Collision evaluation

Abstract

To provide a method of analyzing collision data.SOLUTION: A collision is analyzed by receiving telematics data relating to a collision and determining a feature of the collision. An entity may be notified of the collision based on the determined features.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This invention relates to using data related to a collision in order to evaluate the characteristics of that collision. [Background technology]

[0002] Following a vehicle collision, the vehicles may require repairs, or insurance companies may be required to pay compensation for injuries. To take appropriate action, the circumstances of the collision need to be assessed by the parties involved to determine how responsibility should be shared and to assess the value of the claim.

[0003] Currently, the assessment of responsibility heavily relies on eyewitness testimony, but eyewitness testimony is prone to misrepresentation and driver denial of information. To mitigate this problem, recorded data, which can be recorded by the following means, may be used. - Event data recorders (EDRs) installed by the original equipment manufacturer (often connected to the airbag control circuit). - “Black box” telematics devices that can be used to monitor vehicle operation. Such devices may be installed by insurance providers to set insurance premiums based on how “safe” the driver is. - A "dashcam" (dashboard-mounted camera) that can provide a visual record of the collision. [Overview of the project] [Problems that the invention aims to solve]

[0004] However, the data provided by such recording devices requires analysis by trained personnel, which incurs associated costs. This problem is exacerbated when collision information is presented in a format that is difficult to exchange and interpret, potentially leading to a difficult and time-consuming process of determining responsibility.

[0005] After determining what happened in the collision and who is responsible (or liable), it is also necessary to obtain an estimate of the monetary compensation required by both parties. This usually involves a damage assessor investigating the damage, which is also a long and costly process. In particular, there are cases where simply replacing the vehicle is more cost-effective than repairing it ("total loss"), but this is only determined after a considerable amount of time and expense has been spent on the investigation.

[0006] Therefore, it is desirable to achieve solutions to one or more of the identified problems. [Means for solving the problem]

[0007] According to one aspect of the present invention, a method for analyzing collision data is provided, which includes the steps of: receiving telematics data relating to a collision; determining the severity of the collision based on the data; and notifying the subjects of the collision based on the determined severity of the collision. In this way, the subjects concerned are notified of the collision without requiring user intervention. This method can provide objective factual reports of the collision and the circumstances and effects of the collision. This can overcome the problems of unreliable witnesses and erroneous subjective reports of collisions.

[0008] Optionally, the method includes the step of identifying an impact area based on the received telematics data. In this way, the severity of the collision and subsequent related actions can be more accurately identified. Preferably, the impact area includes one or more zones related to the vehicle or one or more points on the vehicle.

[0009] For accuracy, the step of identifying the impact area may include the step of identifying the rotation of the vehicle and / or the partial or complete inversion of the vehicle.

[0010] The method may further include the step of identifying one or more events related to a collision. The related events may be one or more of hard braking, hard cornering, hard acceleration, hard deceleration, rotation of the vehicle, partial or complete inversion of the vehicle, traction loss, and secondary collision. By identifying the events related to the collision, a report of the course of the events related to the collision can be provided.

[0011] The method may include the step of modeling or simulating the collision, preferably with a resolution of 1 second or less. By modeling or simulating the collision, a report of the course of the events related to the collision can be provided.

[0012] Characteristics of repair For efficiency, the method may further include the step of identifying the repair cost.

[0013] Optionally, the repair cost is identified based on the cost of the parts within the impact area. Optionally, the method further includes the step of identifying the likelihood that a part has been damaged based on the received telematics data.

[0014] For prediction accuracy, the step of identifying the likelihood that a part has been damaged may include a comparison with past impact data. A classifier trained to identify the likelihood that a part has been damaged is used, and preferably the trained classifier is trained by the past impact data.

[0015] For efficiency, if the estimated repair cost is less than a threshold amount, the entity may include a repair entity, and if the estimated repair cost is greater than the threshold amount, the entity may include a scrap recovery entity. Preferably, the threshold amount is the replacement price.

[0016] Characteristics of notification For safety and speed, if the severity of the collision exceeds a threshold level, the entity may include an emergency service.

[0017] The subject may optionally include the occupants of the vehicles involved in the collision.

[0018] To enable a rapid assessment of the collision, the aforementioned entity may be an insurer.

[0019] For safety reasons, the aforementioned notice may include an assessment of the road-use suitability of one or more vehicles involved in the collision.

[0020] Specific characteristics For safety reasons, the method may further include a step of determining the probability of injury. Optionally, the step of determining the probability of injury may include a step of determining the magnitude and direction of the collision. Optionally, the step of determining the probability of injury may include a step of determining the rotation of the vehicle. Optionally, the step of determining the probability of injury may include a step of detecting a partial or complete reversal of the vehicle. Optionally, the probability of injury may relate to the probability of whiplash. Optionally, the method may further include a step of determining the cost of injury compensation.

[0021] For accuracy, the step of identifying the probability of injury may include a comparison with past impact data. Optionally, a trained classifier is used to identify the probability of injury, and the trained classifier is trained using the past impact data.

[0022] Reception characteristics For accuracy, telematics data may include data from gyroscopes, accelerometers, GPS data, video recordings, in-vehicle diagnostic data, and / or audio recordings.

[0023] For a smooth vehicle trajectory, the method may include the step of interpolating GPS and accelerometer data points to determine the vehicle trajectory.

[0024] For speed and / or efficiency, telematics data may be received via a wireless link to the device that collected the data.

[0025] For reliability reasons, the wireless link may include a satellite link.

[0026] Telematics data may include data for a period of 6 to 10 seconds, including the collision. Telematics data (or a portion thereof) may be, for example, 20 Hz, 50 Hz, 100 Hz, 400 Hz, 500 Hz, 1000 Hz, or 5000 Hz, or within a range formed by any two of these exemplary values. Telematics data (or a portion thereof) preferably has a resolution of at least 1 second or less, preferably at least 100 milliseconds, more preferably at least 50 milliseconds, and more preferably at least 10 milliseconds. In some examples, the resolution is 1 millisecond or less. The first portion of the telematics data, preferably accelerometer data and / or gyroscope data, may have a resolution of 1 second or less. The second portion of the telematics data, preferably GPS data, may have a lower resolution than the first portion.

[0027] The step of modeling or simulating the collision may have a resolution of at least 1 second or less (preferably at least 100 milliseconds, more preferably at least 50 milliseconds, and more preferably at least 10 milliseconds). A resolution of at least 1 second may allow for the decomposition of different events before and after the collision (e.g., sudden braking, subsequent collision). The step of modeling or simulating may have a resolution of, for example, 0.2 ms, 1 ms, 2 ms, 2.5 ms, 5 ms, 10 ms, 20 ms, 50 ms, or 100 ms, or be within a range formed by any two of these exemplary values. This can help clarify a series of events that may exceed the limits of human perception and thus provide insights into the collision. A resolution of at least 1 second also makes it possible to identify forces acting on the vehicle with sufficient resolution to identify the area of ​​impact. The resolution may be less than 1 millisecond in some examples. The step of modeling or simulating the collision preferably includes a step of identifying one or more events related to the collision.

[0028] Characteristics of the analysis For efficiency, the method may further include a step of identifying a measure of responsibility. A measure of responsibility may be identified based on priority, road or junction classification, vehicle lane, lane change operation, measure of operational intent, and / or reversing movement. A measure of responsibility may be identified based on road traffic laws, case law, and / or mapping information. The method may further include a step of obtaining data related to road traffic laws, case law, and / or mapping information (preferably from external sources). Preferably, the method includes a step of identifying a measure of responsibility using a cognitive inference model. By considering both telematics data and other data when identifying a measure of responsibility, an objective assessment of responsibility can be provided. Using a cognitive inference model can enable a complex evaluation of numerous factors that may be considered when identifying responsibility.

[0029] Optionally, the telematics data may be used to evaluate the driving of any further entities involved in the collision. Optionally, the telematics data may be used to identify the momentum of any further entities.

[0030] For accuracy, the method may further include the step of obtaining data relating to one or more other vehicles involved in the collision (preferably from an external source). Optionally, the data relating to the additional entities may be obtained using vehicle registration numbers.

[0031] For accuracy and / or to identify additional information, the method may further include the step of determining the mass of the further subject based on the data of the further subject.

[0032] To determine responsibility, the method may further include the step of determining the velocity of the further entity based on its momentum and mass. Optionally, the method may further include the step of determining whether the further entity was traveling at speed prior to the collision.

[0033] For accuracy, the step of determining the severity of the collision may include the step of determining the time of the collision.

[0034] For accuracy, the step of determining the time of the collision may include determining the time of the magnitude of the maximum acceleration based on the received telematics data. Optionally, the step of determining the magnitude of the maximum acceleration may include determining the root mean square of the magnitude of the acceleration in two or three dimensions.

[0035] For accuracy, the step of determining the time of impact may include the step of determining acceleration that exceeds the maximum braking force and / or cornering force.

[0036] Optionally, the step of determining the time of collision may include the step of determining the time of maximum vertical acceleration. Since vertical acceleration is very rare under normal circumstances, detecting maximum vertical acceleration may strongly indicate a collision.

[0037] Optionally, the step of determining the severity of the collision includes the step of determining the forces caused by the collision.

[0038] For accuracy, the step of identifying the forces caused by the collision may include the step of modeling the elasticity of the objects involved in the collision or the step of identifying a model of the elasticity of the objects involved in the collision.

[0039] For accuracy, the model of the elasticity of the objects involved in the collision may include the step of identifying at least one spring constant based on the received telematics data.

[0040] For accuracy, the step of determining the spring constant based on the received telematics data may include the step of determining the frequency of the data from the received telematics data. Preferably, the telematics data includes accelerometer data, and the force is determined according to the accelerometer data.

[0041] For accuracy, the step of identifying the forces caused by the collision may include the step of modeling or identifying the change in momentum of the objects involved in the collision. Optionally, the change in momentum is identified by integrating the acceleration over a time window before and after the identified impact time.

[0042] Steps to identify the severity of a collision and / or related events may include modeling or simulating the collision to assess the severity of the collision and / or related events. Steps to identify other features may include modeling or simulating the collision to assess such features. Such features may include the impact area, repair costs, the likelihood of parts being damaged, the probability of injury, the magnitude or direction of the collision, the rotation of the vehicles, injury compensation costs, the measure of liability, the momentum of further entities, the velocity of further entities, the time of the collision, the time of the magnitude of the maximum acceleration, the forces caused by the collision, the elasticity of the objects involved in the collision, the spring constant, the change in momentum of the objects involved in the collision, events preceding the collision, acceleration above a threshold level, changes in direction and / or speed, and / or traction force loss.

[0043] Interface Features For ease of use, the method may further include the steps of identifying at least one event preceding (or following or related to) the collision based on the received telematics data, and sequentially displaying the at least one event and the collision via a user interface.

[0044] According to another aspect of the present invention, a method for displaying collision data is provided, which includes the steps of: receiving telematics data relating to a collision; identifying at least one event preceding (or following or related to) the collision based on the received telematics data; and sequentially displaying the at least one event and the collision via a user interface. In this way, a user can easily and quickly identify key events leading to a collision, which may then determine further action (e.g., further investigation or approval of a claim).

[0045] For accuracy, the step of identifying at least one event preceding the collision based on the received telematics data may include the step of identifying an acceleration above a threshold level.

[0046] For accuracy, the step of identifying at least one event preceding the collision based on the received telematics data may include the step of identifying a change in direction and / or speed. Optionally, the change in direction includes moving to a different road.

[0047] For ease of use, this method may further include means for the user to change the time period displayed in the interface.

[0048] For ease of use, the method may further include a step of indicating the characteristics of the telematics data relating to the event. Optionally, the indicated characteristics include a speed indicator, a g-force measurement, a braking indication, a cornering indication, a traction loss event indication, a vehicle rotation indication, a partial or complete vehicle reversal indication, an impact area indication, and / or a safe speed (which may depend on statutory speed limits and / or weather and / or road characteristics).

[0049] For ease of use, the method may further include a step of displaying a map of the collision area. Optionally, the method may further include a step of displaying a representation of the vehicle. Optionally, the method may further include a step of displaying the vehicle's path.

[0050] For ease of use, this method may further include the step of displaying the location of one or more events.

[0051] Optionally, the method further includes the step of displaying a situation indicator related to the time of the collision. Optionally, the situation indicator includes indications of weather conditions and / or traffic density.

[0052] For ease of use, the method may further include the step of providing the user with means for annotating one or more of the aforementioned events.

[0053] Optionally, the method further includes a step of providing an indication of damage to a vehicle. Optionally, the indication includes a step of indicating the area of ​​the vehicle that is damaged. Optionally, the indication includes a 3D visualization of the vehicle. In this way, a method for remotely identifying likely damage to a vehicle is provided to the user.

[0054] According to another aspect of the present invention, an apparatus for analyzing collision data is provided, the apparatus comprising means (e.g., in the form of a appropriately programmed processor and associated memory) for receiving telematics data relating to a collision, means (e.g., in the form of a appropriately programmed processor and associated memory) for determining the severity of a collision based on the data, and means (e.g., in the form of a appropriately programmed processor and associated memory) for notifying a subject of the collision based on the determined severity of the collision.

[0055] According to another aspect of the present invention, an apparatus for displaying collision data is provided, the apparatus comprising means (e.g., in the form of a appropriately programmed processor and associated memory) for receiving telematics data relating to a collision, means (e.g., in the form of a appropriately programmed processor and associated memory) for identifying at least one event preceding the collision based on the received telematics data, and means (e.g., in the form of a appropriately programmed processor and associated memory) for sequentially displaying the at least one event and the collision via a user interface.

[0056] The present invention also relates to apparatus adapted to perform any of the methods described herein.

[0057] According to another aspect of the present invention, a system for analyzing collision data is provided, which includes an apparatus described herein and means (for example, in the form of a appropriately programmed processor and associated memory) for transmitting telematics data to the apparatus.

[0058] Optionally, the system further includes means (for example, in the form of a appropriately programmed processor and associated memory) for recording the telematics data.

[0059] Optionally, the means for transmitting and / or recording the telematics data includes a mobile phone.

[0060] According to another aspect of the present invention, a method for modeling a collision from telematics data is provided. The telematics data may be raw numerical data from a general-purpose file. The method may include the step of organizing the telematics data. The method may include the step of calculating a collision model from the telematics data. The method may include the step of identifying collision-related events (leading to and / or after the collision). The method may preferably include the step of indicating the collision-leading (and / or after the collision) events using digital mapping software. The method may include the step of indicating minute vehicle movements (with a resolution of less than 1 second) using digital mapping software. The method may include the step of indicating collision-related vehicle movements and / or events with a resolution of less than 1 second (e.g., a resolution in the range of 1 to 1000 milliseconds, more preferably 10 to 100 milliseconds). The method can provide objective factual reports of collisions and the circumstances and effects of collisions. This can overcome the problems of unreliable witnesses and erroneous subjective reports of collisions.

[0061] According to another aspect of the present invention, a method for analyzing collision data is provided, the method comprising the steps of receiving telematics data relating to a collision, identifying one or more events relating to the collision, and providing a sequence of events relating to the collision. The method may include the features described above. The method can provide objective factual reports of collisions and the circumstances and effects of collisions. This can overcome the problems of unreliable witnesses and erroneous subjective reports of collisions.

[0062] According to another aspect of the present invention, a method for analyzing collision data is provided, comprising the steps of receiving telematics data relating to a collision, identifying a measure of responsibility based on the data, and notifying a subject of the collision based on the identified severity of the collision. The method may include the features described above. The method can provide objective factual reports of collisions and the circumstances and effects of collisions. This can overcome the problems of unreliable eyewitnesses and erroneous subjective reports of collisions. By considering telematics data (and optionally other data) in identifying a measure of responsibility, an objective assessment of responsibility can be provided. A complex evaluation of multiple factors may be considered when identifying responsibility. The method can provide objective factual reports of collisions and the circumstances and effects of collisions. This can overcome the problems of unreliable eyewitnesses and erroneous subjective reports of collisions.

[0063] As used herein, “telematics data” may refer to any data relating to the movement or driving of an entity such as a vehicle.

[0064] As used herein, “operator” may refer to any entity that controls the vehicle, and this operator may be a person, a device, or a combination thereof.

[0065] As used herein, "collision" may refer to any event that exceeds a threshold measurement, and a collision does not necessarily include impact.

[0066] The present invention extends to any novel embodiments or features described and / or illustrated herein.

[0067] Further features of the present invention are characterized by other independent and dependent claims.

[0068] Any feature in one aspect of the present invention may be applied to other aspects of the present invention in any suitable combination. In particular, a method aspect may be applied to an apparatus aspect, and vice versa.

[0069] Furthermore, features implemented in hardware may also be implemented in software, and vice versa. Any references to software and hardware features in this specification should be interpreted appropriately.

[0070] Any apparatus feature described herein may also be provided as a method feature, and vice versa. Means-plus-function features as used herein may also be expressed alternatively with respect to their corresponding structures, such as appropriately programmed processors and associated memory.

[0071] Furthermore, it should be understood that certain combinations of the various features described and defined in any embodiment of the present invention may be carried out and / or supplied and / or used independently.

[0072] The present invention also provides computer programs and computer program products, which include software code adapted to perform any of the methods described herein, including any or all of the configuration steps, when executed by a data processing device.

[0073] The present invention also provides computer programs and computer program products, which include software code that, when executed on a data processing device, includes any of the features of the device described herein.

[0074] The present invention also provides computer programs and computer program products having an operating system that supports a computer program for performing any of the methods described herein and / or for embodying any of the features of the apparatus described herein.

[0075] The present invention also provides a computer-readable medium on which the aforementioned computer program is stored.

[0076] The present invention also provides signals for carrying the aforementioned computer programs, and methods for transmitting such signals.

[0077] The present invention substantially extends to the methods and / or apparatus described herein with reference to the accompanying drawings.

[0078] Next, the present invention will be described as an example with reference to the attached drawings. [Brief explanation of the drawing]

[0079] [Figure 1] This diagram shows a system for recording collision evaluations. [Figure 2] This is a diagram showing a data recording device. [Figure 3] This diagram shows a flowchart for responding to collisions when collision evaluation is required. [Figure 4] This is a diagram illustrating the assessment of collision damage. [Figure 5] This figure shows an example of a graph of accelerometer data. [Figure 6] This is a diagram showing an exemplary model of a vehicle. [Figure 7] This figure illustrates the behavior of a spring that can be used in collision models. [Figure 8] This diagram illustrates the assessment of impact based on a spring model. [Figure 9] This figure illustrates the assessment of impact based on momentum vectors. [Figure 10] This figure shows recorded GPS data that can be used in collision evaluation. [Figure 11] This diagram shows a flowchart for determining whether repairing a vehicle is cost-effective. [Figure 12] This diagram shows a flowchart for identifying the payment amount after a collision. [Figure 13] This diagram shows a flowchart illustrating the transfer of data related to collision evaluation. [Figure 14]This diagram shows a detailed example of data transfer related to collision evaluation. [Figure 15] This diagram shows an interface used to view data related to multiple collisions. [Figure 16a] This diagram shows an interface for viewing events related to collisions. [Figure 16b] This diagram shows an interface for viewing events related to collisions. [Figure 16c] This diagram shows an interface for viewing events related to collisions. [Figure 16d] This diagram shows an interface for viewing events related to collisions. [Figure 17] This diagram shows a view of the interface for displaying a street-level map view of the collision location. [Figure 18a] This diagram shows a view of the interface, including a section for viewing information related to potential claims. [Figure 18b] This diagram shows a view of the interface, including a section for viewing information related to potential claims. [Figure 18c] This diagram shows a view of the interface, including a section for viewing information related to potential claims. [Figure 18d] This diagram shows a view of the interface, including a section for viewing information related to potential claims. [Modes for carrying out the invention]

[0080] This invention relates to a method for evaluating a collision and identifying appropriate actions based on the output of this evaluation. The severity of the collision may be used to identify the entity that should be notified, such as an ambulance or a tow truck.

[0081] The severity of a collision can depend on several factors, including the area of ​​impact. This information makes it possible to determine the repair costs and a list of necessary parts.

[0082] The present invention also relates to a user interface that sequentially presents the key events leading up to a collision so that the user can quickly obtain an accurate picture of the collision.

[0083] The present invention also relates to a model that takes raw numerical data from a general-purpose file as input, organizes it, and calculates it in digital mapping software in order to represent minute movements of a vehicle (i.e., with a resolution of less than one second).

[0084] The claims analysis platform described herein consumes telematics data provided by the parties involved in the conflict via dynamic interfaces (e.g., Representational State (REST) ​​application program interfaces (APIs)) and / or import functions. This data is analyzed and insights are derived.

[0085] Artificial intelligence engines can be used to continuously refine models in order to better understand the nuances surrounding observations of individual accidents, and as a result, the AI ​​engine can continuously improve the accuracy of insights derived from newly encountered telematics datasets, such as the causes and consequences leading to multi-vehicle collisions.

[0086] Figure 1 shows a system for evaluating collisions.

[0087] Vehicle 1000 includes a crew 1100 such as an operator (driver) 112, one or more passengers 114, and / or cargo 116.

[0088] The vehicle may be semi-autonomous ("autonomous driving"), in which case the operator (driver) 112 can be considered to be a processor or operating system that controls the vehicle.

[0089] Throughout the operation of the vehicle, telematics data is recorded using one or more data recording devices 1200, such as a smartphone 112, a dashcam 114, or an event data recorder (EDR) 116, commonly known as a "black box."

[0090] Vehicle 1000 communicates with server 1300 via a network 1400 such as the internet.

[0091] Figure 2 shows the data recording device.

[0092] This data recording device 1200 is - One or more gyroscopes 1202, - One or more accelerometers 1204, - GPS unit 1206, and - Time recording means (e.g., clock 1208), - Processor 1214 for integrating data recording, - Memory 1216 where data is stored, - Antenna 1218 for transmitting recorded data (e.g., to a server where evaluation is performed) or for receiving updated software, - Power supply 1220, for example, an internal battery or a connector (such as a USB connection that transmits power from a vehicle battery), Includes.

[0093] These combinations of components are advantageous as they are commonly found in smartphones or in black boxes installed by car manufacturers or insurance companies. The device may also be an embedded device fixed to the vehicle.

[0094] The means of recording is, - Video recording means 1210 such as a dashcam, and - Audio recording means, i.e., microphone 1212, It is preferable to include it as well.

[0095] Due to the unpredictable nature of collisions, it is necessary to record data throughout the entire process to ensure that data corresponding to the collision is captured. However, out of consideration for user privacy, it is desirable to record and / or store only the necessary data. Therefore, only recorded data related to the temporal period preceding and following the collision, such as data from the start of the process in which the collision occurs, or data from a certain amount of time preceding and following the collision, is retained. In one example, a data file from a 6-10 second excerpt of the collision is stored and analyzed. Data unrelated to the collision is overwritten or deleted. This recording / deletion may occur cyclically so that only the last "X" minutes of data are always available.

[0096] The data stored includes acceleration ("g-force"), GPS location, speed, time, and location data, as well as the location of any impact and at least one of the time intervals in which the impact occurs. Video data and / or audio data are also stored where available. If the vehicle has sufficient sensors, information such as brake operation, steering adjustments, airbag deployment, and engine status may also be recorded.

[0097] If the data recorder includes (or can connect to) the user's mobile phone, data relating to interactions with the user's mobile phone (e.g., whether or not it was on a call) may be recorded. The data recorder may connect to the user's mobile phone using Bluetooth® or another form of wired or wireless connection.

[0098] Other data, such as weather and road conditions, are also stored. Such data may be obtained from sources not connected to the vehicle involved in those conditions, such as weather forecasting services. This additional data is collected using data recording means 1200, and the processor 1214 connects to the area network using antenna 1218 to download the data.

[0099] This additional data may be obtained later using historical weather information related to the collision location. This embodiment is advantageous in situations where the data recording means cannot connect to a network at the time of the collision.

[0100] The stored data is used to assess the conflict, and this assessment is used to identify further actions, such as notifying the subject or evaluating the claim.

[0101] In some embodiments, the data recording device can also be used as an on-board diagnostic (OBD) system, where the device is used to notify the vehicle occupants of any potential problems, such as mechanical failures. The device may also suggest steps to be taken to address any problems.

[0102] Figure 3 shows a flowchart for responding to collisions when collision evaluation is required.

[0103] 1. In the first step 310, a collision is detected using data recording means, and an acceleration (g-force) exceeding a threshold (in a certain direction) indicates a collision. Other measurements, such as calculated energy absorbed by the vehicle, or indications that airbags have deployed, may also be used to detect a collision. It may be advantageous to combine multiple methods to avoid false positives (for example, a sudden acceleration forward might indicate a "rear-end" collision, while a similar acceleration backward could simply be sudden braking).

[0104] Immediately after a collision, the vehicle's insurance company receives a first notice of loss (FNOL), which includes recorded information related to the collision. Obtaining a detailed FNOL quickly allows the insurance company to promptly assess the impact of the collision, such as whether monetary compensation is needed or if a temporary vehicle is required.

[0105] Recorded data related to the conflict will be used later in the evaluation and stored for access if either party deems it necessary; this data may include evidence of decisions, such as denying a claim.

[0106] 2. This collision is evaluated in the second step 320, which will be discussed in more detail in a later section of this explanation. The severity of the collision is the output of the evaluation.

[0107] Telematics data related to the collision is transmitted to a server, stored, and the collision is evaluated. Alternatively, this process may be performed on a local device such as a smartphone 112.

[0108] 3. In step 330, the severity of the collision is queried, and step 4 depends on the severity of the collision as follows:

[0109] 4A. If the severity of the collision is low, the vehicle occupants will be notified in step 4, 2400, and this notification will be used to inquire about the nature of the collision. A collision may be falsely detected if the g-force threshold is exceeded due to strong braking, in which case the occupants can confirm whether a collision occurred and whether assistance is needed.

[0110] Notifications of minor collisions are also used to inform the vehicle occupants of the vehicle's condition. Such notifications may indicate whether the vehicle is safe to continue driving, and may also inform occupants of parts that are expected to be damaged, which can then be examined by the vehicle occupants or a mechanic. This process is described in more detail below.

[0111] 4B. In cases of high severity of collision, emergency services are contacted in Step 4, 350, for example, an ambulance service is notified. This notification incorporates information related to the collision, such as the number of vehicles involved, location, severity, and any anticipated injuries. Depending on the characteristics of the collision, the service to be contacted may be the police, as a high-severity collision occurring on a highway may result in the highway being closed, or the fire department may be contacted to control any resulting fires, as a collision involving an impact to the fuel tank area may result in the fire being extinguished.

[0112] In some embodiments, when a high-severity collision is detected, a notification is transmitted to the occupants containing suggested actions, including calling an ambulance if injuries are anticipated. The occupants can then confirm whether such action is appropriate. If the occupants do not respond within a set time period, the suggested action is taken, and, for example, an ambulance service is notified.

[0113] 5. In Step 5, 360, the assessment is used to estimate the costs associated with the collision, including the cost of repairing the vehicles involved and potential injury claim costs.

[0114] This cost estimate 360 ​​takes into account the severity of the collision and the location of any impact to predict damage to parts or occupants. This assessment is based solely on data obtained from data recording devices, thereby avoiding the costs and delays associated with inspections by mechanics or doctors. This process is described in more detail below.

[0115] These costs are used to inform decisions about actions to be taken, and if high costs are anticipated, further examinations by mechanics and / or doctors may be considered; if lower costs are anticipated, it may be more cost-effective to pay each party involved without further investigation.

[0116] In some embodiments, parts orders are automatically generated, parts are procured along with their costs, and purchase orders for these parts are automatically filled out. An agent may review such orders before they are transmitted.

[0117] 6. In Step 6, 370, the repair costs are used to assess whether it is worthwhile to repair at least one of the vehicles involved, that is, whether the repair costs (including elements such as labor costs) are less than the replacement costs, which is explained in more detail in Figure 7.

[0118] 7A. If it is deemed worthwhile to repair (i.e., the replacement cost exceeds the estimated repair cost), the mechanic will be notified in Step 7, 380. The towing company will also be notified so that the vehicle can be promptly transported to the mechanic.

[0119] 7B. If it is determined that the vehicle is not worth repairing, the scrap yard will be notified in step 7, 390, and the vehicle will be collected and scrapped.

[0120] The stored data is used as input for a user interface that allows users, such as insurance agents, to view details about collisions in order to understand their characteristics (details below). In this way, users can confirm or intervene in automated actions identified from the telematics data, or enter additional information.

[0121] In some cases, the steps described herein are incorporated into such an interface, where, when a collision is detected and assessed, the agent receives a notification and can follow the described method, view the output of the collision assessment, and send notifications to occupants / emergency services / repairmen as needed.

[0122] In some examples, - Inform the families of the vehicle's occupants of the collision. - Enables the parties involved to assess the operator's performance, - Send the location of any occupant to the parties involved, - Data is transmitted to an insurance company that provides a device (e.g., a web-based form) so that the information required for the claim is automatically entered, and an automated claim is created. Further notifications will be sent for this purpose.

[0123] Estimated Cost 360 includes at least one of the expected cost, expected cost range, and expected maximum cost. Different costs may be used in different circumstances; a higher maximum cost may trigger further investigation; a lower expected cost may be used in proposals to the insured party (i.e., if the party accepts this proposal, the case can be resolved quickly); and as a result of the expected cost range, any claim falling within this range will be paid with little to no further investigation, while any claim above this range will be investigated.

[0124] The estimated cost of 360 is then stored and / or transmitted to third parties such as the vehicle occupants and / or insurance agents.

[0125] Some of the features of this system and method are as follows: • Inflexible data from any source can be accepted and formatted for the stated purpose. • GPS and accelerometer data points may be interpolated to smooth the movement. • The bidirectional motion of the accelerometer can be analyzed to identify the impact zone. The severity of a collision can be refined by considering data from vehicle registration, correlating the vehicle type and model with the impact zone and G-force to obtain the vehicle's damage zone and its respective severity. • To determine the value of the vehicle, the amount of damage is estimated based on the damage zone and severity, and compared to the value of the vehicle. • Calculate the unknown speed of the second vehicle using the masses of the two vehicles and the speed / g force of one vehicle.

[0126] Collision evaluation Figure 4 shows the collision damage assessment.

[0127] The data recorded after a collision is used to estimate the location and severity of each impact, as well as the resulting damage to the vehicles and occupants involved.

[0128] The impact area (or "zone") can be identified from the direction of acceleration recorded by the accelerometer. Figure 4 shows a vehicle divided into eight different areas, although more or fewer areas may be appropriate depending on the measurement accuracy of the accelerometer. This process is explained in more detail below with reference to Figures 5-10. Several areas may be affected; for example, areas adjacent to the impact area may also be affected, especially when high accelerations are measured. Similarly, several accelerations may be detected, for example, if a vehicle "rear-ends" a vehicle in front and is then "rear-ended" by a following vehicle.

[0129] By identifying and analyzing the spikes and minute movements of the vehicle's g-force in a given direction, the impact direction and therefore the impact zone can be determined. Once the impact zone is identified, further analysis can quantify the damage to that area, which can then be correlated with the vehicle's value to determine the total loss. If a third-party vehicle is involved, the mass of each vehicle can be assumed (for example, by examining the type, model, and weight of a vehicle with a given vehicle registration number), and the speed of the third-party vehicle can be estimated using the known vehicle's speed and g-force.

[0130] Figure 5 shows an example of a graph of accelerometer data.

[0131] Acceleration is divided into lateral (left-right), vertical (front-back), and vertical components, and further, • The time when the magnitude of acceleration is highest (for example, the time when the root mean square of the 2D or 3D acceleration is at its maximum), • Acceleration exceeding the maximum braking force, and / or • Time of maximum vertical acceleration, The time of impact is determined from one or more of these factors.

[0132] These measures can be combined to form a measure of the probability of the impact time (if all three conditions are met simultaneously, it is highly likely to indicate the impact time).

[0133] Vertical acceleration is also used in the initial assessment of collision severity; a large or sustained vertical acceleration indicates a serious collision.

[0134] To evaluate the velocity and displacement involved in the collision, acceleration data is integrated (once for velocity, twice for displacement). Because large time steps between data points can result in inaccurate acceleration / displacement values, warnings may be issued regarding the calculation if large time steps are present.

[0135] In some embodiments, velocity and displacement may also be obtained using sensor data, with GPS data used to obtain displacement and speedometer data used to obtain velocity. Such data may be combined with accelerometer data, which may be desirable because accelerometer data is often measured more frequently than other data, i.e., GPS location may be recorded at a rate of 1 Hz while acceleration data may be recorded at a rate of 100 Hz. Data from GPS and accelerometers may be interpolated to provide a smoother vehicle trajectory. In some other examples, acceleration data is recorded at rates of 10 Hz, 50 Hz, 100 Hz, 400 Hz, 500 Hz, or 1000 Hz.

[0136] In some embodiments, accelerometer data is used to obtain changes in acceleration, velocity, and position after impact, and initial values ​​are obtained from other recording means; for example, the initial velocity (u0) may be taken from the reading of a speedometer.

[0137] In some embodiments, gyroscope data is used to improve the accuracy of minute vehicle movements by combining gyroscope data with accelerometer data, for example, using an inertial tracking method. This method enhances impact and event detection methods to also take vehicle rotation into account, and this rotational detail can be included in further analysis.

[0138] Furthermore, by including rotational information, it becomes possible to treat the vehicle as a dimensional body that maintains the direction, magnitude, and position of the impact relative to the body's center of gravity. This allows for a more accurate characterization of the impact on the vehicle and a more accurate estimate of the repair costs for such a vehicle.

[0139] This method may include a process for determining the position of the sensor relative to the vehicle's center of gravity. This method may further include a process for estimating the vehicle's moment of inertia.

[0140] In some embodiments, telematics data is used to detect a partial or complete reversal of a vehicle. This method may include detecting changes in acceleration direction by integrating gravity or gyroscope data, or other analyses of telematics or in-vehicle diagnostic data.

[0141] Such events can be incorporated into vehicle damage assessments, particularly those for roofs and windows, which can improve the accuracy of repair estimates.

[0142] In some embodiments, the telematics data is used to detect slip or traction loss events. This may include a method for detecting a difference between the vehicle's orientation (defined by the front and rear axes) and its direction of travel (defined by speed / momentum). This may include ABS signals, as well as information from on-board diagnostics such as steering, accelerator, and brake inputs.

[0143] Figure 6 shows an exemplary model of the vehicle.

[0144] To represent elasticity, the vehicle is modeled as a mass 602 connected to one or more springs 604, with one or more springs for each direction of motion under consideration. For example, in a head-on collision, only longitudinal motion is considered, and only one spring may be used; more complex collisions may use three springs: longitudinal, transverse, and vertical. Each spring has a spring constant used to model the elastoplastic behavior of the vehicle during the collision.

[0145] Such spring combinations can be used to represent elastic behavior together with dampers that represent energy dissipation. A method is used in which one spring is used in each direction, and these springs have a load stiffness k L and unloading stiffness k U It is preferable that the spring constants differ depending on whether or not a load is applied (i.e., compressed), characterized by k. U There are three types of springs, depending on the value of [the variable].

[0146] 1. Elasticity: k L = k U When this is the case, there is a perfectly elastic spring. The spring returns to its initial position without energy dissipation.

[0147] 2. Plasticity: k U = 1, all energy is dissipated, there is no rebound, and the final deformation is equal to the maximum deflection.

[0148] 3. Elastic - plasticity: k U > k L When this is the case, part of the energy is dissipated and the spring is permanently deformed after rebound.

[0149] After each impact, the physical characteristics of the vehicle change, and thus any subsequent collisions need to be modeled using different loading and unloading spring constants.

[0150] Figure 7 shows the behavior of a spring that can be used in a collision model.

[0151] The force F exerted on the vehicle by the collision is decomposed into x and y components F x and F y acting on the x and y axis springs respectively. For each direction i, the magnitude Fi of the force of a single impact in direction i is desired. Hooke's law is used, F i = k L,i d c,i where k L,i is the loading spring constant and d c,i is the compression displacement of the vehicle body during impact. Figure 6 shows a model of a single impact, with each side of the vehicle modeled as a spring system, and the springs being elasto - plastic and thus deforming during impact.

[0152] As shown in Figure 7, when a force is applied, the spring is first compressed 702, which causes a compression displacement d c . Then this force is released 704, and an elastic rebound displacement d e occurs. Here k U>k L Therefore, the result of the impact d c -d e Permanent deformation of length d p This occurs.

[0153] k L,i To calculate this, it is assumed that the motion is sinusoidal,

[0154]

number

[0155] And in the formula s i (t) is the displacement in the i direction, ω e,i It is calculated as the frequency of a sinusoidal curve fitted to impact (e.g., accelerometer) data, and v i imp This is the velocity at the time of impact in direction i.

[0156] This frequency is assumed to be the angular natural frequency, therefore

[0157]

number

[0158] In this equation, m is the mass of the vehicle.

[0159] The total impact of the collisions is,

[0160]

number

[0161] It is calculated by, in the formula Vi postimp is the velocity after impact in direction i.

[0162] therefore,

[0163]

number

[0164] and

[0165]

number

[0166] Here, Fi is the force exerted in the i-direction at the time of impact. The force calculated here is used as a measure of severity.

[0167] The angle of impact is,

[0168]

number

[0169] The angle can be estimated by the coordinate system, and the coordinate system is selected to obtain a detectable angle (preferably the x and y directions are equal to the longitudinal and lateral directions of the vehicle).

[0170] This model allows for the estimation of collision severity even without knowledge of the characteristics of other vehicles involved in the collision. If the characteristics of these other vehicles are known, the calculation can be improved by including these characteristics or by calculating the effective elasticity of the other vehicles involved.

[0171] Severity can be expressed as a numerical measure, such as force in Newtons or energy in Joules, or as a severity category, for example, an insurance company might designate severity as "high," "medium," and "low."

[0172] In some embodiments, multiple impacts are modeled. To model two impacts on the same side, the unloading spring constant k is used. U,i It is also necessary,

[0173]

number

[0174] It is obtained by calculation.

[0175] Figure 9 shows the assessment of impact based on momentum vectors.

[0176] To evaluate a collision, for example, a momentum method may be used. The change in the vehicle's momentum before and after the time of impact is directly related to the magnitude of the impact force, and the direction of the momentum change is related to the zone of impact. These characteristics are identified by calculating the vehicle's motion as a momentum vector over a short time period including the time of impact. The magnitude of this vector can then be used as a measure of the impact force or the severity of the collision.

[0177] The change in the vehicle's momentum Δp is,

[0178]

number

[0179] It is calculated as, and in the formula t imp is the time of impact, and ε is (t imp-ε from t imp+ε Here, a is a small time window (such as one that defines a small time window before and after the impact time), m is the mass of the vehicle, and a is the acceleration (from accelerometer data).

[0180] The magnitude of the momentum vector is,

[0181]

number

[0182] It is calculated as, and in the formula, p x , p y , and p zThese are the momentum components in each direction. Their magnitude is used as a measure of the severity of the impact. The impact zone is obtained by rotating the momentum vector by 180°.

[0183] In some embodiments, multiple models are used to evaluate collisions. For example, contrasting outputs when models do not agree on the severity of the collision may lead to a deeper investigation.

[0184] In some embodiments, the calculations used depend on the data collected; if there is little data, simple calculations based on a simplified model of the collision are used, while if more data is available, more accurate calculations based on a more complex model that takes into account, for example, the characteristics of other vehicles involved, are used. Subsequently, simple calculations are used for initial damage estimation, but more detailed calculations are used as needed to obtain a more detailed reconstruction that may be desired when considering liability.

[0185] Figure 10 shows recorded GPS data that can be used in collision evaluation.

[0186] Figure 10a shows the GPS track obtained from the recorded data.

[0187] Indicator 1012 indicates the time of the collision, and it can be seen that a large y-displacement occurred immediately after the collision, suggesting a large lateral impact.

[0188] Figure 10b shows the same data, but focuses on the period around the time of the accident.

[0189] This graph suggests that the vehicle was likely turning before the collision occurred.

[0190] In some embodiments, collision assessment uses a supervised learning algorithm and recorded data is used to estimate the severity and location of any impact based on past collision data, such recorded / past data may include accelerometer and speed data, as well as information about the vehicles involved in the collision (e.g., model, year of manufacture). Such algorithms may improve in accuracy as more data becomes available and may be used in conjunction with other methods as described herein until an acceptable level of accuracy is reached.

[0191] Damage assessment The vehicle's type and model can be known based on existing insurance information or identified through a database search of registration numbers. In this way, a complete parts list of the vehicle can be created, along with the location of each part within the vehicle.

[0192] Damage calculations take into account the severity and direction / location of the force, and car parts are damaged in a certain order. For example, in a frontal collision, a threshold severity scale is needed before the bumper is damaged, a higher threshold severity scale is needed before the headlights are damaged, then the hood, windshield, and finally structural damage to the chassis with a very high severity.

[0193] Zones containing crumple zones or reinforced components protect the components behind these zones. Therefore, a higher severity collision threshold would need to be met before such protected components are considered damaged.

[0194] To identify a list of parts that are likely to require repair or replacement, historical data from similar vehicles involved in similar collisions may be retrieved.

[0195] A trained classifier can be used to determine the extent of damage and generate more accurate predictions about which parts are likely to need replacing. Training data includes information about past collisions and the parts that needed replacing. This historical data can then be used to create a model that identifies which parts need replacing, based on the collision information. Such a model may include adjustments to threshold levels for when different parts are damaged, or it may be more complex, for example, it may infer interactions between parts (e.g., when the windshield is broken, there is usually some damage to the seat upholstery).

[0196] In one example, the trained classifier may include a neural network. The training may be "supervised," where the user provides feedback on the output of the trained classifier to improve the generated model. Alternatively, the training may be "unsupervised," where the trained classifier generates a model that reduces a given cost function (e.g., mean squared error).

[0197] A scale of damage related to which parts are affected is also calculated; for example, structural damage to the chassis is considered more serious than superficial damage to the bumper. If damage to a particular part, such as the fuel tank or brakes, is suspected, an output is generated indicating that it is dangerous to continue driving and that an assessment by a mechanic is necessary. Such scales also include information about the number and location of the damaged parts (so that the extent of the damage can be estimated).

[0198] Next, the damaged parts and corresponding repair costs and time are identified for each vehicle. Such identification is done using either publicly available cost data obtained via the internet, or data available to insurance companies with relationships and discount deals with parts manufacturers, for example. Labor costs and repair time are estimated using previous repair data, and these estimates may be improved by contacting repair shops.

[0199] In some embodiments, information regarding anticipated damage may be transmitted to a mechanic after a collision, who can then provide an estimated price based on this damage.

[0200] Third-party liability By using the characteristics of other vehicles involved in the collision, it is possible to estimate the likely responsibility. Telematics data is available for the vehicle's speed, mass, and direction, as well as the impulse it received during the collision. Therefore, the momentum of the other vehicles can be calculated. If the mass of the other vehicles is known, estimates of their velocities (speed and direction) can be generated.

[0201] The mass of other vehicles involved can be determined by obtaining their registration numbers, from which their type and model can be identified. The vehicle operator may be required to provide details of other vehicles involved in the collision, or this information may be obtained from reports issued by any party, such as police reports.

[0202] The speed of other vehicles (one or more) at the time of collision is important in identifying the party(s) at fault or responsible for the collision. For example, calculations may reveal that one party was speeding, or recorded GPS / video data may indicate that one party veered off the road.

[0203] Information from external sources such as road traffic laws, case law, and mapping information may be used in combination with telematics data and in-vehicle diagnostics. This may include methods for automatically classifying roads and junctions based on mapping data to determine priority. This may include methods for identifying the lane a vehicle is in and any lane change operations. This may include methods for identifying the vehicle's operational intent. This may include methods for distinguishing reverse movement from forward movement. These methods may also be applicable to third-party vehicles colliding with a vehicle recording telematics data.

[0204] Therefore, using (recorded and calculated) data from the telematics data of a single vehicle, a comprehensive model of the collision (albeit an estimate) can be created, effectively and immediately indicating repair costs and potential liability.

[0205] To determine the total amount to be billed, a similar repair cost calculation may be performed for each vehicle involved in the collision.

[0206] Personal injury claim The data obtained is also used to assess the likelihood and type of injury sustained. In particular, the measured acceleration is used to estimate the probability of whiplash. The combination of direction and magnitude may be relevant in determining whether a whiplash claim will result from the collision. For example, a "rear-end collision" may be more likely to result in a whiplash claim than a side collision of similar magnitude.

[0207] Calculations can be used to counter fraudulent claims; for example, a measured low acceleration can be used as evidence to counter a whiplash claim. The decision to contest a potentially fraudulent claim is likely to depend on the value of the claim and the probability of fraud, and this decision may be made automatically or by some input from the insurance agent, with smaller claim amounts being preferred to be handled with minimal agent input. For example, a claim that includes unexpected injuries or claims for injury costs far higher than estimated may trigger a notification informing the agent of the potential fraud.

[0208] When determining the monetary scale of a potential claim, precedents from past claims can be used. For example, if it is determined that there is a high probability that the vehicle's occupants suffered whiplash, the claim amount may be the average of previous whiplash claims multiplied by the number of vehicle occupants.

[0209] However, more intelligent methods may be employed, and by utilizing more information related to the collision, a more accurate personal injury claim can be calculated.

[0210] For example, past injury claims can be filtered to better represent the collision in question. Such comparable information may include the type and model of vehicles involved, the location and severity of the collision (e.g., based on accelerometer data), or whether airbags deployed. Filtering using as many collision features as possible may be possible until the number of past cases is no longer statistically significant.

[0211] In another example, a trained classifier could be used to quantify the likely level of compensation. In such an example, the trained classifier would be fed historical cases as training data, and thus could build a model of how collision features affect the amount of personal injury compensation paid.

[0212] In one example, the trained classifier may include a neural network. The training may be "supervised," where the user provides feedback on the output of the trained classifier to improve the generated model. Alternatively, the training may be "unsupervised," where the trained classifier generates a model that reduces a given cost function (e.g., mean squared error).

[0213] Total loss / repair assessment The calculated repair cost is used to assess the relative cost of repair and replacement. Collisions in which the repair cost (calculated as described above) exceeds the replacement cost result in the vehicle being immediately scrapped. Replacement costs can be determined from known information about the vehicle, such as its type, model, year, and mileage.

[0214] By using valuations to instantly predict claim costs, insurance companies can achieve savings by investigating only those collisions that are likely to be claimed or are likely to be fraudulent. This saves additional costs that would otherwise be incurred by involving multiple parties, such as insurance agents, mechanics, and doctors, in what would otherwise be low-value claims.

[0215] Immediate claim assessment allows insurance companies and vehicle operators to plan for the final claim, for example, by securing a certain amount of money or renting a replacement vehicle for an appropriate period.

[0216] Actions after evaluation Figure 11 shows a flowchart for determining whether repairing a vehicle is cost-effective.

[0217] 1. The impact is detected and evaluated, and from this evaluation, it is determined whether the vehicle is a total loss or repairable.

[0218] 2A. As a result of a collision causing total loss, the vehicle is replaced. This usually occurs when the severity of the collision exceeds a certain threshold.

[0219] 2B. If the damage caused by the collision is repairable, it will be assessed to determine whether the repair costs will exceed the value of the car. The value of the car may be what is available at any car dealer or may be a discounted price available from the insurance company.

[0220] 3A. If the repair cost exceeds the price of the car, the car will be replaced.

[0221] 3B. If the repair cost is lower than the price of the car, the car will be repaired.

[0222] Figure 12 shows a flowchart for determining compensation after a collision.

[0223] 1. The conflict is assessed to determine who is responsible.

[0224] 2A. If a vehicle, including data recording equipment, is found to be at fault, payment will be made to each party involved under the insurance contract.

[0225] 2B. If this vehicle is not at fault, the costs will be recovered from a third party (such as the insurance company of another vehicle).

[0226] 2C. If negligence is shared, appropriate payments will be made based on the shared responsibility.

[0227] Figure 13 shows a flowchart illustrating the transfer of data related to collision evaluation.

[0228] When a vehicle collision is detected, data related to this situation is recorded and stored using various data sources, and this data is then transmitted to a server.

[0229] The recorded data is used, along with other input data such as figures from the Book of Quantum Personal Insurance (Quantum PI), to assess vehicle collisions. This assessment partially includes steps to determine the severity and location of any impact, as well as to calculate several outputs.

[0230] These outputs are used to provide information related to the conflict, such as cost estimates and liability identification. This information can be sent to one or more parties, who can then use it, for example, to quantify insurance payouts.

[0231] Figure 14 shows a detailed example of data transfer related to collision evaluation.

[0232] 1. A collision is detected.

[0233] 2. The collision data is sent to the server (via satellite) in the form of a comma-separated variable (CSV) file.

[0234] 3. To create a collision model, collisions are evaluated using appropriate algorithms along with other data (location data, recorded vehicle data, and situational data such as weather).

[0235] 4. The data obtained using the model is used to predict the billing.

[0236] This predicted claim data is then used for liability identification, estimating repair costs / time, and assessing potential fraud.

[0237] In one example, liability is inferred by extracting relevant facts from telematics data and processing that data through a cognitive inference model. The model may include information regarding relevant road traffic regulations, case law, and mapping information. Examples of elements that may be considered by the model to identify liability include · classifying roads and junctions based on mapping data to identify precedence, · identifying the lane the vehicle is in and lane change operations, · identifying the vehicle's intent of operation, and · distinguishing reverse movement from forward movement, are included.

[0238] If further information is needed, relevant questions may be dynamically presented to the user. Once a possible liability result is identified, supporting data may be gathered to provide evidence to support the liability identification. This enables an objective analysis of liability and minimizes the risk of fraud or human error.

[0239] Collision evaluation interface Once a collision is detected and an initial automated assessment is made, it may be necessary to analyze the data in more detail to detect anomalies and identify liability. To simplify such processing and improve accuracy, an interface is disclosed that enables an agent to view events preceding the collision as well as the collision itself along with the situational information. This enables the agent to quickly understand how the collision occurred and how injuries / vehicle damage may have occurred.

[0240] Figure 15 shows an interface used to view data corresponding to multiple collisions.

[0241] Interface 4000 may be integrated with existing insurance software, for example, as a plug-in, or it may be provided as a standalone program or web service. This interface can receive data from existing services and existing databases and may have existing data recording devices used by insurance companies to record data.

[0242] Interface 4000 is displayed to the agent, and this interface includes the following:

[0243] - Map 4100 is displayed, and marker 4110 is overlaid on this map. These markers are related to collisions.

[0244] This interface can be used by agents to view multiple collisions, and a selection list is used to limit the accidents shown; for example, collisions can be filtered by date, location, or cost. Markers are color-coded, and this color relates to the characteristics of the collision; for example, collisions within the last week are a different color from even older collisions.

[0245] - Uploader 4200 is provided to upload new accident data, for example, by dragging a file onto the window or by browsing computer files and selecting a collision data file. Certain information is required, and it is impossible to upload a file without, for example, a vehicle identifier (e.g., registration number). If such an upload is attempted, an error notification will be transmitted to the party attempting the upload.

[0246] Uploads occur automatically, and in the event of a collision, it is preferable that the data file is automatically transmitted from the recording means and received by the server. The agent's device checks the server in the event of an event (such as when the agent logs in) and / or at predetermined intervals to detect newly received collision data and loads this data into interface 4000.

[0247] In some embodiments, there is a standard file format in which units are standardized so that, for example, unit indication is not required (i.e., all speeds are in miles per hour, so speed data only contains numbers). The data may be required to be in such a format before being uploaded, or the data may be converted to this format during upload.

[0248] - Further information about each collision is provided, and item 4400 is displayed, including information such as the registration plate number 4410, location 4420, date 4430 and time 4440, and a scale of the collision's age 4450. Other information may be displayed, and the agent can choose which information is displayed.

[0249] Collision markers and items are selectable, which shades the relevant items and changes the color of the markers. A box 4120 containing information related to the selected collision appears when the relevant marker is selected, so that agents can easily identify the collision items corresponding to each marker.

[0250] - A search box 4300 is provided, which allows agents to filter the accidents displayed. Accidents can be filtered by any recorded / calculated features or combination of features of the collision.

[0251] Figure 16 shows various forms of interfaces for viewing events leading up to a collision and events including the collision.

[0252] Within interface 4000, agents can select a conflict to view details.

[0253] Figure 16a shows a view of interface 4000, which displays the status at the start of data recording.

[0254] Interface 4000 includes the following:

[0255] - Identifier 4502, where the vehicle registration number is displayed, indicating which collision is being viewed. Additional situational information such as the location of the collision 4504 and the weather conditions at the time of the collision 4506 is also shown.

[0256] - Various views of the collision scene, such as an overhead view 4508, a "Street View" 4510 (ground-level view), and an in-vehicle view 4512, are viewable. In this embodiment, the means for selecting the in-vehicle view selection is disabled and it is "grayed out" to indicate that the in-vehicle view is not available. The in-vehicle view is only available if a dash cam or other image / video recording means is installed in the vehicle involved in the collision.

[0257] If an in-vehicle video is available, it can be viewed optionally as a video or as a frame with a timestamp, along with another view, and the relevant frames (those related to the event) are automatically detected.

[0258] There is an indication of which view is selected, such as colored / bold font or underlining.

[0259] - The selected view is displayed at 4520. This view is preferably obtained from one or more third-party mapping services such as Google Maps (trademark) or DigitalGlobe (trademark).

[0260] - A representation of the vehicle 4522 involved in the collision is overlaid on the displayed view 4520. Information about the current characteristics of the vehicle, such as speed, is shown along with the representation of the vehicle. An indicator of the vehicle's path 5424 is also shown.

[0261] The path 5424 can be based on recorded / memorized and calculated data, where the GPS locations recorded to obtain a predicted vehicle path can be used together with accelerometer data.

[0262] The millisecond resolution of the recorded / stored and calculated data (potentially through extrapolation or other fitting procedures) may allow minute vehicle movements to be calculated and overlaid on the mapping software.

[0263] This route 5424 can be adjusted by an agent or automatically. Such adjustments may be desired if there is a suspected error in the received GPS data, for example, if the data appears to indicate that the vehicle was driving through a forest adjacent to the road. The agent can adjust the recorded GPS track to obtain a more reliable route. This route can "snap" to the road, meaning that if the route is changed, the device may detect a fit between each point along the route and a point on the road and suggest such an arrangement. An indicator will then be displayed to show that the route has been adjusted.

[0264] - The road speed limit indicator 4582, and where this speed limit begins, are displayed in the selected view 4520.

[0265] - A means 4526 for displaying the vehicle's speed is displayed along with an indication 4528 of the speed limit of the road the vehicle was traveling on at the time of viewing.

[0266] Speed ​​indications may also show safe / dangerous speed ranges, for example, using a variable color scheme. A “dangerous” speed could be considered a speed exceeding the legal limit, or a speed that is above or below that limit by a certain percentage.

[0267] In some cases, the determination of a dangerous speed also takes into account other features of the route, such as rain, curves in the road, or nearby intersections, which may reduce the maximum safe driving speed, resulting in a “dangerous” driving speed below the legal limit.

[0268] - Timeline 4530 displays the period preceding and / or following a collision, along with event time indications 4532, 4534, and 4536. Events are selected based on recorded data, such as exceeding a g-force threshold, an operator making a phone call, an operator interacting with the vehicle (e.g., braking), or a vehicle turning onto a new road.

[0269] The location of each event is displayed as a symbol on the vehicle's route 4522, and in this case, location 4592, "sudden braking," is displayed.

[0270] - Further information related to each event is shown below the timeline, with each event displayed in separate items 4540, 4550, and 4560. These items include information related to the event 4542, 4552, 4562, and 4564. As a result of the collision, a whiplash indicator 4566 is shown here. Other consequence information, such as repair costs, or instructions for responsibility may also be displayed here.

[0271] These events are selectable (for example, by clicking on items 4540, 4550, and 4560 with the computer cursor), and such selection will display information and representations of the events.

[0272] A scrolling mechanism 4570 is also provided, which allows events to be selected over time following the currently selected event, so that the agent can scroll through events preceding and following a collision.

[0273] - An expandable billing window 4600 is provided, which is used to view information about possible claims. This will be explained in more detail in a later section of the description.

[0274] This expandable claims window 4600 allows agents to view conflict-related information to better understand liability before viewing information related to a claim. The expandable claims window 4600 may cover the entire interface 4000, or only certain sections to allow agents to view claim information along with an impact display (to better understand how any damage / injury was suffered).

[0275] Figure 16b shows a view of interface 4000 when the first event (sudden braking) is selected.

[0276] The map 4520, vehicle location 4522, vehicle characteristics (speed 4526), ​​situation (speed limit 4528), and timeline 4530 are updated within the interface so that the values ​​of these characteristics relate to those at the time of the (sudden braking) event. The selected event 4540 is highlighted.

[0277] Figure 16c shows a view of the agent interface 4000 when the second event (sharp cornering) is selected.

[0278] As before, all aspects of the interface are updated. The item events are also updated, and as a result, item 4550 of the second event is minimized (because it occurred before the selected event). An additional scrolling means 4572 is presented, allowing the user to choose to view previous events in chronological order.

[0279] Within the selected view 4520, the speed limit indicator 4584 is displayed and is "grayed out" to indicate that the vehicle has not yet reached the area where this limit applies. The current speed limit is also displayed at 4528.

[0280] Impact location 4596 is displayed on route 4524.

[0281] Figure 16d shows a view of interface 4000 when the third event (impact) is selected.

[0282] As before, all aspects of the interface will be updated. The speed limit indicator 4584 will no longer be grayed out, indicating that this speed limit applies to the vehicle at the selected time.

[0283] The location of the sharp corner 4594 is displayed on the route 4524 of the vehicle representation 4522.

[0284] The event items are also updated, resulting in the first event (sudden braking) no longer being displayed, and the previous event (sharp cornering) item 4550 being minimized. Since there are no subsequent events recorded, the initial scroll means 4570 (for selecting later events) is no longer displayed. The vehicle's speed at the point of impact, as well as the location of the impact, are indicated.

[0285] This time-series display allows agents to view events preceding and following a collision, along with situational data, to better understand the cause of the collision. This enables the quick and accurate identification of responsibility.

[0286] In some embodiments, an agent may annotate interface 4000 to display further information or the agent's opinion. This is useful if the same or another agent reviews the conflict at a later date. In particular, this is used to explain / summarize any aspect of the conflict for other parties, as the data may later be shared with colleagues, legal entities, or anyone involved in the conflict.

[0287] In some embodiments, a virtual reality model is created that precedes the collision, and the user can view (and / or hear) the event, including the collision.

[0288] Figure 17 shows a view of the interface for displaying a street-level map view of the location where the collision occurred.

[0289] Using the selection means 4510, "Street View" is selected. This displays a street-level image 4521 of the location, which does not need to be captured at the time of the collision, and therefore this view does not need to display local weather conditions or vehicles present at the time of the collision.

[0290] Such Street Views are used by agents to improve their understanding of collision situations. For example, Street Views that include intersections allow agents to understand priority and visibility at those intersections.

[0291] Figure 18 shows various views of the interface, including a section for viewing information related to potential claims.

[0292] Figure 18a shows a view of interface 4000, in which the billing section 4600 is used to display information related to potential claims.

[0293] The billing window 4600 includes the following:

[0294] - Information related to the claim, here showing the date of the collision 4602 and the time 4604.

[0295] - Damage assessment items 4610 are included, which display the predicted zones and severity 4612 of any vehicle damage, and the severity may be indicated using a binary system (in which the impacted zones are shown) or a higher-resolution measure such as a color gradient. Damage display may use higher resolution and show a larger number of zones. Damage shown here relates to a vehicle including data recording means. 2D visualizations are shown, but it is also possible to draw 3D visualizations of damage to a vehicle based on a known vehicle type, model, color, and impact area(s). 3D visualizations may make it easier for the user to virtually "inspect" the vehicle.

[0296] In some embodiments, a virtual reality model may be generated and sent to a mechanic for use in assessing a collision.

[0297] A collision characterization 4614, information about the damaged vehicle 4616, and an estimated repair cost 4618 are presented. This repair cost depends on the type of vehicle and the location / severity of the impact.

[0298] - There is an injury item 4620 that displays the estimated cost of an injury claim 4626. Such claims depend on the number of passengers 4622, and means 4624 are provided to change this number. An estimated total claim amount is determined from the estimated injury claim amount per passenger, which may be set by regulation.

[0299] - There is a third-party vehicle item 4630, which includes an input field 4632 into which information about other vehicles involved in the collision is entered.

[0300] - This information is preferably a registration number, which is used to identify the vehicle's type, model, dimensions, and mass (by a vehicle registration search service). It is advantageous that the registration number can be easily obtained from the occupants of the vehicles involved in the collision or from the accident report.

[0301] There may be multiple items 4630 or multiple input fields 4632 into which information about multiple other vehicles can be entered.

[0302] The collision calculation uses information about other vehicles involved to estimate the collision parameters, and the mass of the other vehicles involved is used to obtain the speed of those vehicles at the time of the collision. Subsequently, data relating to the type and model (and any modifications) of each other vehicle is used to estimate the repair costs of these vehicles.

[0303] Information necessary for cost estimation or collision assessment, such as the number of passengers or the registration numbers of any other vehicles involved, - Data recording means, - Video recordings that allow the registration number recognition system to detect the registration number without using input from an agent. - Vehicle-related sensors such as pressure sensors or seat belt detectors that may be used to select the number of passengers or predict injuries, - Vehicle occupants, - Third-party reports such as accident reports from emergency services, It is obtained from.

[0304] In some embodiments, when a collision is detected, a notification is automatically sent to the occupants, and necessary information is requested. The occupants may also have the opportunity to provide further information, such as a recollection of the collision.

[0305] Figure 18b shows a view of interface 4000, where the number of passengers 4622 within injury item 4620 has been updated, and different estimates of potential injury claim amounts 4626 have been obtained.

[0306] Figure 18c shows window 4700, which is used to view the details of the calculation of the estimated repair costs.

[0307] This type of window is opened by clicking on the estimated repair cost of 4618 provided by the agent.

[0308] The parts that need repair are predicted from the collision assessment, the price of replacement parts related to the vehicle is estimated, and the total estimated repair cost is calculated by including painting costs and labor costs. This depends on the location and time of the collision, where mechanics may be more expensive than elsewhere, and there may be peak seasons when the price of mechanics or parts increases.

[0309] Figure 18d shows a view of interface 4000, which has information related to another vehicle involved in the collision.

[0310] Third-party vehicle item 4630 here shows the type / model of other vehicles 4638 involved in the collision, as well as indicators used to assess responsibility. Here, the speed of the other vehicles 4634 at the time of impact is shown, along with indicators, e.g., color coding, that show the speed limit 4636 and recommended safe speed.

[0311] Speed ​​4634 is obtained using known characteristics of the vehicle, including data recording means, as well as the mass and dimensional profiles of other vehicles obtained from the entered registration number (as described above).

[0312] In some embodiments, it is possible to incorporate data from any other involved vehicles, or any vehicles in a similar geographical location, which may also be recording telematics, audio, or visual data. Where available, this data can be used to obtain a more accurate model of the collision.

[0313] In some embodiments, there is an option to include additional data to obtain a more accurate estimate of the speed in question, for example, the load capacity of other vehicles, where a fully loaded vehicle will be much heavier than an unloaded vehicle.

[0314] In some embodiments, a range of evaluations are performed, and a range of possible evaluation inputs are used to account for uncertainties, such as the uncertainty of the load capacity of the vehicles involved. This can be used, for example, to obtain worst-case, best-case, and / or most likely evaluations / costs.

[0315] In some embodiments, the third-party vehicle item 4630 also displays an estimate of the vehicle repair costs and / or injury claim amount.

[0316] In some embodiments, indicators of any other involved vehicles may be included on the map, and third-party vehicles may be displayed along with estimated routes or speeds, as well as vehicles that include data recording means.

[0317] Alternative and modified examples For example, various other modifications will be obvious to those skilled in the art.

[0318] The detailed explanation primarily considers the use of an interface by an (insurance) agent. Such an interface may be used by any user, for example, a party involved in a collision or a mechanic. The interface or the data contained within it may be used jointly so that multiple users can input and view information. The interface shown may be tailored to the party viewing it; for example, vehicle occupants may only be able to view the items for which they need information (regarding the number of passengers or other vehicles involved), and a mechanic may only be able to view a list of damaged parts.

[0319] The detailed explanation primarily considers collisions in which two vehicles are involved in a single impact, but since a collision can involve any number of vehicles or impacts, a collision may be detected when one vehicle impacts one or more vehicles, and separate impacts with the same vehicle or a different vehicle may be detected as events within the same collision or as separate collisions.

[0320] The detailed description primarily considers the use of this method for evaluating collisions, including impacts. The method can similarly be used to evaluate any driving event; for example, a sudden braking event may trigger data storage, where for the purposes of this disclosure, such an event is considered a collision. This data can be used, for example, to evaluate driver performance, and the provided interface can be used to better assess whether the vehicle operator made any dangerous decisions or whether a particular event was caused by factors beyond the operator's control. Such evaluations can be used in insurance decisions, and triggering numerous evaluations may result in increased insurance premiums.

[0321] The methods and systems provided herein can be used in conjunction with any method for estimating impact severity. While the methods presented here are useful for simple estimates where only the characteristics of one vehicle are known, other methods may be used, and collisions can be characterized more accurately if more information is available. Such methods may include computer simulations, such as finite element analysis (FEA) or Monte Carlo simulations. In these methods, objects can be modeled with arbitrary precision; vehicles can be modeled using point sources, or detailed dimensional and mass distribution models can be used. The model used may depend on the available information or the deadline for the severity estimate, and more accurate methods may require considerable computation time.

[0322] In some embodiments, the amount of data recorded during normal driving may decrease, and the amount of data recorded during events may increase thereafter. For example, only acceleration may be recorded during normal driving, and other data may be recorded as a result of events, including exceeding a threshold acceleration. However, this method risks missing potentially important events leading to a collision.

[0323] The methods and systems provided herein can be used in conjunction with autonomous vehicles, and costs and responsibilities may be identified as disclosed. These methods and systems can also be used to assess the performance of autonomous features and may be capable of identifying errors in the control systems of autonomous vehicles. In such use cases, any sensor / driving data from the autonomous vehicle may be used as “telematics data” as referred herein.

[0324] Any processing involving persons disclosed herein may also be performed by a processor, such that any communication or identification occurs using a processor, and these identifications may employ machine learning or artificial intelligence to improve accuracy. Any processing performed by a processor may be transmitted to a person for approval.

[0325] The detailed explanation primarily examines methods used in relation to vehicle collisions, but these methods can be used whenever a threshold scale related to the recorded data is exceeded. For example, such methods can be used in sports where injury or negligence can be identified, or as an injury detection system where a smartphone can detect an injury and notify a third party as appropriate. A specific application could be during skiing, for example, where a collision or fall is detected, resulting injuries are predicted, and if there is a high probability of serious injury, emergency services can be contacted.

[0326] Similar to the explanation regarding vehicles, telematics data from a user's mobile phone could be used to detect collisions and identify responsibility for likely collisions. In the skiing example, it would be possible to identify the type of "course" the user was on (green, red, black, or "off-piste"), the speed at which the skier was skiing leading up to the collision, and whether an object or other skiers were hit. This could affect the level of compensation provided, for example, if the insurance policy does not cover "off-piste" skiing.

[0327] The present invention has been described above merely as an example, and it will be understood that further modifications may be made within the scope of the invention.

[0328] The reference numbers appearing in the claims are illustrative only and do not limit the claims.

Claims

1. A method for analyzing collision data using an analytical device that analyzes collision data, The steps include receiving telematics data of the first vehicle related to the collision of the first vehicle, A step of evaluating the driving of the second vehicle involved in the collision using the telematics data of the first vehicle, Using data on the velocity, mass, and direction of the first vehicle from the telematics data of the first vehicle, the impulse received by the first vehicle during the collision is determined, and the momentum of the second vehicle is determined from the impulse received by the first vehicle. Data relating to the second vehicle is acquired, and the mass of the second vehicle is determined based on the data of the second vehicle. The speed of the second vehicle is determined based on the momentum and mass of the second vehicle. A step of evaluating the operation of the second vehicle involved in the collision, A step of determining the severity of the collision based on the telematics data of the first vehicle, A step of notifying the subject of the collision based on the severity of the identified collision, Methods that include...

2. The method according to claim 1, further comprising the step of determining whether the second vehicle was traveling at speed prior to the collision.

3. The method according to claim 1 or 2, further comprising the step of identifying a measure of responsibility.

4. The method according to any one of claims 1 to 3, comprising the step of modeling the collision with a resolution of one second or less.

5. The step further includes identifying the impact area based on the telematics data of the first vehicle received, The impact area includes one or more zones related to the vehicle or one or more points on the vehicle. The method according to any one of claims 1 to 4, wherein the step of identifying the impact area includes the step of identifying the rotation of the vehicle and / or the partial or complete inversion of the vehicle.

6. The method according to any one of claims 1 to 5, further comprising the step of identifying one or more events related to the collision, wherein the event is one or more of sudden braking, sudden cornering, sudden acceleration, sudden deceleration, rotation of the vehicle, partial or complete reversal of the vehicle, loss of traction, and a secondary collision.

7. The method according to any one of claims 1 to 6, further comprising at least one of the steps of: determining the repair cost based on the cost of parts in the impact area; determining the repair cost based on the likelihood that the parts are damaged, depending on telematics data of the first vehicle received; and determining the repair cost based on a comparison with past impact data.

8. The method according to any one of claims 1 to 7, wherein the subject is one or more of the following: emergency services, the occupants of the vehicles involved in the collision, and the insurance entity, if the severity of the collision exceeds a threshold level.

9. The method according to any one of claims 1 to 8, further comprising the step of determining the probability of injury and / or injury compensation costs depending on one or more of the magnitude of the collision, the direction of the collision, the rotation of the vehicle, the partial or complete reversal of the vehicle, and the probability of whiplash.

10. The method according to any one of claims 1 to 9, wherein the telematics data of the first vehicle includes data from an accelerometer and / or GPS.

11. The method according to any one of claims 1 to 10, wherein the telematics data of the first vehicle includes data from an accelerometer and GPS, and the method further includes the step of interpolating accelerometer and GPS data points to identify the vehicle trajectory.

12. The method according to any one of claims 1 to 11, further comprising the step of determining a measure of responsibility based on one or more of the following: priority, road or junction classification, lane in which a vehicle is located, lane change operation, measure of operational intent, reversing movement, road traffic regulations, case law, mapping information, and cognitive reasoning models.

13. The method according to any one of claims 1 to 12, wherein the step of determining the severity of the collision includes one or more of the steps of determining a change in the momentum of the objects involved in the collision, and determining a change in momentum by integrating the acceleration over a time window before and after the determined impact time.

14. A device for analyzing collision data, Receiving means for receiving telematics data of the first vehicle related to the collision of the first vehicle, An analysis means for identifying collision data based on the telematics data of the first vehicle and for evaluating the driving of the second vehicle involved in the collision using the telematics data of the first vehicle, Using data on the velocity, mass, and direction of the first vehicle from the telematics data of the first vehicle, the impulse received by the first vehicle during the collision is determined, and the momentum of the second vehicle is determined from the impulse received by the first vehicle. Data relating to the second vehicle is acquired, and the mass of the second vehicle is determined based on the data of the second vehicle. The speed of the second vehicle is determined based on the momentum and mass of the second vehicle. An analysis means for evaluating the operation of the second vehicle involved in the collision, Includes an interface module for providing the aforementioned collision data to the user, The collision data is the severity of the collision, and the interface module is for notifying the subject of the collision based on the identified severity of the collision. Device.

Images