| Fatigue driving is the main cause of traffic accidents,especially serious traffic accidents,which will threaten people's lives and cause great losses to social property.Fatigue will reduce the driver's perception of the surrounding environment and the ability to control the vehicle,which has a great potential safety hazard.Fatigue driving is widespread,and many drivers have continued to drive under fatigue.Especially on the highway,the monotonous driving environment is more likely to cause driver fatigue,and the condition of parking and rest can not be obtained anytime and anywhere on the highway.Therefore,it is of great significance for driving safety to explore the generation mechanism of driving fatigue,study the identification method of driving fatigue and design the mitigation means of driving fatigue.At present,there is no unified conclusion on the mechanism of driving fatigue.Based on the characteristics of highway driving environment,this paper establishes a driving fatigue mechanism model based on visual and auditory stimulation.Near infrared spectroscopy(NIRS)was used to collect the data of oxygen-containing hemoglobin concentration(?HBO)in frontal lobe,temporal lobe and occipital lobe of the driver's brain,and then the coupling strength of each brain region before and after fatigue was analyzed based on Dynamic Bayesian inference.The results showed that the coupling strength of ?HBO in temporal and frontal lobes was decreased by continuous auditory stimulation,while the coupling strength of occipital and frontal lobes was not affected by visual stimulation before and after fatigue.In addition,after fatigue,the coupling strength of ?HBO signals in the left and right prefrontal lobes of drivers decreased significantly in the ? and ? frequency bands.According to the research results of driving fatigue mechanism,the ?HBO signal of left temporal lobe and left and right prefrontal lobe is used as the data of driving fatigue identification.Combined with the time-series characteristics of driving fatigue,a double-layer parallel gating circulating unit(GRU)algorithm based on deep learning network was established.The algorithm can make full use of the time-series characteristics of ?HBO signal,and has strong feature extraction ability.It can realize the identification of three types of driving fatigue(awake state,mild fatigue and severe fatigue),and the average accuracy of the algorithm reaches 85%.Then,this paper explores the brain function mechanism of olfactory stimulation on alleviating driving fatigue.The ?HBO signals of the driver's left and right frontal lobes were collected by NIRS technique.The coupling intensity of the left and right frontal lobes before and after fatigue and before and after olfactory stimulation was analyzed by dynamic Bayesian inference.The results show that olfactory stimulation can significantly enhance the coupling strength of the ? and ? frequencies in the left and right frontal lobes of the brain.The increase of coupling intensity indicates that the symmetry of contraction and expansion of the smooth muscle and the coordination of nerve activity are enhanced,which can improve the driver's ability of response,direction judgment and decision-making.The potential danger avoidance experiment further confirmed the effect of olfactory stimulation on the improvement of drivers' obstacle avoidance ability.The results show that the driver who inhales the smell has faster response,more moderate braking and steering and less lane deviation in avoiding potential danger ahead,and the overall obstacle avoidance operation is more reasonable and safe.Finally,the effects of different olfactory stimulation parameters(odor concentration,release time)on driving fatigue were analyzed.Three odor concentrations(5%,10% and 15%)and 2 release time(5s,10s)were set.The results showed that the best fatigue relieving effect was obtained under the condition of 10%concentration and 5 s release time.At the same time,there were certain individual differences between olfactory stimulation parameters and fatigue relief effect,and not all participants obtained the best fatigue relief effect under the condition of 10%concentration and 5s release time.Therefore,based on the objective phenomenon of individual differences,an individual-based olfactory stimulus parameter optimization algorithm was established in this paper.The algorithm can optimize the parameters for individual drivers to achieve the best fatigue relief effect.The experimental results show that the parameter optimization algorithm can improve the fatigue relief effect of olfactory stimulation,and the effect will become more and more obvious in the process of repeated optimization: the average duration of 10 times of fatigue relief is 5.19 minutes higher than that of the control group(fixed 10% concentration and 5S release time),and the average duration of the last five tests is 7.42 minutes higher than that of the control group. |
PDF Full Text Request