A method for extracting driver's real-time reaction time based on the trajectories of front and rear vehicles
The data is collected by vehicle sensors to generate a space-time trajectory diagram, and the reaction time is calculated using the Follow-up model, which solves the problem of low accuracy in fatigue driving detection in the prior art, and accurately judges the driver's fatigue or distraction without adding equipment state, improve driving safety.
Patent Information
- Application Number
- CN202111249371.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-26
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2041-10-26
AI Technical Summary
The existing fatigue driving detection methods are low in accuracy, making it difficult to effectively judge whether the driver is in a fatigue or distracted state without adding equipment, and are greatly affected by driving habits and external conditions.
The driving speed data of the bicycle and the vehicle in front are collected through on-board sensors, and the space-time trajectory diagram is generated. The driver's reaction time is calculated using the follow-up model and micro-traffic flow theory, and the driver's reaction time is determined based on the reaction time threshold.
It realizes accurate judgment of the driver's fatigue or distraction without adding equipment, reduces early warning costs, improves driving safety, and avoids the influence of driving habits and external conditions.
Smart Images

Figure CN114170760B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of active highway traffic safety and intelligent transportation, and in particular to a method for extracting a driver's real-time reaction time based on front and rear vehicle trajectories. Background Art
[0002] According to the traffic police department, about 40% of the current major traffic accidents are caused by fatigue driving, which seriously threatens driving safety. Drivers in a state of fatigue driving will have a series of adverse psychological and physiological reactions, such as slow reaction and decreased judgment ability. In severe cases, they will fall asleep and cause traffic accidents.
[0003] There are mainly the following methods for existing fatigue driving detection and warning. The first method relies on collecting the frequency of the driver's vehicle operations such as turning the steering wheel, stepping on the brakes and the accelerator to determine whether the driver is in a state of fatigue, but the accuracy of this method varies for drivers with different driving habits; the second method uses a wearable body data detector to collect body data to determine whether the driver is in a state of fatigue, but this method is more difficult to complete complex detection with limited on-board equipment; the third method is based on videos and pictures, deep learning of the driver's facial expressions, integrating eye, mouth, and head features, and using computers to detect and identify human actions such as closing eyes and yawning to determine whether the driver is in a state of fatigue, but this method will be affected by the quality of the pictures collected by the camera, and will also be affected by conditions such as lighting, which will interfere with the accuracy of the judgment.
[0004] Therefore, a method for extracting the driver's real-time reaction time based on the trajectories of the front and rear vehicles is designed. The sensor is used to measure the information data such as the speed of the vehicle and the speed of the front vehicle, and the vehicle trajectory data of the vehicle and the front vehicle are extracted to generate the spatiotemporal trajectory diagram of the vehicle and the front vehicle. Then, a new method for calculating the reaction time based on the trajectory data of the front vehicle and the vehicle is proposed to determine whether the driver is fatigued or distracted driving, which can effectively reduce the warning cost and improve driving safety. The difficulty and innovation of this patent lies in: without adding too much extra equipment, fully mining the effective information in the trajectory data of the front and rear vehicles, and combining the microscopic traffic flow theory, a new method for quickly extracting the driver's real-time reaction time is proposed through data expansion and statistical analysis. Summary of the invention
[0005] In order to identify the driving state of the driver and remind the driver who is slow to react, fatigued or distracted, the present invention analyzes the real-time reaction time of the driver. Considering that the rear vehicle will react to the change of the driving state of the front vehicle, a method for extracting the real-time reaction time of the driver based on the trajectories of the front and rear vehicles is invented. The spatiotemporal trajectory diagram of the vehicle and the front vehicle is established, and the key parameters are calculated and analyzed to obtain the optimized reaction time to judge whether the driver is fatigued or distracted.
[0006] A method for extracting a driver's real-time reaction time based on the trajectories of front and rear vehicles comprises the following steps:
[0007] Step 1: Collect the driving speed data information of the vehicle and the front vehicle through the on-board sensors, and output the time-space trajectory diagram of the vehicle and the front vehicle in real time.
[0008] Step 2: Based on the car-following model, analyze the impact of the vehicle in front on the ego vehicle.
[0009] Based on the Newell following model, the position of the vehicle is affected by the vehicle in front, that is,
[0010] x n (t) = x n-1 (t-τ n )+d n
[0011] Among them, the position of the vehicle n at time t is the same as that of the preceding vehicle n-1 at time t-τ n The position difference at time d n , where τ n It can be understood as the reaction time of vehicle n, d n is the minimum interval between vehicle n and the preceding vehicle.
[0012] When the driver receives a signal from the vehicle ahead or a change in road conditions, he does not take action immediately. Instead, he realizes the need to take action after τ1 seconds, and then starts to take action after τ2 seconds. This period of time τ=τ1+τ2 is called the driver's reaction time.
[0013] According to investigation and testing, except for special situations such as distracted driving and fatigue driving, the driver's reaction time is relatively stable over a period of time. Therefore, the change in reaction time can be used to determine whether the driver is in a state of fatigue or distracted driving. The reaction time threshold can be determined based on the average of the driver's input or recorded reaction time, and this patent is set to 1.5s.
[0014] According to the trajectory diagram output in step 1, the spatiotemporal trajectory diagrams of the preceding vehicle and the vehicle within 10 seconds are selected as the input for the subsequent steps.
[0015] Step 3: Extract the parameters of the leading vehicle speed, pre-reaction time, pre-minimum interval, wave speed and passing rate according to the spatiotemporal trajectory diagram of the leading vehicle and the self-vehicle;
[0016] According to the space-time trajectory diagram, for each trajectory point of the ego vehicle, the law of the vehicle following the movement of the front vehicle is used, combined with the front vehicle trajectory, to extract parameters such as the front vehicle speed, pre-reaction time, pre-minimum interval, wave speed and pass rate; and then obtain multiple sets of data sets containing the above parameters.
[0017] Step 4: Analyze the changes in pass rate;
[0018]
[0019]
[0020] In the formula, u represents the speed of the front vehicle, v represents the speed of the vehicle, w is the wave speed, k x is the maximum traffic density; I(v) represents the change in the passing rate r when the speed of the vehicle is v and the speed of the front vehicle is u1 and u2 respectively. From the above formula, it can be seen that the passing rate is related to the wave speed, the maximum traffic density, the front vehicle speed and the speed of the vehicle. When the speed of the vehicle v remains unchanged, the change in the passing rate is related to the speed difference of the front vehicle and the speed of the vehicle and the wave speed difference. When the wave speed w is the same as the speed of the vehicle v, no matter how the speed difference of the front vehicle changes, the change in the passing rate I(v) is always 0. Under the same vehicle speed, compare the changes in the passing rate r when u is different, and the vehicle speed v with the smallest change in the passing rate is the actual wave speed.
[0021] Step 5: Calculate the actual wave speed;
[0022] There are a series of discrete ego-vehicle trajectory points in the ego-vehicle space-time trajectory diagram. The ego-vehicle speed is discretized at intervals of 1 m / s to obtain the ego-vehicle speed set V. For different ego-vehicle speeds v i , select the wave speed w i ∈[v i -1,v i +1] corresponding to the passing rate r i , calculate these pass rates r i Repeat the above calculation to get the standard deviation of different vehicle speeds v i Corresponding pass rate r i The standard deviation of , and then select the v corresponding to the minimum standard deviation i As the actual wave speed w * .
[0023] Step 6: Calculate the optimized reaction time;
[0024] Each point in the trajectory of the vehicle has a set of parameters {u0,τ0,d0,w0,r0; ...; u m ,τ m ,d m ,w m ,r m}, select the corresponding wave speed w from each set of data in turn. * The smallest difference w i , select each w i The corresponding τ i As the reaction time of this point, the reaction time array {τ0; ...; τ m}; The mean of the array is the driver’s real-time reaction time within 10 seconds.
[0025] Step 7: Determine the driver's driving status.
[0026] The vehicle-mounted sensors collect information such as the speed of the vehicle and the vehicle in front, and output the driver's reaction time in real time. If the reaction time exceeds the reaction time threshold of 1.5s, the driver is considered to be in a state of fatigue or distracted driving and is slow to react, and a voice reminder should be given to the driver.
[0027] The beneficial effects of the present invention compared with the prior art are:
[0028] The present invention extracts the driver's real-time reaction time based on the front and rear vehicle trajectories, and uses the driver's reaction time to judge the driver's driving state, effectively avoiding the uncertainty caused by different driving habits of drivers in traditional methods, effectively avoiding the complex vehicle-mounted body detection system, and effectively avoiding interference caused by external conditions such as light. It can output the real-time reaction time of different drivers, and judge whether the driver is in a state of fatigue or distraction based on the driver's reaction time. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 It is a framework for judging the driving state of the driver in the method for extracting the real-time reaction time of the driver based on the trajectories of the front and rear vehicles of the present invention;
[0030] Figure 2 It is a data detection processing diagram in the method for extracting the driver's real-time reaction time based on the front and rear vehicle trajectories of the present invention;
[0031] Figure 3 This is an example diagram of calculating the reaction time based on the vehicle trajectory in the method for extracting the real-time reaction time of the driver based on the trajectories of the front and rear vehicles of the present invention;
[0032] Figure 4 It is a flowchart of obtaining the actual wave speed in the method for extracting the real-time reaction time of the driver based on the trajectories of the front and rear vehicles of the present invention;
[0033] Figure 5 It is a flow chart for obtaining reaction time in the method for extracting driver's real-time reaction time based on the trajectories of front and rear vehicles described in the present invention. DETAILED DESCRIPTION
[0034] The present invention is described in detail below in conjunction with the accompanying drawings:
[0035] A method for extracting the driver's real-time reaction time based on the trajectories of the front and rear vehicles is shown in the attached figure. Figure 1 , including the following steps:
[0036] Step 1: Collect information data such as the driving speed of the vehicle and the front vehicle through the on-board sensors, and output the time-space trajectory diagram of the vehicle and the front vehicle in real time.
[0037] See attached Figure 3 , the time-space trajectory diagram of the vehicle and the front vehicle is output in real time, with the horizontal axis being time and the vertical axis being the position of the vehicle on the road.
[0038] Step 2: Based on the car-following model, analyze the impact of the vehicle in front on the ego vehicle.
[0039] Based on the Newell following model, the position of the vehicle is affected by the vehicle in front, that is,
[0040] x n (t) = x n-1 (t-τ n )+d n
[0041] Among them, the position of the vehicle n at time t is the same as that of the preceding vehicle n-1 at time t-τ n The position difference at time d n , where τ n It can be understood as the reaction time of vehicle n, d n is the minimum interval between vehicle n and the preceding vehicle.
[0042] According to the investigation and test, except for special situations such as distraction and fatigue driving, the driver's reaction time is relatively stable over a period of time, and the change in reaction time can be used to determine whether the driver is in a state of fatigue or distraction. The reaction time threshold is determined based on the average of the driver's input or recorded reaction time. According to the trajectory diagram output in step 1, the 10s space-time trajectory diagram of the preceding vehicle and the vehicle is selected as the input for the subsequent steps.
[0043] Step 3: Extract the parameters of the preceding vehicle speed, pre-reaction time, pre-minimum interval, wave speed and passing rate according to the time-space trajectory diagram;
[0044] See attached Figure 2 , the speed detection device can be used to detect the speed of the vehicle and the front vehicle, and the time-space trajectory diagram of the front vehicle and the vehicle can be obtained. The calculation method of the passing rate is:
[0045] r=Δn / Δt
[0046] The passing rate r represents the number of vehicles that pass through the traffic wave per unit time when the traffic wave is used to cut the space-time trajectory diagram, that is, Δn / Δt. Considering only the front and rear vehicles, the wave speed w = d / τ, and cutting the trajectory diagram with a traffic wave with a wave speed of w, Δn = 1, Δt = τ, r = 1 / τ is the passing rate corresponding to the current speed of the vehicle and the speed of the front vehicle.
[0047] Reference Figure 3, the reaction time τ and the minimum interval d cannot be less than 0, so the area where the vehicle may be affected by the front vehicle at a certain moment is determined by the intersection of the horizontal and vertical lines passing through this point and the trajectory line of the front vehicle. Since the vehicle trajectory is discrete, the data is also stored in a discrete state, and there is a certain time interval between the data. This patent requires a time interval of 0.1s to find all the front vehicle positions that may affect the vehicle, that is, A0 and A in the front vehicle trajectory. n All points between. If B i A i The driver's pre-reaction time of vehicle B is τ i , the minimum interval is d i , wave speed w i =d i / τ i , pass rate r i =1 / τ i The calculation process of this method is simple, and the speed of the front vehicle is the real value. For each point B on the trajectory of the vehicle i We can get a set of parameters {u0,τ0,d0,w0,r0;…;u m ,τ m ,d m ,w m ,r m}, which respectively represent the speed of the preceding vehicle, the pre-reaction time, the pre-minimum interval, the wave speed, and the passing rate.
[0048] Step 4: Analyze the changes in pass rate;
[0049]
[0050]
[0051] In the formula, u represents the speed of the front vehicle, v represents the speed of the vehicle, w is the wave speed, k x is the maximum traffic density; I(v) represents the change in the passing rate r when the vehicle speed is v and the front vehicle speed is u1 and u2 respectively. From the above formula, it can be seen that the passing rate is related to the wave speed, the maximum traffic density, the front vehicle speed and the vehicle speed. When the vehicle speed v is unchanged, the change in the passing rate is related to the front vehicle speed difference, the vehicle speed and the wave speed difference; when the wave speed w is the same as the vehicle v, no matter how the front vehicle speed difference changes, the change in the passing rate I(v) is always 0. Under the same vehicle speed, compare the changes in the passing rate r when u is different, and the vehicle speed v with the smallest change in the passing rate is the actual wave speed.
[0052] Step 5: Calculate the actual wave speed;
[0053] Reference Figure 4 With attached Figure 3There are a series of discrete ego-vehicle trajectory points in the ego-vehicle space-time trajectory diagram, with a speed interval of 1m / s, and the ego-vehicle speed set V is obtained. For different ego-vehicle speeds v i , select the wave speed w i ∈[v i -1,v i +1] corresponding to the passing rate r i , calculate these pass rates r i Repeat the above calculation to get the standard deviation of different vehicle speeds v i Corresponding pass rate r i The standard deviation of , and then select the v corresponding to the minimum standard deviation i As the actual wave speed w is selected * .
[0054] Step 6: Calculate the optimized reaction time;
[0055] Reference Figure 5 With attached Figure 3 , each point B on the trajectory of the vehicle i We can get a set of parameters {u0,τ0,d0,w0,r0;…;u m ,τ m ,d m ,w m ,r m}, which represent the front vehicle speed, pre-reaction time, pre-minimum interval, wave speed, and passing rate respectively. Select the corresponding wave speed w from each set of data in turn. * The smallest difference w i , select each w i The corresponding τ i As the reaction time of this point, the reaction time array {τ0; ...; τ m}; The mean of the array is the driver’s real-time reaction time within 10 seconds.
[0056] Step 7: Determine the driver's driving status.
[0057] The real-time reaction time of the driver within 10 seconds is output in real time. If the real-time reaction time is greater than the reaction time threshold of 1.5 seconds, the driver is considered to be in a state of fatigue or distracted driving and has a slow reaction, and a voice reminder is given to the driver, otherwise no reminder is given.
Claims
1. A method for extracting driver's real-time reaction time based on the trajectories of the front and rear vehicles. It is characterized in that The actual reaction time of the driver is extracted according to the trajectories of the vehicle and the preceding vehicle, and compared with the reaction time threshold to judge the driving state, including the following steps: Step 1: Collect the driving speed data of the vehicle and the preceding vehicle through the vehicle-mounted sensors, and output the spatiotemporal trajectory diagram of the vehicle and the preceding vehicle in real time; Step 2: Based on the car-following model, analyze the impact of the vehicle in front on the ego vehicle. When the driver receives a signal from the vehicle ahead or a change in road conditions, he or she does not take action immediately, but takes a period of time before taking action, which is called reaction time. According to the following model, the ego vehicle will be affected by the vehicle ahead, and the reaction time threshold is determined based on the driver input or the average of the recorded reaction time, which is set to a fixed value of 1.5s. According to the trajectory diagram output in step 1, the spatiotemporal trajectory diagram of the 10s preceding vehicle and the ego vehicle is selected as the input for the subsequent steps. Step 3: Extract the parameters of the leading vehicle speed, pre-reaction time, pre-minimum interval, wave speed and passing rate according to the spatiotemporal trajectory diagram of the leading vehicle and the self-vehicle; According to the time-space trajectory diagram, for each trajectory point of the ego vehicle, the law of the vehicle following the movement of the front vehicle is used, combined with the front vehicle trajectory, to extract the parameters of the front vehicle speed, pre-reaction time, pre-minimum interval, wave speed and passing rate; and then multiple sets of data sets containing the above parameters are obtained; Step 4: Analyze the changes in pass rate; The passing rate is related to the speed of the preceding vehicle, the speed of the vehicle itself, the wave speed and the maximum traffic density. When the speed of the vehicle itself is fixed, if the speed of the vehicle itself is equal to the wave speed, the change in the passing rate is independent of the change in the speed of the preceding vehicle, and the actual wave speed is calculated accordingly. Step 5: Calculate the actual wave speed; Discretize the vehicle speed in the space-time trajectory diagram, calculate the standard deviation of the passing rate change under different vehicle speeds, and select the vehicle speed with the smallest standard deviation as the actual wave speed; Step 6: Calculate reaction time; Among the corresponding wave speeds at any point of the vehicle trajectory, the wave speed that is closest to the actual wave speed and within the threshold range is selected, and the pre-reaction time corresponding to the wave speed is used as the reaction time of the point; the average reaction time of all points is taken as the actual reaction time of the driver during the period; Step 7: Determine the driving status of the driver; Output the driver's real-time reaction time within 10 seconds in real time. If the driver's actual reaction time is greater than the reaction time threshold of 1.5 seconds, the driver is considered to be in a state of fatigue or distracted driving and a voice reminder is given, otherwise no reminder is given; Step 3: Extract the parameters of the preceding vehicle speed, pre-reaction time, pre-minimum interval, wave speed and pass rate according to the time-space trajectory diagram; wherein the pass rate is calculated as follows: r = △n / △t The passing rate r represents the number of vehicles that pass through the traffic wave per unit time when the traffic wave is used to cut the space-time trajectory diagram, that is, △n / △t; considering only the front and rear vehicles, the wave speed w=d / τ, and the trajectory diagram is cut with the traffic wave with a wave speed of w to obtain △n=1,△t=τ; r=1 / τ is the passing rate corresponding to the current vehicle speed and the front vehicle speed; The values of reaction time τ and minimum interval d cannot be less than 0. Therefore, the area where the ego vehicle may be affected by the front vehicle at a certain moment is determined by the intersection of the horizontal and vertical lines passing through this point and the trajectory line of the front vehicle. Since the vehicle trajectory is discrete, the data is also stored in a discrete state. There is a certain time interval between the data. The time interval is required to be 0.1s. Find all the positions of the front vehicles that may affect the ego vehicle, that is, A in the front vehicle trajectory 0 and A n All points between i A i The driver's pre-reaction time of vehicle B is τ i , the minimum interval is d i , wave speed w i =d i / τ i , pass rate r i =1 / τ i ; The calculation process of this method is simple, and the speed of the front vehicle is the real value; for each point B on the trajectory of the vehicle i We get a set of parameters {u 0 ,τ 0 ,d 0 ,w 0 ,r 0 ;…;u m ,τ m ,d m ,w m ,r m }, which represent the front vehicle speed, pre-reaction time, pre-minimum interval, wave speed, and passing rate respectively; In step 4, the change of the pass rate is analyzed; the calculation method of the pass rate and the change of the pass rate is: In the formula, u represents the speed of the front vehicle, v represents the speed of the vehicle, w is the wave speed, k x is the maximum traffic density; I(v) means when the vehicle speed is v, the speeds of the preceding vehicles are u 1 、u 2 When the speed of the vehicle is the same as that of the vehicle, the change of the passing rate r is always 0. ... The change of the passing rate r is always 0. When the speed of the vehicle is the same as that of the vehicle, the change of the passing rate r is 0. The change of the passing rate r is 0. The change of the passing rate r is 0. The change of the passing rate r is 0. The change of the passing rate r is In step 5, the actual wave speed is calculated. There are a series of discrete self-vehicle trajectory points in the self-vehicle space-time trajectory diagram. The self-vehicle speed is discretized at an interval of 1 m / s to obtain the self-vehicle speed set V. For different self-vehicle speeds v i , select the wave speed w i ∈[v i -1,v i +1] corresponding to the passing rate r i , calculate these pass rates r i The standard deviation of the vehicle speed v is obtained by repeating the calculation. i Corresponding pass rate r i The standard deviation of , and then select the v corresponding to the minimum standard deviation i As the actual wave speed w is selected * ; In step 6, the optimal reaction time is calculated; each point on the ego-vehicle space-time trajectory graph has a set of parameters {u 0 ,τ 0 ,d 0 ,w 0 ,r 0 ;…;u m ,τ m ,d m ,w m ,r m }; Select the w with the smallest difference from the actual wave speed w* from each set of data in turn i , select each w i The corresponding τ i As the reaction time of this point, the reaction time array {τ 0 ;…;τ m }; Calculate the mean of the array as the driver’s real-time reaction time within 10 seconds.
