A driver steering mode modeling method under icy and snowy road surface

Abstract

The application provides a driver manipulation mode modeling method under ice and snow road surface. In the case of comprehensively considering the road surface condition under extreme ice and snow weather and the operation characteristics of the driver, a driver model considering the mental stress factor of the driver under ice and snow environment is established, so that the driver model can fully reflect the different behavior abilities of the driver under ice and snow road surface, and aims to solve the problem that the existing driver model under ice and snow road surface is insufficient to depict different behavior abilities, so as to obtain a better control effect for subsequent man-machine sharing direct control. The application comprises the following steps: regarding the driver model to be researched as a whole, utilizing the visual perception and motion perception of the driver, establishing a complete driver manipulation model through a two-point visual anticipation module and a compensation module, and selecting and adjusting each module and specific parameters in the model according to the different behavior abilities and manipulation characteristics of different drivers.

Description

Technical Field

[0001] This invention belongs to the field of mechanical engineering technology. Specifically, it designs a pre-aiming-control driver control mode modeling method that can clearly and accurately characterize different driver behaviors on icy and snowy roads for researchers to use in the future. Background Technology

[0002] Research on driver control mode modeling has a history of at least fifty years. To realistically simulate driver behavior, many scholars both domestically and internationally have conducted related research, resulting in numerous driver models, based on different modeling mechanisms, methods, and application scenarios. The establishment of driver models plays a crucial role in the development of vehicle active safety control systems, vehicle closed-loop system simulation, and the evaluation of vehicle handling stability systems. Since live-action experiments are not feasible to better ensure driver safety, simulation experiments using relevant software are particularly important.

[0003] Driving in snowy weather significantly increases the challenge to drivers' abilities. When a car is driving on icy or snowy roads, it may be subject to external interference, causing changes in the vehicle's lateral stability and leading to skidding. Drivers will instinctively take measures to correct the car's direction. However, if a driver lacks composure or experience, they may become nervous under these adverse driving conditions, potentially leading to errors in vehicle operation, which is extremely dangerous.

[0004] Current research on vehicle operation on icy and snowy roads primarily focuses on models of specific driver control modes, but it falls short in characterizing the diverse behavioral abilities of different drivers. Therefore, this invention aims to accurately characterize different driver behaviors on icy and snowy roads within a human-vehicle-road closed-loop system by establishing a driver preview-control model under extreme conditions on icy and snowy roads.

[0005] Based on the above considerations, this invention proposes a driver control mode modeling method. Driver behavior on icy and snowy roads is the result of the interaction between vision and motion; generally, driver behavior is characterized by a specific driver control mode, but this mode has limited diversity and is unsuitable for researchers to develop and use. This invention proposes a driver control mode modeling method for icy and snowy roads. On the one hand, it integrates driver behavior into the control design process to verify driver behavior on icy and snowy roads. On the other hand, it considers that a pre-aiming-control driver model can accurately describe the characteristics of drivers with different behavioral abilities to meet research needs. Summary of the Invention

[0006] The technical problem solved by this invention is to address the shortcomings of existing research by proposing a driver control mode modeling method suitable for icy and snowy roads. This method integrates driver behavior into the control design process to verify driver behavior on icy and snowy roads. Furthermore, by selecting various modules and varying specific parameters in the driver model, it characterizes differences in driver style, thus providing a more suitable human-machine shared control model for subsequent control research under icy and snowy road conditions.

[0007] The driver control mode modeling method for icy and snowy roads provided in this invention is achieved through the following technical solution:

[0008] Step 1: Treat the driver model to be studied as a whole, without using single-point pre-aiming information, without using pre-aiming information from the left, right and rear sides of the vehicle, without considering the transformation of the obtained pre-aiming information into the desired steering wheel angle through linear or nonlinear control theory, and at the same time, assume that the road friction coefficient of the icy and snowy road surface is uniform and constant.

[0009] Step Two: Based on the system obtained in Step One, utilize the driver's vision and motion perception to integrate driver behavior into the control design process. Establish a complete driver dynamics model through a two-point visual expectation module and a compensation module. Further refine the driver model for icy and snowy roads based on the different behavioral abilities and operational characteristics of different drivers. This will better reflect the different behavioral abilities of different drivers under extreme icy and snowy road conditions. Figure 1 As shown;

[0010] Step 3: Based on the dynamic expression obtained in Step 2, and combined with the mathematical relationship between the driver and the road under icy and snowy conditions, an explicit expression for the anticipation information under the current road conditions is given, resulting in a more complete driver model for icy and snowy roads. The established driver model's input signals include longitudinal velocity, heading angle, lateral deviation, and yaw rate, and its output is the driver's torque, which can fully meet the needs of researchers for subsequent work and research.

[0011] Real-world drivers are complex and exhibit numerous uncertainties and diversity. This invention treats the driver model as a whole, integrating driver behavior into the control design process. It avoids using single-point anticipation information, anticipation information from the left, right, and rear of the vehicle, and the conversion of obtained anticipation information into the desired steering wheel angle using linear or nonlinear control theory. Furthermore, it assumes the road friction coefficient on icy or snowy surfaces is uniform and constant. Utilizing driver vision and motion perception, it integrates driver behavior into the control design process, establishing a complete driver dynamics model through a two-point visual anticipation module and a compensation module.

[0012] The driver model block diagram considering two points of visual expectation and compensation is as follows: Figure 2 As shown, it utilizes both visual and motion perception, integrating driver behavior into the control design process. The model is based on the assumption that the driver uses a distant visual cue to predict road curvature and proximity visual information to compensate for lateral positioning errors. The outputs of two modules are added to form the reference signal for the neuromuscular system. The neuromuscular system module represents the muscles being activated by both α and γ motor neuron signals, calculating an appropriate steering torque and applying it to the steering wheel.

[0013] The anticipatory control in the predictive control section is a simple proportional controller with a gain K. a From θ far supply.

[0014] For compensatory control, the driver utilizes their vision and kinesthetic sense to compensate for instantaneous changes in the trajectory. The driver uses information from the near zone (near point of view) to maintain center position and correct the vehicle's current position within the lane. The associated driver actions are determined by… Figure 2 The gain parameter K is represented by the transfer function in the compensation control. c This represents the driver's proportional action relative to the near-viewpoint error. The driver adjusts these three parameters to adapt to the system's dynamics. Among them, the visual compensation control K... c This indicates the driver's caution regarding getting too close to lane signs. As speed increases, K... c This will reduce reliance on near-vision information, meaning the dependence on near-vision information is lower at this time. Among them, T... l and T i These are the lead and lag time constants, respectively. And T i It is also a useful indicator for measuring driver fatigue, representing the impact of the driver's psychological state. The arm's inertia, passive damping, and passive stiffness are represented by the neuromuscular dynamics system module, where T... n This is the neuromuscular time lag constant. The output of the neuromuscular system represents the torque T provided by the driver. d .

[0015] The system dynamics corresponding to the above two pre-aiming-operator models are shown below:

[0016] x d =A d x d +B d u d y d =C d x d (1)

[0017] The state vector is defined as x d =[xd1 x d2 ] T and u d =[θ near θ far ] T State x d and Figure 2 This relates to the compensation control block within. d1 This is a state related to the compensation module in the diagram, which can be interpreted as considering θ near The driver's perception of the impending steering wheel correction under the change. Second state x d2 The driver's torque, i.e., x d2 =T d The output is also y. d =T d The above system matrix is ​​defined as follows:

[0018]

[0019] Among them, T l T i They are respectively Figure 2 The lead time constant and lag time constant of the compensation control block. The driver's neuromuscular lag time constant is represented by T. n It means that the obtained K a and K c These represent visual expectation control and visual compensation control, respectively.

[0020] Step Two involves selecting and adjusting various modules and specific parameters in the driver control mode model based on the different behavioral abilities and operational characteristics of different drivers on icy and snowy roads. The specific method is as follows:

[0021] Because this invention can accurately and clearly represent the different behavioral abilities of different drivers, it uses the "Prediction Error Method" (PEM). This method determines appropriate driver parameters by measuring or simulating response data to characterize the behavior of a typical driver in a specific vehicle. Missing parameters in the driver model are obtained through gray-box identification of the input and output data. Based on this, driver parameters for three different driving strategies are obtained. It is important to note that T... n This represents a rough estimate of neuromuscular dynamics and is not different across different objects; that is, the parameters of the neuromuscular system can be considered constant. However, icy and snowy roads can affect the driver's psychological state, which is reflected in the lag time constant T in the driver's operating mode. i The influence factor under icy and snowy road conditions is set as σ, while the influence factor σ = 1 under normal road conditions when the driver is driving normally.

[0022] The first type of driver, a recent licensed driver experiencing a curve for the first time, attempts to stay in the center of the lane, but this results in the largest distance from the curve tangent, potentially leading to an ineffective turn. These drivers lack experience and operate with a novice's mindset, therefore their expectations for anticipated control actions and approach angles are not high. Consequently, their K-axis performance is low. a and K c The parameters are all close to the middle level, which means that the driving style is unfamiliar and inexperienced. However, driving in icy and snowy conditions will make the driver feel nervous, and for novice drivers, the influencing factor σ > 1.

[0023] The above two points of pre-aiming and control driver model diagram are shown below. Figure 3 As shown, the corresponding system dynamics are as follows:

[0024]

[0025] Where u d1 =[θ near θ far ] T x d1 =[x d11 x d12 ] T Second state x d12 The driver's torque, i.e., x d12 =T d1 The output is also y. d1 =T d1 The above system matrix is ​​defined as follows:

[0026]

[0027] The second type of driver attempts to stay in the center with minimal lateral deviation. This can be considered a very cautious cornering approach, representing an experienced but conservative driver. The second type of driver tries to travel along a line that minimizes lateral deviation. That is, the driver model stays near the centerline, therefore the driver must compensate through steering wheel movements. As can be seen from the parameters, K... c Very high, which means the driver pays great attention to the approach angle θ. near .

[0028] The third type of driver attempts to follow the path closest to the tangent when turning. This behavior of minimizing lateral acceleration represents an experienced yet aggressive and impatient driver. The third type of driver cuts off the curve, taking a straighter path to reduce lateral acceleration. This behavior is achieved by minimizing the field of vision at the curve's tangent point. More experienced drivers have higher expectations for their anticipated control actions, which is K... a The larger the angle θ, the lower the expected control actions required by the other two drivers. This can be seen from the driver parameters, and it also means that the driver must pay close attention to the far angle θ. far .

[0029] While the driving behaviors of the second and third types of drivers differ significantly under icy and snowy conditions, both are experienced by the inexperienced first type of driver. Therefore, in terms of module representation, they can be considered as one type—experienced drivers—and have similar expressions in the dynamic equations. Regarding the driver's anxiety in icy and snowy environments, considering that they are all experienced drivers, extreme conditions may actually stimulate their more focused driving skills than usual. Therefore, the influence factor σ of the icy and snowy environment on the operating modes of these two types of drivers in the model is <1.

[0030] The block diagrams of the second and third types of driver models are as follows: Figure 4 As shown. The system dynamics corresponding to its pre-aiming-operator model are as follows:

[0031]

[0032] The state vector is defined as x d2 =[x d21 x d22 ] T and u d2 =[θ near θ far ] T State x d2 and Figure 4 This relates to the compensation control block within. d21 This is a state related to the compensation module in the diagram, which can be interpreted as considering θ near The driver's perception of the impending steering wheel correction under the change. Second state x d22 The driver's torque, i.e., x d22 =T d2 The output is also y. d2 =T d2 The above system matrix is ​​defined as follows:

[0033]

[0034] As described in step four, since the driver model, vehicle model, and road need to be jointly simulated in the subsequent control study, a visual near angle θ is required. near and visual distance θ far The explicit expression defines the origin of the vehicle-road coordinate system at the center of the road, with the vehicle's forward direction as the positive X-axis and the Y-axis perpendicular to the X-axis and parallel to the ground, pointing to the left of the vehicle. For example... Figure 5 As shown, the near angle θ near The heading angle ψ can be used L and lateral deviation y L The angle of near vision can be calculated using a function.

[0035]

[0036] Among them, l d T is the distance from the foreground point to the near point. p For the driver's preview time, ψ L It's the navigational point. V x V is the longitudinal velocity of the vehicle's center of mass. y Let l be the lateral velocity of the vehicle's center of mass. p It is the vertical distance from the forward-looking point to the lateral deviation. Therefore, in l d Lateral deviation error y at the location d for:

[0037] y d =y L +(l d -l p )ψ L (8)

[0038] From the above two equations, we can obtain...

[0039]

[0040] The second input θ far This represents the angle between the vehicle's direction of travel and the point of tangency. When the driver is at a distance L... f When tracking a distant point, the driver predicts the road curvature as follows:

[0041]

[0042] Where, ω r It is the yaw rate of the vehicle at the point of tangency. R tp The radius of curvature of the road can be approximated as... This gives us

[0043]

[0044] The advantages of this invention compared to existing technologies are as follows: This invention provides a driver modeling method for icy and snowy roads, integrating the driver's visual and kinematic perception control design under extreme icy and snowy weather conditions. Furthermore, the adjustment of modules and selection of parameters fully reflect the different behavioral capabilities of drivers on icy and snowy roads, further demonstrating the diversity of driver behavior and the accuracy of the model. This invention can be effectively integrated with CarSim's built-in vehicle model, and can quickly and accurately calculate the driver's output and apply it to the vehicle under icy and snowy double lane change and icy and snowy serpentine driving conditions. It has strong portability and versatility, and can clearly provide accurate references for related research tasks conducted by researchers. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the driver model structure provided by the present invention.

[0046] Figure 2 This is a schematic diagram of a general driver model provided by the present invention;

[0047] Figure 3 This is a schematic diagram of a novice driver model provided by the present invention;

[0048] Figure 4 A schematic diagram of the driver experience model provided by the present invention;

[0049] Figure 5 This is a schematic diagram of the model of the present invention from a certain perspective;

[0050] Figure 6 This is a comparison diagram between the model of this invention and the model included with CarSim;

[0051] Figure 7 Schematic diagram of double-track shifting operation in ice and snow conditions

[0052] Figure 8 Output torque diagram for a Class 1 driver model under dual lane change conditions in ice and snow;

[0053] Figure 9 Output torque diagrams for two types of driver models under the dual lane change conditions in ice and snow;

[0054] Figure 10 Output torque diagrams for three types of driver models under the double lane change condition in ice and snow;

[0055] Figure 11 This is a schematic diagram of a serpentine working condition in ice and snow.

[0056] Figure 12 Output torque diagram for driver model 1 under icy and snowy serpentine conditions;

[0057] Figure 13 Output torque diagrams for two types of driver models under icy and snowy serpentine conditions;

[0058] Figure 14 Output torque diagrams for three types of driver models under icy and snowy serpentine conditions. Detailed Implementation

[0059] The proposed modeling method for driver operation modes on icy and snowy roads will be further elaborated and explained below with reference to the accompanying drawings.

[0060] This invention provides a method for modeling driver operation patterns applicable to icy and snowy roads, implemented according to the following steps:

[0061] 1. The driver model to be studied is treated as a whole, and the driver's behavior is integrated into the control design process. Single-point anticipation information is not used, nor is anticipation information from the left, right, and rear sides of the vehicle. The conversion of the obtained anticipation information into the desired steering wheel angle through linear or nonlinear control theory is not considered. Simultaneously, the road friction coefficient of icy and snowy surfaces is assumed to be uniform and constant. Figure 1 As shown;

[0062] 2. The driver model block diagram considering two-point visual expectations and compensation in icy and snowy environments is as follows: Figure 2 As shown, it utilizes both visual and motion perception, integrating driver behavior into the control design process. The model is based on the assumption that the driver uses a distant visual cue to predict road curvature and proximity visual information to compensate for lateral positioning errors. The outputs of two modules are added to form the reference signal for the neuromuscular system. The neuromuscular system module represents the muscles being activated by both α and γ motor neuron signals, calculating an appropriate steering torque and applying it to the steering wheel.

[0063] The anticipatory control in the predictive control section is a simple proportional controller with a gain K. a From θ far supply.

[0064] For compensatory control, the driver utilizes their vision and kinesthetic sense to compensate for instantaneous changes in the trajectory. The driver uses information from the near-field point to maintain center position and correct the vehicle's current position within the lane. The associated driver actions are... Figure 2 The transfer function in the compensation control is used to represent this.

[0065] Gain parameter K c This represents the driver's proportional action relative to the near-viewpoint error. The driver adjusts these three parameters to adapt to the system's dynamics. Among them, the visual compensation control K... c This indicates the driver's caution regarding getting too close to lane signs. As speed increases, K... cThis will reduce reliance on near-vision information, meaning the dependence on near-vision information is lower at this time. l and T i These are the lead and lag time constants, respectively. And T i This could be a useful indicator for measuring driver fatigue. The arm's inertia, passive damping, and passive stiffness are represented by a neuromuscular dynamics system module, where T... n This is the neuromuscular time lag constant. The output of the neuromuscular system represents the torque T provided by the driver. d .

[0066] The system dynamics corresponding to the above two pre-aiming-operator models are shown below:

[0067]

[0068] The state vector is defined as x d =[x d1 x d2 ] T and u d =[θ near θ far ] T State x d and Figure 2 This relates to the compensation control block within. d1 This is a state related to the compensation module in the diagram, which can be interpreted as considering θ near The driver's perception of the impending steering wheel correction under the change. Second state x d2 The driver's torque, i.e., x d2 =T d The output is also y. d =T d The above system matrix is ​​defined as follows:

[0069]

[0070] Among them, T l T i They are respectively Figure 2 The lead time constant and lag time constant of the compensation control block. The driver's neuromuscular lag time constant is represented by T. n It means that the obtained K a and K c These represent visual expectation control and visual compensation control, respectively.

[0071] 3. Because this invention can accurately and clearly represent the different behavioral abilities of different drivers, it uses the "Prediction Error Method" (PEM) to determine appropriate driver parameters by measuring or simulating response data to characterize the behavior of a typical driver in a specific vehicle. Missing parameters in the driver model are obtained through gray-box identification of the input and output data. Based on the identification of driver parameters and simulation results, driver parameters for three different driving strategies were obtained, as shown in Table 1. It should be noted that T... n This represents a rough estimate of neuromuscular dynamics and is not different across different objects; that is, the parameters of the neuromuscular system can be considered constant. However, icy and snowy roads can affect the driver's psychological state, which is reflected in the lag time constant T in the driver's operating mode. i The influence factor under icy and snowy road conditions is set as σ, while the influence factor σ = 1 under normal road conditions when the driver is driving normally.

[0072] The first type of driver, a recent licensed driver experiencing a curve for the first time, attempts to stay in the center of the lane, but this results in the largest distance from the curve tangent, potentially leading to an ineffective turn. These drivers lack experience and operate with a novice's mindset, therefore their expectations for anticipated control actions and approach angles are not high. Consequently, their K-axis performance is low in these parameters. a and K c The parameters are all at an intermediate level, which indicates an unfamiliar and inexperienced driving style. Its model block diagram is as follows: Figure 3 As shown, driving in icy and snowy conditions can cause psychological tension for drivers. For novice drivers, the external environment has a greater physiological impact on their lag characteristics, so the influence factor σ > 1.

[0073] The system dynamics corresponding to the above two points of the pre-aiming-controlling driver model are as follows:

[0074]

[0075] Where u d1 =[θ near θ far ] T x d1 =[x d11 x d12 ] T Second state x d12 The driver's torque, i.e., x d12 =T d1 The output is also y. d1 =T d1 The above system matrix is ​​defined as follows:

[0076]

[0077] The second type of driver attempts to stay in the center with minimal lateral deviation. This can be considered a very cautious steering approach when cornering, representing an experienced but conservative driver. The second type of driver tries to travel along a line that minimizes lateral deviation. In other words, the driver model stays near the centerline, therefore the driver must compensate through steering wheel movements. As can be seen from the parameters, K... c Very high, which means the driver pays great attention to the approach angle θ. near .

[0078] The third type of driver attempts to follow the path closest to the tangent when turning. This behavior of minimizing lateral acceleration represents an experienced yet aggressive and impatient driver. The third type of driver cuts off the curve, taking a straighter path to reduce lateral acceleration. This behavior is achieved by minimizing the field of vision at the curve's tangent point. More experienced drivers have higher expectations for their anticipated control actions, which is K... a The larger the angle θ, the lower the expected control actions required by the other two drivers. This can be seen from the driver parameters, and it also means that the driver must pay close attention to the far angle θ. far .

[0079] While the driving behaviors of the second and third types of drivers differ significantly under icy and snowy conditions, both are experienced by the inexperienced first type of driver. Therefore, in terms of module representation, they can be considered as one type—experienced drivers—and have similar expressions in the dynamic equations. Regarding the driver's anxiety in icy and snowy environments, considering that they are all experienced drivers, extreme conditions may actually stimulate their more focused driving skills than usual. Therefore, the influence factor σ of the icy and snowy environment on the operating modes of these two types of drivers in the model is <1.

[0080] The block diagrams of the second and third types of driver models are as follows: Figure 4 As shown. The system dynamics corresponding to its two-point pre-aiming-operator model are as follows:

[0081]

[0082] The state vector is defined as x d2 =[x d21 x d22 ] T and u d2 =[θ near θ far ] T State x d2 and Figure 4 This relates to the compensation control block within. d21 This is a state related to the compensation module in the diagram, which can be interpreted as considering θ near The driver's perception of the impending steering wheel correction under the change. Second state x d22 The driver's torque, i.e., x d22 =T d2 The output is also y. d2 =T d2 The above system matrix is ​​defined as follows:

[0083]

[0084] 4. Because subsequent control studies require joint simulation of the driver model, vehicle model, and road, a visual near angle θ is needed. near and visual distance θ far The explicit expression defines the origin of the vehicle-road coordinate system at the center of the road, with the vehicle's forward direction as the positive X-axis and the Y-axis perpendicular to the X-axis and parallel to the ground, pointing to the left of the vehicle. For example... Figure 5 As shown, the near angle θ near The heading angle ψ can be used L and lateral deviation y L The function is used to calculate.

[0085] The angle of the near point can be expressed as:

[0086]

[0087] Among them, l d T is the distance from the foreground point to the near point. p For the driver's preview time, ψ L It's the navigational point. V x V is the longitudinal velocity of the vehicle's center of mass. y Let l be the lateral velocity of the vehicle's center of mass. d Lateral deviation error y at the location d for:

[0088] y d =y L +(l d -l p )ψ L (8)

[0089] From the above two equations, we can obtain...

[0090]

[0091] The second input is the far-angle θ. far This represents the angle between the vehicle's direction of travel and the point of tangency. When the driver is at a distance L...f When tracking a distant point, the driver predicts the road curvature as follows:

[0092]

[0093] At the tangent point At this point, it can be approximated as This gives us

[0094]

[0095] 5. The simulation model established based on the above steps needs to consider human-machine co-driving in subsequent research. Therefore, joint simulation of the two models needs to be performed in CarSim software. The vehicle model selected is a C-class sedan, a 2012 hatchback. The simulation vehicle parameters are shown in Table 2. The input of the vehicle model is the steering wheel torque. The input of the driver model is the vehicle's longitudinal velocity, heading angle, lateral deviation, and yaw rate, and the output is the driver torque.

[0096] This invention establishes a driver model for icy and snowy road conditions, using two lane-change and serpentine driving scenarios. The two-lane-change scenario involves a vehicle moving from one lane to another and back to its original lane; the serpentine scenario involves a vehicle navigating an "S"-shaped lane. After setting the scenarios, the road surface was also configured accordingly, using a straight road heading east. The road surface adhesion coefficient significantly affects tire lateral characteristics, so the impact of icy and snowy conditions on the road surface is primarily reflected in the road surface adhesion coefficient. Different road surface types have different adhesion coefficients, thus it is necessary to determine the adhesion coefficient corresponding to each road surface type. Different types of icy and snowy road surfaces have different adhesion coefficients, as shown in Table 3. Because the simulated environment involves icy and snowy conditions, the scenario is one where snow has fallen on the road surface, causing vehicle traffic and compaction, and the snowfall does not affect normal vehicle operation. This results in snow accumulation and lightly compacted snow on the road surface. A road surface adhesion coefficient of 0.35 was selected for simulation testing. The road surface was adjusted in CarSim, selecting a snowy road with a centerline.

[0097] The output of CarSim's built-in driver model and the designed driver control mode model at a speed of 54 km / h under double lane change conditions is as follows: Figure 6 Since CarSim's built-in driver model can only output steering angle, the output torque of the designed driver model was converted into steering angle through the steering system. It is evident that under icy and snowy conditions, compared to CarSim's built-in driver model, the steering angle produced by the driver model designed in this paper is more reasonable in both magnitude and clockwise / counterclockwise variation; therefore, choosing the entire driver model is correct.

[0098] Because the driver model used must reflect different driver behaviors, comparative experiments need to be conducted under different working conditions on icy and snowy roads. After testing the simulation model, it can be seen that the driver model can maintain the vehicle's forward trajectory on low-adhesion icy and snowy roads, reaching the set speed of 54 km / h, and the torque and steering angle generated under different working conditions are discussed.

[0099] Operating Condition 1: Double Line Transfer Condition in Ice and Snow. Figure 7 This is a schematic diagram of the double-track shifting operation. Figures 8-10 The output of the driver model is shown in the double lane change condition on the icy and snowy road surface. The vehicle model in the joint simulation is kept at a speed of 54 km / h.

[0100] Operating Condition 2: Snow and Ice Serpentine Operating Condition. Figure 11 This is a schematic diagram of a serpentine working condition. Figures 12-14 The output of the driver model under the serpentine condition on the icy and snowy road is shown. The vehicle model in the joint simulation maintains a speed of 54 km / h.

[0101] This invention comprehensively considers road conditions and driver behavior characteristics under extreme icy and snowy weather, establishing a pre-aiming-control driver model. This model includes proactive control, compensatory control, and a neuromuscular module, integrating driver behavior into the control design process to verify driver behavior on icy and snowy roads. Compared to existing driver models for icy and snowy roads, the driver model established in this invention is significantly superior to typical single-driver models. By selecting and varying specific parameters of each module within the driver model, it characterizes differences in driver styles, including inexperienced novices, experienced and cautious drivers, and experienced but bold and impatient drivers. This lays the foundation for achieving better control effects in subsequent auxiliary control system design and better reflects the actual operating characteristics of drivers on icy and snowy roads.

[0102] Table 1

[0103]

[0104] Table 2

[0105]

[0106]

[0107] Table 3

[0108] Road conditions Adhesion coefficient range Smooth ice film 0.05～0.15 Smooth compacted snow 0.10～0.20 Ice boards, ice boards under the snow 0.15～0.20 ice film 0.15～0.30 There are ice plates under the snow. 0.20～0.30 Snow accumulation, lightly compacted snow 0.25～0.35 Asphalt or concrete pavement (dry) 0.80～0.90

Claims

1. A method for modeling driver operation patterns on icy and snowy roads, comprising the following steps: Step 1: Treat the driver model to be studied as a whole, integrate the driver's behavior into the control design process, do not use single-point preview information, do not use preview information of the left, right and rear sides of the vehicle, do not consider converting the obtained preview information into the desired steering wheel angle through linear or nonlinear control theory, and at the same time consider the road friction coefficient of icy and snowy road surface to be uniform and constant. Step Two: Based on the system obtained in Step One, the driver's behavior is integrated into the control design process using the driver's vision and motion perception. The system dynamics corresponding to the two-point preview-control driver model are established through a two-point visual anticipation module and a compensation module, as shown below: (3) in , Second state The driver's torque, i.e. The output is also The above system matrix is ​​defined as follows: (4) The above-mentioned complete driver dynamics model is further refined based on the different behavioral abilities and operational characteristics of different drivers on icy and snowy roads. Step 3: Based on the dynamic expression obtained in Step 2, and combined with the mathematical relationship between the driver and the road when driving in the vehicle on icy and snowy roads, give an explicit expression for the pre-aiming information under the current road conditions and obtain the driver model on icy and snowy roads. The input signals of the established driver model include longitudinal velocity, heading angle, lateral deviation and yaw rate, and the output is driver torque; A driver model that reflects the different driving abilities of drivers is established on icy and snowy roads. The model is based on the assumption that drivers use visual information of a distant visual cue to predict road curvature in parallel to compensate for lateral positioning errors. The outputs of two modules are added to form a reference signal for the neuromuscular system; The neuromuscular system module represents how muscles are activated by signals from both α and γ motor neurons, which calculate a steering torque and apply it to the steering wheel. The anticipatory control section's lead control is a proportional controller with a gain of [missing information]. From a distance supply; For compensatory control, the driver utilizes their vision and kinesthetic sense to compensate for instantaneous changes in the trajectory; the driver uses information from the near-field point to maintain center position and correct the vehicle's current position within the lane; the associated driver actions are represented by the transfer function in the compensatory control, with gain parameters... This represents the driver's proportional action relative to the near-viewpoint error; the driver adjusts these three parameters to adapt to the system's dynamics; among them, visual compensation control... This indicates to the driver the need for caution regarding getting too close to lane signs; as speed increases, It will decrease, of which and These are the lead and lag time constants, respectively; and It is an indicator for measuring driver fatigue; icy and snowy roads can affect the driver's psychological state, which is related to the lag time constant in the driver's operating mode. The influence factor under icy and snowy road surfaces is set to Influencing factors when the driver is driving normally on a regular road surface The inertia, passive damping, and passive stiffness of the arm are represented by the neuromuscular dynamics system module, in which... This represents the neuromuscular time lag constant; the output of the neuromuscular system represents the torque provided by the driver. ; The driver model described in step three is a driver model that can reflect the different behavioral abilities of different drivers under extreme icy and snowy road conditions; The driver model is obtained by selecting and adjusting various modules and specific parameters; The Prediction Error Method (PEM) is used to determine appropriate driver parameters by measuring or simulating response data to characterize the behavior of a typical driver in a specific vehicle; missing parameters in the driver model are obtained by gray-box identification from the input and output data. Based on this, driver parameters for three different driving strategies were obtained; The This represents an estimate of neuromuscular dynamics that will not differ from different objects; that is, the parameters of the neuromuscular system are considered constant.

2. The method for modeling driver operation modes on icy and snowy roads according to claim 1, characterized in that, The first type of driver, after obtaining their license, attempts to stay in the middle of the lane when cornering for the first time, but this results in the largest distance from the corner tangent, leading to an ineffective turn. These drivers lack driving experience, and therefore, their performance in terms of parameters... and The parameters are all close to the middle level, which indicates an unfamiliar and inexperienced driving style; however, driving in icy and snowy conditions can cause psychological tension for the driver, which is an influencing factor for novice drivers. .

3. The method for modeling driver operation modes on icy and snowy roads according to claim 1, characterized in that, The second type of driver attempts to stay in the center with minimal lateral deviation, which is considered a very cautious way of navigating corners; that is, an experienced but conservative driver. This second type of driver tries to drive along a line that minimizes lateral deviation; that is, the driver model stays near the centerline, so the driver must compensate through steering wheel movements, as can be seen from the parameters. Very high, which means the driver pays great attention to the near angle. .

4. The method for modeling driver operation modes on icy and snowy roads according to claim 1, characterized in that, The third type of driver attempts to follow the path closest to the tangent when turning; this behavior of minimizing lateral acceleration represents an experienced yet aggressive and impatient driver; the third type of driver cuts corners, taking a straighter path to reduce lateral acceleration; this behavior is achieved by reducing the field of vision to the minimum of the curve's tangent point; that is... The larger the angle, the lower the requirements for the other two drivers' anticipated control actions; this also means that the driver must pay close attention to the long-range angle. .

5. The method for modeling driver operation modes on icy and snowy roads according to claim 3 or 4, characterized in that, While the driving behaviors of the second and third types of drivers differ significantly in icy and snowy conditions, both are considered experienced by the inexperienced first type of driver. Therefore, in the module representation, they are considered as one type—experienced drivers—and have similar expressions in the dynamic equations. Regarding the driver's anxiety in icy and snowy environments, considering that they are all experienced drivers, extreme conditions may actually stimulate their skills to be more focused than usual. Therefore, the influence factors of icy and snowy environments on the operating modes of these two types of drivers in the model are... ; The system dynamics corresponding to the second and third type of driver models are as follows: (5) The state vector is defined as follows: and ,state Related to the compensation control block. This is a state related to the compensation module in the diagram, and is interpreted as being under consideration. Under the change, the driver's perception of the upcoming steering wheel correction, the second state. The driver's torque, i.e. The output is also The above system matrix is ​​defined as follows: (6) 。 6. The method for modeling driver operation modes on icy and snowy roads according to claim 1, characterized in that, Solving the vehicle's positioning problem on the road involves setting the origin of the vehicle-road coordinate system at the center of the road, with the vehicle's direction of travel as... Positive axis direction Axis perpendicular to The axis is parallel to the ground and points to the left of the vehicle.

7. The method for modeling driver operation modes on icy and snowy roads according to claim 1, characterized in that, In step four, the driving model needs to obtain a display representation of the aiming information to refine the established driver model; near angle Use heading angle and lateral deviation The angle of near vision is calculated using a function, and then expressed as: （7） Among them, The distance from the forward-looking point to the near point. Preview time for drivers, It is the guiding angle. Let the longitudinal velocity be the vehicle's center of mass. Let be the lateral velocity of the vehicle's center of gravity. It is the vertical distance from the forward-looking point to the lateral deviation; therefore, in Lateral deviation error at the location for: (8) From the above two equations, we obtain... (9) Second input This represents the angle between the vehicle's direction of travel and the point of tangency. It is the yaw rate of the vehicle, when the driver is at a distance When tracking a distant point, the driver predicts the road curvature as follows: (10) Among them, at the tangent point place, The radius of curvature representing the road approaches 1. That's how (11)。

Images