METHOD AND SYSTEM FOR ESTIMATING AN ACCIDENT RISK OF AN AUTONOMOUS VEHICLE

• US 20190308619A1
• Filed: 04/10/2019
• Published: 10/10/2019
• Est. Priority Date: 04/10/2018
• Status: Active Grant

• Current Assignee: Walter Steven Rosenbaum
• Original Assignee: Walter Steven Rosenbaum
• Inventors: ROSENBAUM, WALTER STEVEN
• Granted Patent: US 11,407,410 B2
• Time in Patent Office: Days
• CPC Class Codes: B60W 2556/05 Big data
• CPC Class Codes: B60W 30/0953 the prediction being responsive to vehicle dynamic parameters
• CPC Class Codes: B60W 40/09 Driving style or behaviour
• CPC Class Codes: G05D 1/0088 characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours using knowledge based models G06N5/00
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: G06Q 50/40 Business processes related to the transportation industry shipping G06Q10/083

Abstract

A method for estimating an accident risk of an autonomous driving unit produces helpful results with fewer autonomous driving cycles. From driving values, which have been determined from monitoring at least one driving parameter of the driving unit during autonomous driving, an autonomous-driving quantity is determined quantifying an autonomous-driving quality of the driving of the driving unit. The autonomous-driving quantity is associated with a plurality of manual-driving quantities, which have been determined from the same driving parameter during manual driving periods of different driving units. An autonomous-driving accident rate value is determined from accident rate values associated to those manual-driving quantities.

Claims

1. A computer-implemented method for estimating an accident risk of an autonomous driving unit, the method comprising:

monitoring with a sensor in the driving unit at least one driving parameter of the driving unit during autonomous driving and determining driving values with an evaluation unit;

determining an autonomous-driving quantity which quantifies an autonomous-driving quality of the driving of the driving unit;

associating the autonomous-driving quantity with a plurality of manual-driving quantities, which have been determined from the at least one driving parameter during manual driving periods of different driving units; and

determining an autonomous-driving accident rate value from accident rate values associated with the manual-driving quantities.

2. The method according to claim 1, which comprises deriving the manual-driving quantities in the same way as the autonomous-driving quantity.

3. The method according to claim 1, wherein a driving unit is a combination of a specific driver with a specific vehicle, the driver of the autonomous driving unit is an artificial intelligence driving software of the autonomous driving unit, and the drivers of the different driving units are humans.

4. The method according to claim 1, which comprises generating driving values for a plurality of driving units from manual driving over a predefined period, and determining for each driving unit a manual-driving quantity from the driving values gathered over the entire predefined period.

5. The method according to claim 1, wherein each of the manual-driving accident rate values are a mileage per accident and are determined from an accident statistic of one driving unit.

6. The method according to claim 1, which comprises determining from one or more driving parameters being monitored during manual driving of the different driving units a plurality of manual-driving quantities of different categories from which only those are chosen for association with the autonomous-driving quantity which belong to the same category as the autonomous-driving quantity.

7. The method according to claim 6, wherein the categories are a plurality of different locale categories where the driving parameter was monitored.

8. The method according to claim 1, which comprises selecting only those manual-driving quantities for association with the autonomous-driving quantity which have a bad driving rating value below a predefined threshold compared with a group of manual-driving quantities.

9. The method according to claim 1, wherein the autonomous-driving quantity is associated with the manual-driving quantities only if the autonomous-driving quantity has a bad driving rating value below a predefined threshold compared with a group of other autonomous-driving quantities.

10. The method according to claim 1, which comprises selecting from a multitude of manual-driving quantities for association with the autonomous-driving quantity only those manual-driving quantities which show a predetermined degree of similarity with the autonomous-driving quantity.

11. The method according to claim 10, wherein the predetermined degree of similarity is a value difference of the autonomous-driving quantity to the manual-driving quantity below a predefined threshold.

12. The method according to claim 1, which comprises selecting only those manual-driving quality metrics for deriving the autonomous driving accident risk which show a predetermined degree of statistical similarity with the autonomous-driving quantity, wherein the degree of statistical similarity is tested by Null-Hypothesis.

13. The method according to claim 1, which comprises determining the autonomous-driving quantity and the manual-driving quantities by the following steps:

during operation of the driving unit the driving speed of the driving unit is determined at random instants of time;

the locations of the driving unit is determined for those time points where the speed is determined;

the driving unit speed values are compared with a geostationary speed value associated to the respective locations;

a speed difference between the driving speed and the geostationary speed value is determined; and

a plurality of such resulting values are combined to the quantity.

14. The method according to claim 1, which comprises determining the autonomous-driving quantity and the manual-driving quantities by the following steps:

during operation of the driving unit a speedup acceleration of the driving unit is detected by an acceleration sensor in the driving unit;

wherein a maximum speed, which occurs during or without deceleration after an acceleration event in which the driving unit acceleration is above a predetermined threshold, is stored as an acceleration event value;

the acceleration event value is compared with a geostationary speed value of a location where the acceleration event value occurred; and

a plurality of such resulting values are combined to the quantity.

15. The method according to claim 1, which comprises determining the autonomous-driving quantity and the manual-driving quantities by the following steps:

during operation of the driving unit a deceleration of the driving unit is detected by an acceleration sensor in the driving unit;

wherein from a speed difference between the speed occurring at the beginning and the speed at the end of an acceleration event in which the driving unit deceleration is above a predetermined threshold a resulting value is determined; and

a plurality of such resulting values are combined to the quantity.

16. The method according to claim 1, which comprises determining the autonomous-driving quantity and the manual-driving quantities by the following steps:

during operation of the driving unit a positive and negative acceleration of the driving unit, each above a predetermined threshold, are detected by an acceleration sensor in the driving unit;

a time span between the positive and negative acceleration is stored as an erratic driving event value or resulting value, respectively; and

a plurality of such resulting values are combined to the quantity.

17. The method according to claim 1, which comprises determining a plurality of autonomous-driving quantities for different quantity categories, and associating each autonomous quantity with a plurality of manual-driving quantities of the same category of different driving units, and determining a respective autonomous-driving accident rate value for each quantity category.

18. The method according to claim 17, which comprises determining an overall autonomous-driving accident rate value from a combination of the autonomous-driving accident rate values which are weighted according to respective driving distance in the combination.

19. A system for estimating an accident risk of an autonomous driving unit, the system comprising:

a sensor system of the driving unit configured for monitoring a driving parameter during a driving period driven autonomously by the driving unit;

an evaluation unit configured to receive from said sensor system signals related to the driving parameter, said evaluation being configured for determining from the driving parameter a quantity quantifying a quality of an autonomous driving of the driving unit, for relating the autonomous-driving quantity to a plurality of manual-driving quantities, which have been determined from the same driving parameter during manual driving periods of different driving units, and for determining from accident rate values associated to the manual-driving quantities an autonomous driving accident rate value.