Control method for deploying airbag

A control method for deploying airbags in a vehicle, the control method comprising:

• a model-securing step of acquiring learning data, via a controller, reflecting a state of an in-vehicle person and securing a deep learning model, via the controller, by learning a deep learning network based on the acquired learning data;
• an imaging step of imaging, via a detector, the in-vehicle person;
• a determining step of determining, via the controller, the state of the in-vehicle person by inputting image data of the imaged in-vehicle person into the deep learning network and by performing operations by means of the controller to determine the state of the in-vehicle person based on an output of the deep learning network which is based on the inputted image data;
• and
• a deployment control step of deploying an airbag in different ways, via the controller, depending on the state of the in-vehicle person when a collision accident occurs,
• further comprising:
• securing joint positions of the in-vehicle person by obtaining x-axial, y-axial, and z-axial coordinate values of joints of the in-vehicle person and determining a posture of the in-vehicle person by connecting the joint positions in the determining step;
• and
• deploying a front airbag in front of the in-vehicle person while adjusting (a) one or more of a point-in-time of deployment of the airbag, (b) whether to deploy the airbag, and (c) a deployment pressure of the airbag in correspondence to the determined posture of the in-vehicle person in the deployment control step.

A control method for deploying airbags in a vehicle, the control method comprising:

• a model-securing step of acquiring learning data, via a controller, reflecting a state of an in-vehicle person and securing a deep learning model, via the controller, by learning a deep learning network based on the acquired learning data;
• an imaging step of imaging, via a detector, the in-vehicle person;
• a determining step of determining, via the controller, the state of the in-vehicle person by inputting image data of the imaged in-vehicle person into the deep learning network and by performing operations by means of the controller to determine the state of the in-vehicle person based on an output of the deep learning network which is based on the inputted image data;
• and
• a deployment control step of deploying an airbag in different ways, via the controller, depending on the state of the in-vehicle person when a collision accident occurs,
• further comprising:
• determining whether a seatbelt in the vehicle has been fastened by detecting the seatbelt on the basis of image data input to the deep learning network in the determining step;
• and
• deploying a front airbag in front of the in-vehicle person earlier than a predetermined preset deployment time in the deployment control step when it is determined that the seatbelt has not been fastened.

A control method for deploying airbags in a vehicle, the control method comprising:

• a model-securing step of acquiring learning data, via a controller, reflecting a state of an in-vehicle person and securing a deep learning model, via the controller, by learning a deep learning network based on the acquired learning data;
• an imaging step of imaging, via a detector, the in-vehicle person;
• a determining step of determining, via the controller, the state of the in-vehicle person by inputting image data of the imaged in-vehicle person into the deep learning network and by performing operations by means of the controller to determine the state of the in-vehicle person based on an output of the deep learning network which is based on the inputted image data;
• and
• a deployment control step of deploying an airbag in different ways, via the controller, depending on the state of the in-vehicle person when a collision accident occurs,
• further comprising at least one of a following series of steps:
• (A) determining personal information of the in-vehicle person by recognizing a face of the in-vehicle person in the determining step;
• and
• deploying a front airbag in front of the in-vehicle person while adjusting a deployment pressure of the front airbag in correspondence to a body shape of the in-vehicle person based on the personal information in the deployment control step;
• or
• (B) determining personal information of the in-vehicle person by recognizing the face of the in-vehicle person in the determining step;
• and
• deploying a front airbag in front of the in-vehicle person while adjusting a deployment pressure of the front airbag in correspondence to medical information of the in-vehicle person based on the personal information in the deployment control step;
• or
• (C) determining whether a driver of the vehicle neglects looking toward a front area of the vehicle based on a gaze of the driver and states of pupils and eyelids of the driver in the determining step;
• and
• decreasing a collision determination-critical value for deploying an airbag in the deployment control step when it is determined that the driver is driving without looking toward the front area of the vehicle;
• or
• (D) determining whether a driver of the vehicle has dozed off at a steering wheel of the vehicle based on a gaze of the driver and states of pupils and eyelids of the driver in the determining step;
• and
• increasing a collision determination-critical value for deploying an airbag in the deployment control step when it is determined that the driver has dozed off at the steering wheel.