System and method for sensor fusion based collision and quadrant detection, and severity of accident determination
The present disclosure discloses systems and methods that can be implemented on the edge system in a vehicle for near real-time detection of collisions with multi-level confidence (i.e., low confidence, moderate confidence, high confidence), automated gathering and report generation of collision information required for post-accident analysis in a manner suitable for insurance company process workflow and storing the confidential collision details securely on the vehicle (edge) system. Confidence is determined by preconfigured adaptive thresholds measured by sensor data.
1 . One or more non-transitory computer readable media storing computer-executable instructions that upon execution cause one or more processors on a vehicle to perform acts comprising:
activating an Inertial Measurement Unit Sensor Processor (IMUSP) at a beginning of a trip in the vehicle for collection of IMUSP data;
activating a Time to Collision Sensor Processor (TTCSP) at the beginning of the trip for collection of TTCSP data wherein the TTCSP data comprises distance measurements and velocity measurements from one or more external-facing sensors used to compute a Time to Collision (TTC) value or a Distance to Collision (DTC) value for one or more objects external to the vehicle;
analyzing, on an IMU Accident Detector (IMUAD), the collected IMUSP data for an abnormal event exceeding a first preconfigured threshold;
creating an IMUAD record using the analyzed IMUSP data;
transmitting, upon detection of the abnormal event, a first accident confidence flag to an Accident Confidence Analyzer (ACA) for further analysis;
analyzing, on a TTC Data Collector (TTCDC), the collected TTCSP data for the Time to Collision (TTC) value or the Distance to Collision (DTC) value below a second preconfigured threshold, wherein the TTC value or the DTC value indicates whether a collision with the one or more objects external to the vehicle is imminent before the collision occurs;
creating a TTCAD record using the analyzed TTCSP data;
analyzing, by the ACA, IMUAD and TTCAD records; and
transmitting, upon detection of the TTC value or the DTC value below the second preconfigured threshold, a second accident confidence flag to an Accident Confidence Analyzer (ACA), wherein the first accident confidence flag is generated by the IMUAD based on the IMUSP data and the second accident confidence flag is generated based on the TTCSP data, such that the ACA receives independent confidence flags from both an IMU-based detection path and a TTC-based detection path.
2 . The one or more non-transitory computer-readable media of claim 1 , wherein the IMUSP and the TTCSP operate continuously during the trip.
3 . The one or more non-transitory computer-readable media of claim 2 , wherein the TTCDC captures the collected TTCSP data at a high frequency when the time to collision or distance to collision are below a preset threshold.
4 . The one or more non-transitory computer-readable media of claim 1 , wherein a confidence of an occurrence of an abnormal event is Moderate when an accident confidence from the TTCAD record is high and the first accident confidence flag from the IMUAD for a corresponding quadrant has not been detected.
5 . The one or more non-transitory computer-readable media of claim 1 , wherein a confidence of an occurrence of an abnormal event is Moderate when an accident confidence from the IMUAD is high and the first accident confidence flag from the TTCAD for a corresponding quadrant has not been detected.
6 . The one or more non-transitory computer-readable media of claim 1 , wherein a confidence of an occurrence of an abnormal event is Low when an accident confidence from the IMUAD is Moderate or Low and the first accident confidence flag from the TTCAD for a corresponding quadrant has not been detected.
7 . The one or more non-transitory computer-readable media of claim 1 , wherein there is no recorded occurrence of an abnormal event in the ACA when the accident confidence from the TTCAD is Moderate and there is no accident confidence flag from the IMUAD for the corresponding quadrant.
8 . The one or more non-transitory computer-readable media of claim 1 , wherein the IMUAD and the TTCAD operate continuously.
9 . The one or more non-transitory computer-readable media of claim 1 , wherein the TTCSP data and the IMUSP data are received from one or more internal sensors of the vehicle.
10 . The one or more non-transitory computer-readable media of claim 9 , wherein the one or more internal sensors includes data from at least one of before, during, and after the abnormal event.
11 . The one or more non-transitory computer-readable media of claim 9 , wherein the one or more internal sensors are located in a left quadrant, a right quadrant, a front quadrant, and a rear quadrant of the vehicle.
12 . The one or more non-transitory computer-readable media of claim 11 , wherein multiple collisions in different quadrants are detected simultaneously.
13 . The one or more non-transitory computer-readable media of claim 1 , wherein an occurrence of the abnormal event is high when the first accident confidence flag from the IMU-based detection path and the second accident confidence flag from the TTC-based detection path are both received by the ACA, thereby corroborating detection of the abnormal event through independent sensor modalities.
14 . The one or more non-transitory computer-readable media of claim 13 , wherein the ACA waits to receive records from both the TTCAD and IMUAD for a preset threshold before determining the confidence level of the abnormal event.
15 . The one or more non-transitory computer-readable media of claim 13 , wherein the ACA utilizes the IMUAD data and the TTCAD data to determine the confidence level of the abnormal event.
16 . The one or more non-transitory computer-readable media of claim 15 , further comprising:
determining, by the ASA, a severity of damage caused by the abnormal event, using data from the ACA, IMUAD, TTCAD, IMUSP, and TTCSP.
17 . The one or more non-transitory computer-readable media of claim 16 , wherein a severity of damage to an ego vehicle is classified as Low, Moderate, or High.
18 . The one or more non-transitory computer-readable media of claim 17 , wherein a severity of injury to passengers in an ego vehicle is classified as Low, Moderate, or High.
19 . The one or more non-transitory computer-readable media of claim 18 , wherein determining the severity of injury to the passengers in the ego vehicle includes use of seatbelts by the passengers in the ego vehicle.
20 . The one or more non-transitory computer-readable media of claim 16 , wherein the calculation of severity of injury to passengers in an ego vehicle is performed on a local device in or on the vehicle.
21 . The one or more non-transitory computer-readable media of claim 17 , wherein a target object type is used in determining a severity of damage.
22 . The one or more non-transitory computer-readable media of claim 17 , wherein a severity of injury to the target vehicle's passengers is classified as Null, Low, Moderate, or High.
23 . The one or more non-transitory computer-readable media of claim 17 , wherein the calculation of severity of injury to passengers in a target vehicle is performed on a local device in or on the vehicle.
24 . The one or more non-transitory computer-readable media of claim 23 , wherein the calculation of severity of injury to passengers in a target vehicle is performed on a local device in or on the vehicle.
25 . The one or more non-transitory computer-readable media of claim 23 , wherein the severity of damage is classified as Low, Moderate, or High.
26 . The one or more non-transitory computer-readable media of claim 25 , wherein the calculation of severity of damage is performed on a local device in or on the vehicle.
27 . The one or more non-transitory computer-readable media of claim 25 , wherein the calculation of severity of damage is performed in or on the vehicle.
28 . The one or more non-transitory computer-readable media of claim 25 , further comprising generating a collision report using data from the ACA, TTCSP, IMUSP, TTCDC, TTCSP, IMUAD, and TTCAD.
29 . The one or more non-transitory computer-readable media of claim 28 , wherein the collision report is generated by a local device in or on the vehicle.
30 . The one or more non-transitory computer-readable media of claim 28 , further comprising storing the collision report to a local storage device in or on the vehicle.
31 . The one or more non-transitory computer-readable media of claim 28 , further comprising transmitting the collision report to a connected user device.
32 . The one or more non-transitory computer-readable media of claim 28 , further comprising transmitting the collision report to a cloud.