define the first front probability, the weighted voting module includes a first equation as follows:

[0146]

[0147] where P(X)_front_fusion is the first front probability of the first windshield class (X), P(X)_front is the front windshield perception, P(X)_rear is the rear windshield perception, P(X)_road is the environmental perception, and wl, w2, and w3 are the weights of the detection sources of the front windshield perception, the rear windshield perception, and the environmental perception, respectively. The magnitudes of wl, w2, and w3 are calibrated and can vary based on the environmental perception produced by the CWP module.

[0148] For example, the first windshield class (rain) includes a front windshield perception of 0.92, a rear windshield perception of 0.9, and an environmental perception of 0.91, and

[0149] If P(X)_road > 85%, then wl= 1, w2 = 0, w3 = 0 (thereby emphasizing the front windshield perception);

[0150] If then wl= 0.9, w2 = 0.05, w3 = 0.05 (thereby scaling down the confidence or probability of the front windshield perception); and

[0151] If P(X)_road < 60%, then wl= 0.8, w2 = 0, w3 = 0.2 (thereby further scaling down the confidence of the front windshield perception).

[0152] The first front probability can then be calculated as follows when P(X)_road > 85%:

[0153]

[0154] ​Further, the weighted voting module is arranged to fuse the front windshield perception, the rear windshield perception, and the environmental perception of the first combination of detection sources. The fusion of the front windshield perception, the rear windshield perception, and the environmental perception defines a first rear probability of the first windshield class.

[0155] In one example of the fusion to define the first rear probability, the weighted voting module includes a second equation as follows:

[0156] ,

[0157] where P(X)_rear_fusion is the first rear probability of the first windshield class (X), P(X)_front is the front windshield perception, P(X)_rear is the rear windshield perception, P(X)_road is the environmental perception, and wl, w2, and w3 are weights of the detection sources of the front windshield perception, the rear windshield perception, and the environmental perception, respectively. As noted above, the magnitudes of wl, w2, and w3 are calibrated and can vary based on the environmental perception produced by the CWP module or from the processed information 25. As further shown, the first rear probability can be calculated in the same manner as the first front probability is shown to be calculated.

[0158] The processed information 25 includes V2V and / or V2X information 42 and road condition information 44. It should be appreciated that the V2V information includes environmental information received and transmitted between vehicles as known in the art. Further, it can be appreciated that the V2X information includes environmental information received and transmitted between infrastructure and vehicles as known in the art. Further, it should be appreciated that the road condition information 44 includes road condition information from network-based sources (e.g., the Internet).

[0159] Further, the ECU 40 is arranged to determine a first front intensity and a first rear intensity of the first windshield class. The first front intensity is based on a drop rate within the area of the front windshield 14. Further, the first rear intensity is based on a drop rate within the area of the rear windshield 18. In one example, the ECU 40 is arranged to determine each of the first front intensity and the first rear intensity by applying a linear fit equation as follows:

[0160]

[0161] where In(drop) is the normalized intensity, V_drop is the drop rate, N_drop is the number of drops per area of the windshield, and kl and k2 are calibration coefficients.

[0162] Further, the ECU 40 includes an activation module arranged to activate the front wipers 16, 16 when the first front probability is greater than a front threshold. Alternatively, the activation module is arranged to activate the front wipers 16, 16 when the first front probability is greatest in magnitude of at least one windshield class. Moreover, the activation module is arranged to activate the rear wipers 20 when the first rear probability is greater than a rear threshold. Alternatively, the activation module is arranged to activate the rear wipers 20 when the first rear probability is greatest in magnitude of at least one class.

[0163] Further, the ECU 40 is arranged to adjust the speed of the front wipers 16, 16 based on the first front intensity when the front wipers 16, 16 are activated. Additionally, the ECU 40 is arranged to adjust the speed of the rear wipers 20 based on the first rear intensity when the rear wipers 20 are activated. The speed of the front wipers 16, 16 and the rear wipers 20 can be adjusted based on the respective intensity in any suitable manner, such as a rule-based algorithm, without departing from the scope or spirit of the present disclosure.

[0164] Figure 2 A flowchart of an overall method 110 of windshield wiper activation of the system 10 in accordance with one example of the present disclosure is shown. Figure 1 As shown, the method 110 includes a step 112 of providing existing information (vehicle interior and exterior temperatures 26, local weather information 28, geographic climate 30, global positioning system (GPS) information 32 of the vehicle for condition assessment); a step 114 of condition assessment; a step 116 of conditional windshield perception (CWP); and a step 118 of arbitration and decision making. As will be described in greater detail below, the step 114 of condition assessment pre-assesses the windshield condition and determines what type of condition to detect. As a result, some conditions can be isolated to reduce computational workload, increase efficiency, and improve accuracy.

[0165] Further, the step 116 of conditional windshield perception can focus on visual images of the front windshield 14 and the rear windshield 18 of the vehicle 12 based on the output of the step 114 of condition assessment (described above). Machine learning models can be used to determine the probability of the windshield condition and class. Moreover, the step 118 of arbitration and decision making determines whether the front wipers 16 and / or the rear wipers 20 can be activated based on the output of the CWP step and the result of the weighted voting module. The speed of the wipers can also be determined and adjusted based on the probability of the windshield condition.

[0166] Figure 3 A flowchart of an overall method 110 of windshield wiper activation of the system 10 in accordance with one example of the present disclosure is shown. Figure 1A flowchart of a method 210 of automatically activating a windshield wiper implemented by the system 10. In this example, the method 210 automatically activates a windshield wiper of a vehicle 12 having a front windshield 14 with a front wiper 16 and a rear windshield 18 with a rear wiper 20. As shown, the method 210 includes a step 112 of providing raw information 24 for a condition assessment. As described above, the cloud server 22 provides the raw information 24, which includes a vehicle interior and exterior temperature 26, local weather information 28, a geographic climate 30, global positioning system (GPS) information 32 of the vehicle for a condition assessment.

[0167] As shown, the method 210 also includes a step 214 of assessing at least one windshield classification of a road condition with a condition assessment module of the ECU 40 based on the raw information 24. The condition assessment module pre-assesses the windshield condition and determines what type of condition to detect.

[0168] Further, the method 210 also includes a step 216 of capturing a front image of the front windshield 14, a rear image of the rear windshield 18, and an ambient image of the environment. At least one sensor, here the front sensor 34 and the rear sensor 36 Figure 1 ), preferably captures the images and correspondingly sends image signals to the ECU 40 for processing.

[0169] With reference to Figure 2 , the method 210 then includes a step 218 of classifying the front image, the rear image, and the ambient image into at least one windshield classification with the ML by the ECU 40, as described above. The method 210 also includes a step 220 of determining a front windshield perception, a rear windshield perception, and an ambient perception of the first windshield class with the CWP module of the ECU 40 Figure 1 ) to define a first combination of detection sources with the ML, as described above. As described above, the CWP module of the ECU 40 Figure 1 ) is arranged to determine the front windshield perception, the rear windshield perception, and the ambient perception of the first windshield class via the ML, thereby defining the first combination of detection sources.

[0170] In one example, the method 210 can include estimating a first front intensity of the first windshield class with the conditional windshield perception module. As described above, the CWP module is arranged to estimate the first front intensity of the first windshield class. The first front intensity can be estimated by an algorithmic model based on a rate of precipitation within a region of the windshield.

[0171] In this aspect of the disclosure, the method 210 includes a step 222 of fusing the first combination of the front windshield perception, the rear windshield perception, and the environmental perception of the detection sources. As previously described, the weighted voting module is arranged to fuse the front windshield perception, the rear windshield perception, and the environmental perception to define the first front probability of the first windshield class.

[0172] As described above, one example of fusing to define the first front probability is achieved through a weighted voting module. As discussed, the weighted voting module includes a first equation as follows:

[0173] where P(X)_front_fusion is the first front probability of the first windshield class (X), P(X)_front is the front windshield perception, P(X)_rear is the rear windshield perception, P(X)_road is the environmental perception, and w1, w2, and w3 are the weights of the detection sources for the front windshield perception, the rear windshield perception, and the environmental perception, respectively. The magnitudes of w1, w2, and w3 are calibrated and can vary based on the environmental perception produced by the CWP module.

[0174] For example, the first windshield class (rain) includes a front windshield perception of 0.92, a rear windshield perception of 0.9, and an environmental perception of 0.91, and

[0175] If P(X)_road > 85%, then w1= 1, w2 = 0, w3 = 0 (thereby emphasizing the front windshield perception);

[0176] If then w1= 0.9, w2 = 0.05, w3 = 0.05 (thereby scaling down the confidence or probability of the front windshield perception); and

[0177] If P(X)_road < 60%, then w1= 0.8, w2 = 0, w3 = 0.2 (thereby further scaling down the confidence of the front windshield perception).

[0178] Accordingly, when P(X)_road > 85%, the first front probability can be calculated as follows:

[0179]

[0180] In one example, the method 210 further includes utilizing the ECU 40 Figure 1) to fuse the first combination of front windshield perception, rear windshield perception, and environmental perception of the detection sources to define a first rear probability of the first windshield class. Further, the weighted voting module is arranged to fuse the first combination of front windshield perception, rear windshield perception, and environmental perception of the detection sources. The fusion of the front windshield perception, rear windshield perception, and environmental perception defines the first rear probability of the first windshield class.

[0181] In one example of the fusion to define the first rear probability, the weighted voting module includes a second equation as follows:

[0182] ,

[0183] where P(X)_rear_fusion is the first rear probability of the first windshield class (X), P(X)_front is the front windshield perception, P(X)_rear is the rear windshield perception, P(X)_road is the environmental perception, and w1, w2, and w3 are the weights of the detection sources for the front windshield perception, rear windshield perception, and environmental perception, respectively. As described above, the magnitudes of w1, w2, and w3 are calibrated and can vary based on the environmental perception produced by the CWP module.

[0184] Further, the method 210 can include determining a first front intensity and a first rear intensity of the first windshield class. The first front intensity and the first rear intensity are based on a drop rate within the area of the front windshield 14. Further, the ECU 40 is arranged to determine the first front intensity and the first rear intensity of the first windshield class. The first front intensity is based on a drop rate within the area of the front windshield 14. Further, the first rear intensity is based on a drop rate within the area of the rear windshield 18. In one example, the ECU 40 is arranged to determine each of the first front intensity and the first rear intensity by applying a linear fit equation as follows:

[0185]

[0186] where In(drop) is the normalized intensity, V_drop is the drop rate, N_drop is the number of drops per area, and k1 and k2 are calibration coefficients.

[0187] The method 210 then comprises a step 224 of activating the front wiper 16 when one of the first front probability is greater than the front threshold and the first front probability is maximum in magnitude of at least one classification. The method 210 can then comprise activating the rear wiper 20 when one of the first rear probability is greater than the rear threshold and the first rear probability is maximum in magnitude of at least one classification. As mentioned in the description of the system, the ECU 40 comprises an activation module arranged to activate the front wiper 16 when the first front probability is greater than the front threshold. Alternatively, the activation module is arranged to activate the front wiper 16 when the first front probability is maximum in magnitude of at least one windshield classification. Furthermore, the activation module is arranged to activate the rear wiper 20 when the first rear probability is greater than the rear threshold. Alternatively, the activation module is arranged to activate the rear wiper 20 when the first rear probability is maximum in magnitude of at least one classification.

[0188] Furthermore, the method 210 can also comprise adjusting the speed of the front wiper 16 based on the first front intensity when the front wiper 16 is activated and adjusting the speed of the rear wiper 20 based on the first rear intensity when the rear wiper 20 is activated. As previously mentioned, the ECU 40 is arranged to adjust the speed of the front wiper 16 based on the first front intensity when the front wiper 16 is activated. In addition, the ECU 40 is arranged to adjust the speed of the rear wiper 20 based on the first rear intensity when the rear wiper 20 is activated. The speed of the front wiper 16 and the rear wiper 20 can be adjusted by any suitable means, for example a rule-based algorithm, without departing from the scope or spirit of the invention.

[0189] According to another aspect of the present disclosure, Figure 4 A flowchart depicting a method 310 of automatically activating a windshield wiper system of a vehicle 12 having a front windshield 14 with a front wiper 16 and a rear windshield 18 with a rear wiper 20 is depicted. Figure 4 The method 310 is implemented by Figure 1 the system 10.

[0190] As shown, the method 310 comprises a step 312 of providing raw information 24 for the assessment of the conditions and comprises a step 314 of assessing at least one windshield classification of the road conditions with a condition assessment module based on the raw information 24. The method 310 further comprises a step 316 of capturing a front image of the front windshield 14, a rear image of the rear windshield 18 and an environment image of the environment. Furthermore, the method 310 also comprises a step 318 of classifying the front image, the rear image and the environment image into at least one windshield classification with a condition windshield perception module defining a first windshield classification.

[0191] In this example, the method 310 further includes a step 320 of determining, with the conditional windshield perception module, a front windshield perception, a rear windshield perception, and a road perception of the first windshield class to define a first combination of detection sources. The method 310 further includes a step 322 of fusing, with the weighted voting module, the first combination of detection sources with respect to the front windshield perception, the rear windshield perception, and the road perception to define a first front probability of the first windshield class.

[0192] In this example, the weighted voting module includes a first equation as follows:

[0193] ,

[0194] where P(X)_front_fusion is the first front probability of the first windshield class (X), P(X)_front is the front windshield perception, P(X)_rear is the rear windshield perception, P(X)_road is the road perception, and w1, w2, and w3 are the weights of the detection sources for the front windshield perception, the rear windshield perception, and the road perception, respectively.

[0195] The method 310 further includes a step 324 of fusing, with the weighted voting module, the front windshield perception, the rear windshield perception, and the road perception of the first combination of detection sources to define a first rear probability of the first windshield class. The weighted voting module includes a second equation as follows:

[0196] ,

[0197] where P(X)_rear_fusion is the first rear probability of the first windshield class (X).

[0198] Further, the method 310 includes a step 326 of activating the front wiper 16 when one of a first front probability is greater than a front threshold and a first front probability is greatest in magnitude of at least one classification. Further, the method 310 includes a step 328 of activating the rear wiper 20 when one of a first rear probability is greater than a rear threshold and a first rear probability is greatest in magnitude of at least one classification.

[0199] The description of the present disclosure is merely exemplary in nature and variations that do not depart from the gist of the present disclosure are intended to be within the scope of the present disclosure. Such variations are not to be regarded as a departure from the spirit and scope of the present disclosure.

Claims

1. A method of automatically activating a windshield wiper system of a vehicle having a front windshield with a front wiper and a rear windshield with a rear wiper, the method comprising: providing raw information for condition assessment; assessing at least one windshield classification of road conditions with a condition assessment module based on the raw information; capturing a front image of the front windshield, a rear image of the rear windshield, and an ambient image of an environment; classifying the front image, the rear image, and the ambient image into the at least one windshield classification with a conditional windshield perception module defining a first windshield class; determining a front windshield perception, a rear windshield perception, and an ambient perception of the first windshield class with a conditional windshield perception module to define a first combination of detection sources; fusing the front windshield perception, the rear windshield perception, and the ambient perception of the first combination of detection sources with a weighted voting module to define a first front probability of the first windshield class; and activating the front wiper when one of a first front probability greater than a front threshold and the first front probability being greatest in magnitude of at least one classification occurs, wherein the weighted voting module comprises a first equation as follows: P(X)_front _ fusion = w1 * P(X)_front + w2 * P(X)_rear + w3 * P(X)_road, where P(X)_front_fusion is a first front probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an ambient perception, and wl, w2, and w3 are weights of detection sources of the front windshield perception, the rear windshield perception, and the ambient perception, respectively.

2. The method of claim 1, wherein, the raw information comprises a vehicle interior temperature, a vehicle exterior temperature, local weather information, a geographic climate, global positioning system (GPS) information of the vehicle for condition assessment.

3. The method of claim 1, wherein, the at least one windshield classification is a road condition enabling activation of one of the front wiper and the rear wiper of the vehicle.

4. The method of claim 1, further comprising: estimating a first front intensity of the first windshield class with a conditional windshield perception module; determining the first front intensity of the first windshield class, the first front intensity being based on a drop rate within a region of the front windshield; and adjusting a speed of the front wiper based on the first front intensity when the front wiper is activated.

5. The method of claim 4, wherein, determining the first front intensity comprises applying a linear fitting equation as follows: In(drop) = k1*V _ drop + k2*N _ drop where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area, and kl and k2 are calibration coefficients.

6. The method of claim 1, further comprising: fusing the front windshield perception, the rear windshield perception, and the ambient perception of the first combination of detection sources with a weighted voting module to define a first rear probability of the first windshield class; and activating the rear wiper when one of a first rear probability greater than a rear threshold and the first rear probability being greatest in magnitude of at least one classification occurs.

7. The method of claim 6, wherein, the weighted voting module comprises a second equation as follows: P(X)_rear _ fusion = w1 * P(X)_front + w2 * P(X)_rear + w3 * P(X)_road, where P(X)_rear_fusion is a first rear probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environment perception, and wl, w2, and w3 are weights of detection sources of the front windshield perception, the rear windshield perception, and the environment perception, respectively.

8. The method of claim 6, further comprising: estimating a first rear intensity of a first windshield class with a conditional windshield perception module; determining the first rear intensity of the first windshield class, the first rear intensity being based on a drop rate within a region of the rear windshield; and adjusting a speed of the rear wiper based on the first rear intensity when the rear wiper is activated.

9. The method of claim 8, wherein, Determining the first rear intensity includes applying a linear fitting equation as follows: In(drop) = k1*V _ drop + k2*N _ drop where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area of the rear windshield, and kl and k2 are calibration coefficients.

10. A system for automatically activating windshield wipers of a vehicle, the vehicle including a front windshield having a front wiper and a rear windshield having a rear wiper, the system comprising: a cloud server disposed remotely from the vehicle, the cloud server arranged to provide raw information for a condition assessment; at least one sensor disposed on the vehicle and arranged to capture a front image of the front windshield, a rear image of the rear windshield, and an environment image of an environment; and an electronic control unit (ECU) disposed in the vehicle, the ECU in communication with the cloud server and the at least one sensor, the ECU including: a condition assessment module arranged to assess at least one windshield classification of a road condition based on the raw information; a conditional windshield perception module arranged to classify the front image, the rear image, and the environment image into the at least one windshield classification defining a first windshield class, the conditional windshield perception module arranged to determine a front windshield perception, a rear windshield perception, and an environment perception of the first windshield class to define a first combination of detection sources; a weighted voting module arranged to fuse the front windshield perception, the rear windshield perception, and the environment perception of the first combination of detection sources, thereby defining a first front probability of the first windshield class; and an activation module arranged to activate the front wiper when one of a first front probability is greater than a front threshold and the first front probability is maximum in magnitude of the at least one windshield classification, wherein the weighted voting module includes a first equation as follows: P(X)_front _ fusion = w1 * P(X)_front + w2 * P(X)_rear + w3 * P(X)_road, where P(X)_front_fusion is a first front probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environmental perception, and wl, w2, and w3 are weights of detection sources for the front windshield perception, the rear windshield perception, and the environmental perception, respectively.

11. The system of claim 10, wherein, The raw information includes a vehicle interior temperature, a vehicle exterior temperature, local weather information, a geographic climate, global positioning system (GPS) information of the vehicle for condition assessment.

12. The system of claim 10, wherein, The conditional windshield perception module is arranged to estimate a first front intensity of the first windshield class, wherein the ECU is arranged to determine the first front intensity of the first windshield class, the first front intensity being based on a drop rate within an area of the front windshield, and wherein the ECU is arranged to adjust a speed of the front wiper based on the first front intensity when the front wiper is activated.

13. The system of claim 12, wherein, The ECU is arranged to determine the first front intensity by applying a linear fitting equation as follows: In(drop) = k1*V _ drop + k2*N _ drop where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area, and kl and k2 are calibration coefficients.

14. The system of claim 10, wherein, The weighted voting module is arranged to fuse the front windshield perception, the rear windshield perception, and the environmental perception of the first combination of detection sources to define a first rear probability of the first windshield class, and wherein the weighted voting module is arranged to activate the rear wiper when one of a first rear probability greater than a rear threshold and the first rear probability being greatest in magnitude of at least one classification occurs.

15. The system of claim 14, wherein, The weighted voting module includes a second equation as follows: P(X)_rear _ fusion = w1 * P(X)_front + w2 * P(X)_rear + w3 * P(X)_road. where P(X)_rear_fusion is a first rear probability of a first windshield class (X), P(X)_front is a front windshield perception, P(X)_rear is a rear windshield perception, P(X)_road is an environmental perception, and wl, w2, and w3 are weights of detection sources for the front windshield perception, the rear windshield perception, and the environmental perception, respectively.

16. The system of claim 14, wherein, The conditional windshield perception module is arranged to estimate a first rear intensity of the first windshield class, wherein the ECU is arranged to determine the first rear intensity of the first windshield class, the first rear intensity being based on a drop rate within an area of the rear windshield, and wherein the ECU is arranged to adjust a speed of the rear wiper based on the first rear intensity when the rear wiper is activated.

17. The system of claim 16, wherein, The ECU is arranged to determine the first rear intensity by applying a linear fitting equation as follows: In(drop) = k1*V _ drop + k2*N _ drop where In(drop) is a normalized intensity, V_drop is a drop rate, N_drop is a number of drops per area of the rear windshield, and kl and k2 are calibration coefficients.

18. The system of claim 10, wherein, The cloud server is arranged to provide the processed information for system arbitration.

Images