Method and device for detecting damage to a window of a vehicle

Abstract

The present method and the associated device and the use thereof in a system and a vehicle are therefore intended to provide a possibility by way of which damage to a window of a vehicle, for example a windscreen, can be detected at an early stage. To this end, a time curve of a noise is first captured by means of at least one microphone. The impact of an object is detected on the basis of the signal of the time curve, whereupon a damage classification is carried out. If a relevant crack formation, or a damage class which makes repair necessary in general, is identified, damage information is generated accordingly.

Description

[0001] Description

[0002] title

[0003] Method and device for detecting damage to a vehicle window

[0004] The invention relates to a method and a device as well as their application in a corresponding system or vehicle with which damage to a window of a vehicle can be detected.

[0005] State of the art

[0006] Stones kicked up by vehicles or objects like sheets of ice falling from them can easily damage a vehicle's windshield. Most of these impacts cause little to no damage to the glass. However, if the stone or object strikes the windshield at a certain angle and with sufficient force, it can easily cause chips or cracks. If these chips or cracks are small, the actual damage may only become apparent over time as they enlarge.

[0007] Such a delayed visibility of the damage is a particular problem for fleet vehicles such as rental cars, since it is not the current user but a later user who has to report the damage and, if necessary, have it repaired via the insurance.

[0008] German patent application DE 10 2016 010 705 A1 discloses a method for detecting stone impacts using a structure-borne sound sensor. The present invention aims to provide a means of detecting damage to a windshield caused by impacting objects at an early stage, in order to minimize further damage and thus costs.

[0009] Disclosure of the invention

[0010] The present method, along with the associated device and its application in a system and a vehicle, aims to provide a means of detecting damage to a vehicle window, such as a windshield, at an early stage. To this end, a sound, particularly one emanating from the windshield, is recorded over time using at least one microphone. Based on this signal, the impact of an object is detected. Additionally, a velocity parameter is recorded, specifically the vehicle speed and / or wind speed. Using the first microphone signal and the velocity parameter, a damage classification is then performed.If relevant cracking or a damage class is detected that requires repair, damage information is generated accordingly.

[0011] The invention is based on the use of sensors already present in or available in the vehicle. For example, a microphone from a hands-free device or a smartphone can be used to capture the first microphone signal. The microphone of the hands-free device is typically mounted near the windshield or front window so that it can detect sounds in the vicinity of a potential impacting object. Further damage classification allows the consequences of an impact to be differentiated. Furthermore, an object impact can be documented so that subsequent users or a fleet operator can be informed about the incident.Furthermore, a second microphone signal can be recorded, representing wind noise or, more generally, wind speed at the vehicle and / or windshield, particularly while the vehicle is in motion. The damage classification can then be further refined based on this second microphone signal or the derived (wind) speed value. By considering wind speed and / or headwind while the vehicle is in motion, the impact energy and / or windshield damage can be determined more precisely.

[0012] The first microphone, previously used for the first microphone signal, can be used for the second microphone signal. Optionally, a second microphone can be used, for example, mounted on the vehicle body to capture external sounds. Alternatively, a pressure sensor can be used to capture the second microphone signal. Suitable pressure sensors can be integrated into the bumper, for example, and are used particularly for pedestrian protection and / or as part of a parking assistance system.

[0013] Additionally or optionally, consumption information related to the vehicle's movement can be recorded. This includes recording the vehicle's fuel consumption and / or electricity consumption during travel and using changes in these values ​​to calculate the headwind acting on the vehicle. The headwind speed determined or derived in this way can then be used to classify the damage.

[0014] In a further development of the invention, it is provided that environmental data of the vehicle is recorded, for example, to prove the origin of the object that struck the windshield. For this purpose, at least some of the environmental data is added to the damage information. This environmental data can include, for example, camera information (images or video sequences), radar information, or location information. Using this environmental data, the vehicle ahead, which is presumably responsible for the object, can be identified. This identification, for example, a license plate number or other identifying characteristic, can also be added to the damage information.

[0015] Using at least two primary microphones, a differential analysis of the two microphone signals can be performed. This differential analysis, or the consideration of the primary microphone signals from at least two different primary microphones, can be used to detect, specify, or classify the impact of the object. Localizing the point of impact is also possible when using multiple microphone signals.

[0016] In one embodiment of the invention, an energy value representing the energy input into the disk can also be derived from the first and / or second microphone signals. This energy value can then be used to classify the damage and detect the crack in the disk. Optionally, an impact can also be identified that, while not yet causing a crack, has left a sufficiently deep chip that could potentially lead to crack formation.

[0017] In a further development of the invention, the first microphone signal can be integrated for damage classification and / or crack detection. In this context, exceeding a threshold value by the integration value can, in particular, indicate crack formation.

[0018] In a particular embodiment of the invention, it is provided that the first microphone signal or the temporal progression of the noise recorded by the first microphone is only detected when the speed has exceeded a speed threshold.

[0019] Since the vibration characteristics of the disc change when a crack forms or propagates, the initial microphone signal can also be used to track its propagation. Thus, after detecting the crack formation, the propagation of the crack can be recorded by the initial microphone (at least one microphone) through further sound recording. The analysis of the sounds or the signal data from the sounds, as well as the damage classification, can be performed in a control unit within the vehicle. Alternatively, the sounds or the signal data from the sounds can be sent to a central server outside the vehicle, which then analyzes them. The results of this analysis can then be transmitted back to the vehicle or a control unit within the vehicle to generate further damage information.Alternatively, the damage classification and damage information can also be generated on the central server, so that only this damage information is sent to the vehicle.

[0020] It is also possible to take individual disc parameters into account when evaluating the noise or the noise signal data. This can be done, for example, by considering specific disc parameters during integration or by defining the comparison thresholds.

[0021] The damage information can also be transmitted to a fleet operator, an insurance company, and / or a repair shop. For example, the fleet operator or insurance company can automatically initiate a repair order based on the damage information and notify the vehicle's driver. Additional parameters such as proximity or cost can also be considered when selecting a suitable repair shop.

[0022] Brief description of the drawings

[0023] The diagrams in Figures 1a and 1b show possible signal waveforms of noises when an object strikes a disc. Figure 2 illustrates a possible device for implementing the invention. Figure 3 shows an embodiment of the invention within a process as a flowchart. Figure 4 illustrates an application for detecting damage to the disc. Embodiments of the invention

[0024] When driving on highways and country roads, it is not uncommon for vehicles ahead to kick up stones that then strike the windshield. The same applies to trucks where bulk goods come loose from the cargo bed. Another source of danger for objects hitting the windshield in winter is ice sheets breaking off the roofs of vehicles. Most such impacts cause no or only very minor damage to the windshield and are likely to go undetected without a closer inspection. Even if small, localized damage is initially unproblematic, it, like cracks, can worsen due to vibrations and wind resistance while driving, eventually requiring a windshield replacement. In order to charge the responsible party for the cost of such a windshield replacement, the incident must be documented, ideally with proof of the event.Such data collection is also beneficial for fleet operators or car rental companies, ensuring that subsequent users of a vehicle are not held responsible for damage caused by a previous user. The following descriptions outline a method, device, system, or vehicle that enables this type of data collection and documentation of damage incidents.

[0025] When an object strikes a window, for example a stone chip, at least some of the impact energy is converted into an (elastic) deformation of the window, depending on the angle of impact, which creates a sound. Another part of the impact energy can lead to damage to the window, for example in the form of chips, holes or cracks.

[0026] Figure 1a shows an example of the sound level profile following a stone impact without cracking. In this case, the stone is elastically scattered by the windshield, so that the impact energy is essentially returned to the deflection of the stone, thus ending the energy exchange between the stone and the windshield at time t1. As can be seen from the diagram, there is essentially a symmetrical energy exchange with the windshield, which leads to a vibration. However, this vibration dissipates very quickly, so that the windshield returns to rest after only a few oscillations.

[0027] In contrast, the diagram in Figure 1b shows an impact of a stone with crack formation at time t2. Due to the crack formation, the disk loses the ability to elastically return the applied energy to the stone, resulting in further elastic deformation of the disk after time t2. Only at the later time t3 has the energy been distributed throughout the disk, so that the vibrations of the disk are significantly reduced.It should be noted that similar vibration behavior results from a stone impact with a resulting hole or chip, although the behavior after time t2 is less pronounced. However, the decisive factor is the energy input when the pane is damaged, which can be demonstrated as a longer and more intense vibration of the pane or as sound compared to elastic scattering. This difference is exploited by the present invention to detect a crack or, more generally, damage to the pane due to the impact of an object or a stone.

[0028] The present invention can be implemented in a device according to Figure 2. For this purpose, a processing unit 100 or an evaluation unit is provided, which performs a corresponding method according to the invention for detecting damage to a vehicle windshield and can, for example, be part of an airbag control system. This processing unit 100 can have a memory 110 in which the detected sounds or microphone signals from a first microphone 120 and / or a second microphone 130 are stored, for example, in the form of a ring buffer. Additionally, further information can be stored in the memory 110, for example, individual physical parameters that can be assigned to the windshield.Based on the captured first and / or second microphone signals, processing unit 100 performs a damage classification to detect a crack or other damage to the windshield. If a crack or sufficient damage is detected to require action, corresponding damage information is generated. This information can be forwarded within the vehicle to a corresponding system 160 for further processing or consideration, for example, to an external fault memory. It is also possible that the vehicle's operation will be restricted due to the detected windshield damage. For instance, a speed limit could be imposed on a bumpy road to slow the propagation of a crack.Alternatively or additionally, the driver can be informed via an acoustic or visual indicator (170) that the windshield has been damaged or requires repair / replacement. It is also possible for the information about the damaged windshield to be sent to a fleet operator, an insurance company, or directly to a repair shop (180). For the fleet operator, this has the advantage of allowing them to assign the damage to a specific user, estimate when a windshield replacement will be necessary, and / or obtain quotes for windshield replacements in advance. The insurance company can use the damage information, and possibly other information about the responsible party, to attempt to recover the costs of the damage from them. Furthermore, by requesting potential cost estimates from repair shops, the insurance company can automatically negotiate the most favorable price before the vehicle user even visits the repair shop.Furthermore, the insurance company can, if necessary, direct the vehicle user to a preferred repair shop. Transmitting the damage information to the (potentially commissioned) repair shop allows them to order a replacement windshield in advance if one is not in stock. In addition, the recorded microphone signals and / or the damage information can be sent to a cloud application (190). There, the microphone signals can be analyzed and, if necessary, combined with further information, such as individual windshield parameters or risks of damage propagation, to create the damage information, which is then sent back to the processing unit (100) for further processing.Alternatively, it is also possible to send the damage information as such to the cloud application 190, which then processes and / or forwards this information, for example to the fleet operator, the insurance company and / or a workshop.

[0029] The processing unit 100 can optionally store the captured microphone signals in a ring buffer and monitor the noise level. If the noise level of the microphone signal exceeds a predetermined threshold, a copy of the captured microphone signals and vehicle data surrounding this detected exceedance can be stored in memory 110 for further processing in damage detection and classification.

[0030] To perform damage classification or to generate damage information, the processing unit 100 can acquire at least one additional microphone signal from another first or second microphone 130. This at least second microphone signal allows, in the case of an additional first microphone signal, the precise location of the damage or crack on the pane to be determined by differential analysis. Furthermore, an additional microphone signal can be used to confirm the sound heard when the damage or crack occurred. Advantageously, the at least two first microphones are spatially separated from each other, so that they can, if necessary, cover different areas of the pane for monitoring.

[0031] By using an additional second microphone 130, the wind speed at the vehicle and / or the windshield can be determined from the second microphone signals and the sounds thus captured. The wind speed thus captured, as a measure of velocity, can provide information about the angle at which the object impacts, the vehicle's motion relative to the object, the energy input the object generates into the windshield, and / or the strength of the headwind that throws / accelerates the object against the windshield. Damage classification can also be improved based on the wind speed thus captured in relation to the second microphone signals. The underlying mapping of the sounds and / or the strength of the second microphone signal to the wind speeds can be recorded under laboratory and real-world conditions and stored as reference values ​​or tables in memory 100.Optionally and / or alternatively, the second microphone signals can also be sent to a server in the cloud (190) for evaluation and retrieval of the results. In addition to information about the occurrence of damage to the windshield, the damage information can include further details related to the damage. For example, the time and location of the incident can be documented. Furthermore, the responsible party can be recorded and identified using a suitable device (140). For instance, if a stone chip is detected, camera images could be captured and saved along with the damage information to identify a vehicle traveling ahead. Optionally, it is also possible to identify the vehicle directly from the camera data, for example, via its license plate number.In addition to a camera, other sensors could also be used to identify the perpetrator, for example, a radar system with sufficient resolution. However, the device 140 can also be used for any other type of object identification.

[0032] For both initiating the process of detecting damage to the windshield and verifying the findings, additional sensor data from the vehicle can be acquired and considered by integrated sensors. For example, the sensors of an airbag control unit can be used to detect the vehicle's speed, so that the recording of noises with at least one microphone, or the detection of windshield damage, only begins when a minimum speed is exceeded. Furthermore, the vehicle's movement can also be factored into the generation of damage classification or damage information. For instance, vehicle movement can influence the extent of the damage and thus the timing of the replacement.

[0033] The vehicle's sensor system 150 can also be used to record the vehicle's fuel consumption in order to derive a value representing the headwind. For example, a consumption value representing the amount of fuel or electricity consumed over time can be recorded, and a headwind speed value can be derived from its change. This headwind speed value can also be used to determine damage classification. The headwind speed value can be determined, for example, by comparing the current speed with the target speed at a known fuel consumption. Alternatively or optionally, the elevation profile or the route can also be taken into account.

[0034] The method for detecting damage to a vehicle windshield is explained below with reference to the flowchart in Figure 3. After the method is started, for example, depending on a minimum vehicle speed, a sound is detected in step 320 using at least one microphone. This can be a microphone from a hands-free system in the vehicle or an additional microphone, preferably located inside the vehicle and near the rear. Optionally, the microphone of a mobile device, such as a smartphone, can also be used for detection. In step 320, or optionally in a previous step 300, a speed parameter is also detected. This can be the speed of the vehicle and / or the wind acting on the vehicle and / or the windshield.

[0035] In step 340, the system uses the recorded time series of the initial microphone signals to detect when an object, such as a stone, strikes the windshield. Based on this impact detection, a damage classification is performed using the initial microphone signals and the speed of the object. If sufficient damage to the windshield is detected, requiring immediate or later action, damage information is generated in step 370 and forwarded, for example, to the vehicle user. Alternatively, the damage information can be saved for later service purposes. In another application, the damage information can also be sent to a fleet operator, an insurance company, and / or a repair shop. If no damage is detected in step 340, the process can be terminated or restarted.

[0036] To detect the object impact and to derive the

[0037] Damage classification involves evaluating the temporal evolution of the noise generated during impact. As already explained in Figures 1a and 1b, the impact produces a more pronounced, asymmetrical noise in the case of damage than in the case of a primarily elastic collision. Therefore, the temporal evolution of the noise, or the initial microphone signal, is used to evaluate the object impact and, above all, to identify damage and its extent. Since the energy transferred to the disc also differs in both processes, an integration of the noise signal, representing a measure of the energy input into the disc, can also be used to distinguish damage from an elastic collision. A classification of the object impact can also be performed by considering the temporal evolution and / or the energy input.By taking into account the physical parameters of the disc, the damage classification can be further refined.

[0038] In a further embodiment of the invention, the method may first detect the vehicle's speed in step 300 and then compare it with a speed threshold in a subsequent step 310. The damage detection process is only carried out if the vehicle is moving at a speed above the speed threshold. Otherwise, the process is terminated or restarted with step 300.

[0039] As already described for the device according to Figure 2, the method can also capture the noise or microphone signal of at least one additional microphone in step 320. These at least two first microphone signals can be compared with each other in step 340 to perform damage detection and / or damage classification. For example, a difference between the microphone signals can be used to identify the specific location on the windshield where the object struck. Furthermore, the second microphone signal can be used to verify the first microphone signal and / or the impact. It is also conceivable that the second first microphone signal is used to suppress background noise. In a further embodiment, it can be provided that in step 330, a second microphone is used to capture second microphone signals representing the wind noise at the vehicle and / or the windshield.This second microphone signal can be used in step 340 as a velocity parameter to further determine the damage classification. By recording wind noise, the strength or speed of the wind can be determined, and thus the force or energy with which an object is thrown against the windshield can be derived.

[0040] Optionally, in step 340, when determining the damage classification, the headwind speed from the vehicle's consumption data can also be used as an additional speed parameter. For this purpose, in step 330, the temporal profile or change of a consumption parameter representing the fuel or electricity consumption for powering the vehicle can be recorded or imported. By comparing this with the target speed at a known consumption level, the headwind and, in particular, its strength can be determined. Optionally, navigation data, elevation profiles, and / or route data can also be considered to evaluate consumption.

[0041] After the object impact and damage to the windshield are detected in step 340, environmental data from the vehicle can be collected in a subsequent step 340. This data can be used to identify the driving situation and / or a potential vehicle ahead that may have caused the damage. In addition to the time and location coordinates of the incident, an image or video from a camera showing the vehicle ahead or possibly the impacting object can also be captured. It is also conceivable that the license plate of the vehicle ahead could be identified in step 340 based on the camera images. The environmental data or findings collected in this way can be at least partially added to the damage information in step 370 and / or transmitted to external recipients.

[0042] In further optional steps 330 and 360, additional sensor parameters or vehicle operating parameters can be recorded, which can be helpful in refining damage detection and information. For example, step 330 could record vehicle parameters that improve the plausibility of damage detection or support damage classification. Here, for instance, pre-existing damage to the windshield could be detected and taken into account. Furthermore, an uneven road surface could be detected, which would likely accelerate the propagation of a crack, thus necessitating faster action when replacing the windshield. The additional data recorded in step 360, on the other hand, can be used to specify the environmental data from step 350 or to refine the damage information with further details. This could potentially prove that the driver is not at fault for the damage.Optionally, data can also be recorded here that might be relevant for the fleet operator, the insurance company and / or the repair shop after the identified damage has occurred, for example the type of windshield.

[0043] Further development of the procedure could include, as a step 380, sending the damage information directly to a selected or nearby repair shop so that an appointment can be booked for the windshield replacement. The repair shop can then order a suitable windshield even before the vehicle arrives. Optionally, it could also be provided that an autonomously driving vehicle, upon detecting damage, drives itself to a repair shop.

[0044] As already explained, it is also possible to perform the detection of an object impact or a stone chip according to the invention not only in the processing unit 100 in the vehicle. Accordingly, a cloud application is shown with the device according to Figure 4. The starting point here is an evaluation unit 400, which acquires corresponding microphone signals and, if applicable, further vehicle data from corresponding vehicle systems. Instead of performing an independent analysis of the microphone signals, the evaluation unit 400 sends the acquired signals and data to a server 420 in the cloud. This server performs the detection of windshield damage and damage classification according to the aforementioned method and reports the damage information back to the evaluation unit 400.The evaluation unit 400 can then forward the damage information, possibly along with further data such as environmental data, to the fleet operator 430 or the insurance company. Depending on the damage information, the fleet operator 430 or the insurance company can directly commission the repair shop and report the appointment back to the evaluation unit 400. Alternatively, the evaluation unit 400 can also automatically inform the repair shop 440 and schedule an appointment.

[0045] Book an appointment for windscreen replacement.

Claims

1. Method for detecting damage to a window of a vehicle, in particular a windshield, wherein the method • at least one initial microphone signal is captured (320), which represents a temporal progression of a sound, and • depending on the microphone signal, detects the impact of an object, in particular a stone, on the disc (330), and • a speed measurement is recorded, and • performs a damage classification of the impact of the object depending on the temporal course of the first microphone signal and the speed magnitude (330), • generates damage information when a crack is detected in the disc (360).

2. The method according to claim 1, characterized in that the method • at least one second microphone signal is captured, representing the wind noise at the vehicle while driving, and • additionally performs the damage classification depending on the second microphone signal, whereby it is specifically provided that the method determines the velocity and / or the impact energy of the object by means of the at least one second microphone signal 3. Method according to claim 1 or 2, characterized in that the method • at least one temporal consumption data of the vehicle is recorded, in particular a fuel and / or electricity consumption figure, and • determined a headwind speed value from the change in the consumption information, which represents the headwind acting on the vehicle, and The damage classification is additionally performed depending on the magnitude of the opposing speed.

4. Method according to claim 1 or 2, characterized in that the method • Environmental data of the vehicle was recorded (340) to prove the origin of the object, and • adds at least some of the contextual data to the damage information (360), in particular providing that the procedure • as environmental data, data from at least one camera and / or radar system are recorded (340), and • at least one preceding vehicle was identified from this data (340), and • adds this data and / or the identification of the vehicle ahead to the damage information (360).

5. Method according to one of the preceding claims, characterized in that the method • at least two first microphone signals captured (320) and • when detecting crack formation, a differential analysis of the at least two first microphone signals is performed (330), wherein it is particularly provided that the location of the crack formation is derived by means of the differential analysis.

6. Method according to one of the preceding claims, characterized in that the method • in the damage classification, derives an energy quantity from the at least one first microphone signal (330) that represents the energy input into the disk, and • detects crack formation depending on the energy quantity (330).

7. Method according to one of the preceding claims, characterized in that the method • that at least one initial microphone signal is detected (320) when the velocity quantity exceeds a velocity threshold.

8. Device for detecting damage to a window of a vehicle, in particular a windshield, comprising a processing unit (100) performing a method according to any one of claims 1 to 7, wherein the processing unit (100) • at least one initial microphone signal is captured (320), which represents a temporal progression of a sound, and • depending on the first microphone signal, detects the impact of an object, in particular a stone, on the disc (330), and • a speed measurement is recorded, and • performs a damage classification of the impact of the object depending on the temporal course of the first microphone signal and the speed magnitude (330), • generates damage information when a crack is detected in the disc (360).

9. System for detecting damage to a window of a vehicle, in particular a windshield, for carrying out a method according to one of claims 1 to 7, comprising a processing unit (100) and an evaluation unit (190), wherein • the processing unit (100) o detects at least one first microphone signal representing a temporal progression of a sound, and o detects a velocity quantity, and o sends the at least one first microphone signal and the velocity quantity to the evaluation unit, • the evaluation unit (190) recognizes a damage classification of the impact of the object, in particular a stone, on the disc depending on the first microphone signal and the speed parameter (330), and o if a crack is detected in the disc, damage information is generated (360).

10. System according to claim 9, characterized in that the • the processing unit (100) o records environmental data of the vehicle to prove the origin of the stone, and o depending on the detection of a crack formation, sends this environmental data to the evaluation unit, and • the evaluation unit (190) o generates damage information that contains at least some of the environmental data.

11. Vehicle with a device according to claim 8 or a processing unit (100) in a system according to one of claims 9 or 10.

Images