Method and system for vehicular collision reconstruction
A system for accident reconstruction can include and/or be configured to interface with any or all of: a set of models, a set of modules, a processing system, client application, a user device (equivalently referred to herein as a mobile device), a set of sensors, a vehicle, and/or any other suitable components. A method for accident reconstruction includes collecting a set of inputs; detecting a collision and/or one or more features of the collision; reconstructing the collision; and producing an output based on the reconstruction. Additionally or alternatively, the method can include training a set of models and/or modules, and/or any other suitable processes.
1. A method for collision detection and reconstruction, the method comprising:
at a remote computing system in communication with a mobile device associated with a driver, wherein the driver is arranged within a vehicle, receiving a set of inputs, wherein the set of inputs comprises:
a location dataset collected at a location sensor of the mobile device;
a motion dataset collected at a motion sensor the mobile device;
a pressure dataset collected at a pressure sensor of the mobile device;
processing the set of inputs with a model, wherein the model is a machine learning model, wherein at least a portion of the machine learning model is a deep learning model, wherein the model comprises a set of algorithms configured to produce a set of probabilistic outputs, the set of algorithms comprising:
a collision detection algorithm configured to produce a collision parameter;
a severity detection algorithm configured to produce a severity parameter; and
a direction of impact detection algorithm configured to produce a direction of impact parameter;
processing a set of modules based on the set of probabilistic outputs to determine a set of module outputs, wherein the set of modules comprises:
a confidence module;
a fraud module;
a direction of impact module; and
a severity module;
implementing an emergency response based on at least one of the collision detection algorithm and the confidence module;
providing a subset of module outputs to an insurance entity, wherein the subset of module outputs comprises outputs from at least one of the fraud module, the direction of impact module, and the severity module; and
updating at least one of the set of algorithms and the set of modules based on the set of module outputs;
further comprising determining a collision risk score associated with a road segment based on at least one of the set of module outputs, wherein the road segment is arranged proximal to the location of the collision.
2. The method of claim 1 , further comprising processing the collision detection algorithm multiple times.
3. The method of claim 2 , wherein each of the severity detection algorithm and the direction of impact detection algorithm is performed a single time.
4. The method of claim 1 , wherein a first subset of the set of modules comprises machine learning models and wherein a second subset of the set of modules comprises classical rule-based models.
5. The method of claim 4 , wherein the second subset comprises the fraud module.
6. The method of claim 1 , further comprising:
receiving a second set of inputs associated with a second driver; and
processing the second set of inputs with an updated set of algorithms, the updated set of algorithms determined based on the first driver.
7. The method of claim 1 , wherein a second portion of the model implements traditional machine learning.
8. The method of claim 7 , wherein the traditional machine learning comprises a gradient boosting machine.
9. The method of claim 1 , wherein the model is further configured to produce as outputs a set of auxiliary outputs, wherein the set of auxiliary outputs comprises a set of gravitational force equivalent values and a speed of the vehicle just before the collision.