CN-122009224-A - Dynamic risk visualization method and system based on Bayesian network

CN122009224ACN 122009224 ACN122009224 ACN 122009224ACN-122009224-A

Abstract

The invention discloses a dynamic risk visualization method and a system based on a Bayesian network, and relates to the technical field of risk visualization; inputting the multi-element driving monitoring data into a Bayesian network model to obtain a current human factor risk level, visually displaying the current human factor risk level, activating a human-computer intelligent decision-making allocation mechanism, and executing human-computer collaborative management according to the human-computer intelligent decision-making allocation mechanism based on the current human factor risk level. The invention solves the technical problems of insufficient driving safety and man-machine suitability caused by difficult real-time perception, non-visual presentation and incapability of dynamically adjusting man-machine interaction according to risks in the driving process in the prior art, achieves the technical effects of realizing real-time perception, visual presentation and dynamic collaborative management and control of the man-machine risk in the driving process, and improves the driving safety and man-machine interaction suitability.

Inventors

Hu Lunhu
HAI YUN
LIU PENGFEI
LIU ZHIHUI
BAOYINHEXI

Assignees

内蒙古工业大学

Dates

Publication Date: 20260512
Application Date: 20251203

Claims (8)

1. A bayesian network-based dynamic risk visualization method, the method comprising: Real-time monitoring is carried out on a driver according to the multi-mode monitoring component, and multi-element driving monitoring data are obtained; inputting the multivariate driving monitoring data into a Bayesian network model to obtain a current human factor risk level; Visually displaying the current human factor risk level and activating a human-computer intelligent decision-making allocation mechanism; And executing man-machine collaborative management according to the man-machine intelligent decision-making allocation mechanism based on the current human factor risk level.
2. A bayesian network based dynamic risk visualization method according to claim 1, wherein the man-machine intelligent decision apportionment mechanism comprises: When the current personal factor risk level is a low risk level, acquiring a driver leading instruction; When the current human factor risk level is a middle risk level, acquiring a human-computer collaborative leading instruction; And when the current personal factor risk level is a high risk level, acquiring a system leading instruction.
3. A bayesian network based dynamic risk visualization method according to claim 2, wherein the man-machine co-dominant instruction comprises highlighting recommended decision options; If the driver selects the recommendation decision option, directly executing the recommendation decision option; and triggering a secondary confirmation mechanism if the driver does not select the recommended decision option.
4. A bayesian network based dynamic risk visualization method according to claim 2 wherein the system lead instructions include automatically executing optimal decision options and informing the driver that the driver retains overrules on automatically executed results.
5. The bayesian network-based dynamic risk visualization method of claim 1, wherein performing human-machine collaborative management according to the human-machine intelligent decision-making apportionment mechanism based on the current human-machine risk level comprises: recording decision feedback and results of the driver under different risk levels; and updating the prior probability parameters of the Bayesian network model by utilizing the decision feedback and the result.
6. The bayesian network-based dynamic risk visualization method of claim 1, wherein visually exposing the current personal risk level comprises: activating a risk display instrument panel according to the current personal factor risk level; and carrying out self-adaptive pointer driving management on the risk display instrument panel according to the current personal factor risk level.
7. The bayesian network-based dynamic risk visualization method of claim 1, wherein the multi-modal monitoring component comprises a steering wheel torque sensor, a driver status monitoring camera, a seat physiological sensor, a vehicle dynamics sensor, a high-precision map, a GPS, an on-board computer, a digital dashboard, and a speech synthesizer.
8. A bayesian network based dynamic risk visualization system for implementing the bayesian network based dynamic risk visualization method according to any of claims 1-7, the system comprising: the driving monitoring data acquisition module is used for monitoring a driver in real time according to the multi-mode monitoring component to acquire multi-element driving monitoring data; The human factor risk level acquisition module is used for inputting the multi-element driving monitoring data into a Bayesian network model to acquire the current human factor risk level; The visual display module is used for visually displaying the current human factor risk level and activating a human-computer intelligent decision-making allocation mechanism; And the man-machine collaborative management module is used for executing man-machine collaborative management according to the man-machine intelligent decision-making allocation mechanism based on the current human factor risk level.

Description

Dynamic risk visualization method and system based on Bayesian network Technical Field The invention relates to the technical field of risk visualization, in particular to a dynamic risk visualization method and a dynamic risk visualization system based on a Bayesian network. Background In a driving scenario with safety requirements, the cognitive state of the driver and the human risk caused by the change of the driving environment are core factors influencing the driving safety. However, the existing driving assistance technology has obvious limitations that on one hand, the monitoring of human factor risks is dependent on single dimension data, multiple information such as physiological states, operation behaviors, vehicle operation parameters and environmental road conditions of a driver are difficult to integrate, so that risk perception is delayed and potential safety hazards cannot be captured in real time on the other hand, the risk results are presented in a simple alarm mode, visual hierarchical display is lacking, the driver is difficult to judge the risk degree quickly and accurately, meanwhile, a human-computer interaction mode is fixed, a system only passively provides data or basic alarms, decision authority allocation cannot be dynamically adjusted according to the risks, and finally driving safety management and control is lack of pertinence and synergism, so that accidents caused by cognitive errors or decision delays are difficult to be effectively avoided. In the prior art, the risk of a driver is difficult to sense in real time, presentation is not visual, and man-machine interaction cannot be dynamically adjusted according to the risk, so that the technical problems of insufficient driving safety and man-machine suitability are caused. Disclosure of Invention The application provides a dynamic risk visualization method and a dynamic risk visualization system based on a Bayesian network, which are used for solving the technical problems that in the prior art, the driving safety and man-machine suitability are insufficient because the risk is difficult to sense in real time, the presentation is not intuitive, and the man-machine interaction cannot be dynamically adjusted according to the risk. In view of the above, the present application provides a bayesian network-based dynamic risk visualization method and system. In a first aspect of the present application, there is provided a bayesian network based dynamic risk visualization method, the method comprising: The method comprises the steps of carrying out real-time monitoring on a driver according to a multi-mode monitoring component to obtain multi-element driving monitoring data, inputting the multi-element driving monitoring data into a Bayesian network model to obtain a current human factor risk level, carrying out visual display on the current human factor risk level and activating a human-computer intelligent decision-making and distributing mechanism, and carrying out human-computer collaborative management according to the human-computer intelligent decision-making and distributing mechanism based on the current human factor risk level. In a second aspect of the present application, there is provided a bayesian network based dynamic risk visualization system, the system comprising: The system comprises a driving monitoring data acquisition module, a human factor risk level acquisition module, a visual display module and a human-computer collaborative management module, wherein the driving monitoring data acquisition module is used for carrying out real-time monitoring on a driver according to a multi-mode monitoring component to obtain multi-element driving monitoring data, the human factor risk level acquisition module is used for inputting the multi-element driving monitoring data into a Bayesian network model to obtain a current human factor risk level, the visual display module is used for carrying out visual display on the current human factor risk level and activating a human-computer intelligent decision allocation mechanism, and the human-computer collaborative management module is used for executing human-computer collaborative management according to the human-computer intelligent decision allocation mechanism based on the current human factor risk level. One or more technical schemes provided by the application have at least the following technical effects or advantages: The method comprises the steps of carrying out real-time monitoring on a driver according to a multi-mode monitoring component to obtain multi-element driving monitoring data, inputting the multi-element driving monitoring data into a Bayesian network model to obtain a current human factor risk level, carrying out visual display on the current human factor risk level and activating a human-computer intelligent decision-making and distributing mechanism, and carrying out human-computer collaborative management according to the human-computer intellige