CN-114684053-B - Apparatus and method for controlling an airbag of a vehicle

CN114684053BCN 114684053 BCN114684053 BCN 114684053BCN-114684053-B

Abstract

An apparatus and method for controlling an airbag of a vehicle can ensure robustness of airbag deployment logic and more effectively protect a passenger. The apparatus and method achieve this by determining whether to deploy an airbag based on a human injury posterior probability calculated through a human injury probability model and bayesian network learning (feedback learning). The apparatus includes a human body injury probability calculator configured to calculate a human body injury conditional probability and a human body injury prediction probability based on vehicle motion information measured by the sensing apparatus, a learner configured to calculate a human body injury posterior probability by performing probability-based real-time feedback machine learning based on the human body injury conditional probability and the human body injury prediction probability, and an airbag deployment determiner configured to determine whether to deploy an airbag based on the human body injury posterior probability.

Inventors

LI SHENGYU
Hong Xihao
JIN SHENGZHU

Assignees

现代自动车株式会社
起亚株式会社

Dates

Publication Date: 20260505
Application Date: 20211209
Priority Date: 20201229

Claims (18)

1. An apparatus for controlling an airbag of a vehicle, the apparatus comprising: A human body injury probability calculator configured to calculate a human body injury condition probability and a human body injury prediction probability, respectively, based on the vehicle motion information measured by the sensing device; a learner configured to calculate a human injury posterior probability by performing probability-based real-time feedback machine learning based on the human injury conditional probability and the human injury prediction probability, and An airbag deployment determiner configured to determine whether to deploy an airbag based on the human injury posterior probability.
2. The apparatus of claim 1, wherein the learner is configured to: calculating a human body injury prior probability based on the human body injury conditional probability obtained in the previous time and the human body injury prediction probability obtained in the previous time, and The human body injury posterior probability is calculated by multiplying the human body injury conditional probability obtained through current calculation by the human body injury prior probability.
3. The apparatus of claim 2, wherein the probability-based real-time feedback machine learning is configured to update the human injury prior probability by feeding back a current human injury posterior probability to a previous human injury prior probability.
4. The apparatus of claim 2, wherein the human injury conditional probability is configured to be calculated by equation 1: [ formula 1] , wherein, Is a human injury prediction probability at a current time point (t) of the actual human injury probability at a previous time point (t-1) according to the measured value of the sensing device, Is based on the human injury prediction probability predicted by simulation of the measured passenger injury, Is the actual human injury event that occurs at the current point in time (t), Is the measurement of the sensing device at the current point in time (t), Is the actual human injury event occurring at the previous time point (t-1), and Is a human injury prediction event occurring at the current point in time (t).
5. The apparatus of claim 2, wherein the prior probability of human injury is configured to be calculated using equation 2: [ formula 2] , wherein, Is a human injury prediction probability at the current time point (t) and a measured value of the sensing device according to the human injury probability at the previous time point (t-1), Is the posterior probability of human injury at the previous time point (t-1), Is the actual human injury event that occurs at the current point in time (t), Is the measurement value of the sensing device at the current point in time (t), and Is the actual human injury event that occurred at the previous time point (t-1).
6. The apparatus of claim 2, wherein the human injury posterior probability is configured to be calculated using equation 3: [ formula 3] , wherein, Η is the normalization factor and is a factor of the normalization, Is based on the human injury prediction probability predicted by simulation of the measured passenger injury, Is the prior probability of human body injury, Is the actual human injury event occurring at the current point in time (t), and Is a human injury prediction event occurring at the current point in time (t).
7. The apparatus of claim 1, wherein the airbag deployment determiner is configured to determine to deploy the airbag based on the human injury posterior probability exceeding a preset reference value.
8. The apparatus of claim 1, wherein the vehicle motion information includes an acceleration value and an angular velocity value, a collision value, a pressure value, a roll value, a pitch angle value, and a yaw angle value of the vehicle.
9. A method of controlling an airbag of a vehicle, the method comprising: A human body injury probability calculator for calculating a human body injury condition probability and a human body injury prediction probability based on the vehicle motion information measured by the sensing device; Calculating, by a learner, a human injury posterior probability by performing probability-based real-time feedback machine learning based on the human injury conditional probability and the human injury prediction probability, and Whether to deploy the airbag is determined by an airbag deployment determiner based on the human injury posterior probability.
10. The method of claim 9, further comprising: calculating, by the learner, a human injury prior probability based on the human injury conditional probability obtained in the previous time and the human injury prediction probability obtained in the previous time, and And the learner calculates the human injury posterior probability by multiplying the human injury prior probability by the human injury conditional probability calculated currently.
11. The method of claim 10, wherein the probability-based real-time feedback machine learning is configured to update the human injury prior probability by feeding back a current human injury posterior probability to a previous human injury prior probability.
12. The method of claim 10, wherein the human injury conditional probability is configured to be calculated by equation 1: [ formula 1] , wherein, Is a human injury prediction probability at a current time point (t) of the actual human injury probability at a previous time point (t-1) according to the measured value of the sensing device, Is based on the human injury prediction probability predicted by simulation of the measured passenger injury, Is the actual human injury event that occurs at the current point in time (t), Is the measurement of the sensing device at the current point in time (t), Is the actual human injury event occurring at the previous time point (t-1), and Is a human injury prediction event occurring at the current point in time (t).
13. The method of claim 10, wherein the prior probability of human injury is configured to be calculated using equation 2: [ formula 2] , wherein, Is a human injury prediction probability at the current time point (t) and according to the measured value of the sensing device of the human injury probability at the previous time point (t-1), Is the posterior probability of human injury at the previous time point (t-1), Is the actual human injury event that occurs at the current point in time (t), Is the measurement value of the sensing device at the current point in time (t), and Is the actual human injury event that occurred at the previous time point (t-1).
14. The method of claim 10, wherein the human injury posterior probability is configured to be calculated using equation 3: [ formula 3] , wherein, Η is the normalization factor and is a factor of the normalization, Is based on the human injury prediction probability predicted by the simulated measurement of the injury of the passenger, and Is the prior probability of human body injury, Is the actual human injury event occurring at the current point in time (t), and Is a human injury prediction event occurring at the current point in time (t).
15. The method of claim 9, wherein the airbag deployment determiner is configured to determine to deploy the airbag based on the human injury posterior probability exceeding a preset reference value.
16. The method of claim 9, wherein the vehicle motion information includes acceleration and angular velocity values, collision values, pressure values, roll values, pitch angle values, and yaw angle values of the vehicle.
17. An apparatus for controlling an airbag of a vehicle, the apparatus comprising: A human body injury probability calculator configured to calculate a human body injury condition probability and a human body injury prediction probability, respectively, based on the vehicle motion information measured by the sensing device; A learner, the learner configured to Calculating a human injury prior probability based on the human injury conditional probability and the human injury prediction probability, Calculating a human injury posterior probability by multiplying the human injury conditional probability by the human injury prior probability, and The human body injury prior probability is updated by feeding back the current human body injury posterior probability to the previous human body injury prior probability through probability-based real-time feedback machine learning, and An airbag deployment determiner configured to determine whether to deploy an airbag based on the human injury posterior probability.
18. A method of controlling an airbag of a vehicle, the method comprising: A human body injury probability calculator for calculating a human body injury condition probability and a human body injury prediction probability based on the vehicle motion information measured by the sensing device; Calculating, by a learner, a human injury prior probability based on the human injury conditional probability and the human injury prediction probability; Calculating, by the learner, a human injury posterior probability by multiplying the human injury conditional probability by the human injury prior probability; the learner updates the human injury prior probability by feeding back the current human injury posterior probability to the previous human injury prior probability through probability-based real-time feedback machine learning, and Whether to deploy the airbag is determined by an airbag deployment determiner based on the human injury posterior probability.

Description

Apparatus and method for controlling an airbag of a vehicle Technical Field The present disclosure relates to a vehicle, and more particularly, to airbag deployment control in the vehicle. Background Airbags are devices for protecting passengers from impact at the time of vehicle collision, and are typical passenger protection devices of vehicles along with seat belts. When a vehicle collision is detected by the sensor, the working gas device is triggered, and the airbag is instantaneously inflated (deployed) by the explosive gas to protect the passenger. Therefore, the shorter the time between the vehicle collision and the airbag deployment time, the better. However, it is necessary to determine whether to deploy the airbag according to whether the impact caused by the vehicle collision is strong enough to deploy the airbag or whether the impact reaches a level where the airbag does not need to be deployed. The passenger cannot be protected when the airbag is not deployed in the case where the airbag must be deployed to protect the passenger. Conversely, if the impact is to the extent that the airbag is not required to be deployed, it is undesirable because the use of the airbag is unnecessary and there is a cost of rearranging the airbag (e.g., replacement, repair). In other words, it is necessary to accurately determine whether to deploy an airbag according to the degree of injury to a passenger in a vehicle collision situation. Disclosure of Invention An aspect of the present disclosure is to provide an airbag control apparatus and method capable of determining whether an airbag is to be deployed based on a human injury posterior probability calculated through a human injury probability model and bayesian network learning (feedback learning) to ensure robustness of airbag deployment logic and more effectively protect a passenger. Additional aspects of the disclosure will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the disclosure. According to one aspect of the present disclosure, an apparatus for controlling an airbag of a vehicle is provided. The apparatus includes a human body injury probability calculator configured to calculate a human body injury conditional probability and a human body injury prediction probability based on vehicle motion information measured by the sensing apparatus, a learner configured to calculate a human body injury posterior probability by performing probability-based real-time feedback machine learning based on the human body injury conditional probability and the human body injury prediction probability, and an airbag deployment determiner configured to determine whether to deploy an airbag based on the human body injury posterior probability. The learner may be configured to calculate a human injury prior probability based on the human injury conditional probability and the human injury prediction probability, and calculate a human injury posterior probability by multiplying the human injury conditional probability by the human injury prior probability. The probability-based real-time feedback machine learning may be configured to update the human injury prior probability by feeding back the current human injury posterior probability to the previous human injury prior probability. The human injury conditional probability may be configured to be calculated by the following equation 1. [ Formula 1] P(xt|ut,xt-1),P(zt|xt) The expression P (x t|ut,xt-1) represents the human injury prediction probability at the current time point (t) according to the measured values of the collision sensors 102 and 106 and the human injury probability at the previous time (t-1). The expression P (z t|xt) represents a human injury prediction probability predicted from the passenger injury measured through simulation. The term x t represents the actual human injury probability for each of the six regions of the head, neck and chest at the current point in time (t). The term u t denotes the measurement of the sensing device 250 at the current point in time (t). The term x t-1 denotes the actual human injury probability at the previous time point (t-1). The term z t represents the human injury prediction probability for each of the six regions of the head, neck and chest at the current point in time (t). The human injury prior probability may be configured to be calculated by the following equation 2. [ Formula 2] The expression P (x t|ut,xt-1) represents the human injury prediction probability at the current time point (t) and the measured values of the collision sensors 102 and 106 according to the previous (t-1) human injury probability. The expression bel (x t-1) represents the posterior probability of human injury from the previous time (t-1). The human injury posterior probability may be configured to be calculated by the following equation 3. [ Formula 3] The term η denotes a normalization factor.