A method for establishing a driving risk field model

Abstract

The present invention discloses a method for establishing a driving risk field model, comprising: (1) establishing an x-axis along the road line direction, establishing a y-axis perpendicular to the road line direction, constructing a plane environment field and establishing a z-axis representing the field strength perpendicular to the plane environment field to construct a three-dimensional environment field; (2) for a moving object i, if a driving style coefficient is not required, calculating the field strength of the moving object i in the environment field; (3) for a moving object i, if the moving object i is a vehicle and the driving style coefficient needs to be considered, calculating the field strength of the moving object i in the environment field; (4) calculating the field strength of a stationary object i in the environment field; step (5), combining steps (2)-(4) to construct a unified model of driving risk field. The method for establishing a driving risk field model of the present invention increases the influence of road boundaries on driving, as well as the consideration of the influence of object speed and spacing, providing more effective support for driving warning and path planning.

Description

Technical Field

[0001] The invention relates to the field of driving risk fields, and in particular to a method for establishing a macroscopic driving risk field model. Background Art

[0002] Driving risk field refers to the use of a mathematical model analogous to the "physical field" to quantitatively describe the driving risks faced by a vehicle based on the acquisition of dynamic traffic information around the vehicle. The earliest such research was Khatib's proposal of a mobile robot trajectory planning method based on the concept of artificial potential field in the field of robot trajectory planning, but this method has the problem of local minimum. Later, Wolf et al. proposed an artificial potential field, which is a superposition of different functions of autonomous vehicles such as lane keeping, road stay, speed preference, and avoidance and passing objects. A set of potential energy functions composed of lanes, roads, cars and speed potential can effectively simulate the road environment around autonomous vehicles. Wang Jianqiang et al. proposed a new concept, namely driving risk field, which uses field theory to represent traffic risk factors related to drivers, vehicles and road environments. However, it does not consider the impact of road boundaries on driving, as well as the impact of object speed and spacing. The risk estimation has a certain deviation, which affects the accuracy. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method for establishing a driving risk field model in response to the above-mentioned deficiencies in the prior art. The method for establishing a driving risk field model adds consideration of the impact of road boundaries on driving, as well as the impact of object speed and spacing, to provide more effective support for driving warning and path planning.

[0004] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0005] A method for establishing a driving risk field model, comprising:

[0006] Step (1), establishing an x-axis along the direction of the road line, establishing a y-axis perpendicular to the direction of the road line, constructing a plane environment field and establishing a z-axis representing the field strength perpendicular to the plane environment field, so as to construct a three-dimensional environment field;

[0007] Step (2): For the moving object i, let its center of mass coordinates be (x i ,y i ), if the driving style coefficient is not required, the field strength of the moving object i in the environmental field is:

[0008]

[0009] Among them, G, p are both unknown constants greater than 0; R i is (x i ,y i) Road conditions at M affect the model; i is the equivalent mass of object i; TI i is the influence model of the speed and distance of object i; θ i is the deflection angle of object i and the x-axis, d ij is the distance between object i and other objects j, σ i Represents the convergence factor that determines the area affected by the obstacle; r ij Represents the distance vector between object i and other objects j;

[0010] Step (3), for the moving object i, let its center of mass coordinates be (x i ,y i ), if the moving object i is a vehicle, and the driving style coefficient corresponding to the vehicle needs to be considered, then the field strength of the moving object i in the environmental field is:

[0011]

[0012] Among them, Style i represents the driving style coefficient corresponding to object i as a vehicle;

[0013] Step (4): For a stationary object i, let its centroid coordinates be (x i ,y i ), then the field strength of the stationary object i in the environmental field is:

[0014]

[0015] Step (5), the unified model of driving risk field is:

[0016]

[0017] Among them, n is the number of field strengths of objects of the type in step (2) in the environment field, l is the number of field strengths of objects of the type in step (4) in the environment field, and m is the number of field strengths of objects of the type in step (3) in the environment field.

[0018] As a further improved technical solution of the present invention, (x i ,y i ) Road conditions affect model R i Specifically:

[0019]

[0020] Where: i is the road visibility, δ * is the standard road visibility; μ i is the road adhesion coefficient, μ * is the standard road adhesion coefficient; ρi is the road curvature, ρ * is the standard road curvature; τ i is the road slope, τ * is the standard road slope; γ1, γ2, γ3, γ4, γ5 are all unknown constants; Line represents the impact model of the road dividing line on driving.

[0021] As a further improved technical solution of the present invention, the equivalent mass M of the object i is i for:

[0022]

[0023] Among them, m i is the mass of object i; T i is the type of object i; α k , β k are all unknown constants; k is the number of terms in the velocity polynomial.

[0024] As a further improved technical solution of the present invention, the influence model TI of the speed and distance of object i i for:

[0025]

[0026] Among them, q1 and q2 are unknown constants, v i is the real-time velocity of object i, model TI i The above formula is applicable to objects i with a speed greater than or equal to 80 km / h; model TI i The following formula is applicable to objects i with a speed less than 80 km / h;

[0027] Among them, d ij for:

[0028]

[0029] Among them, δ1 is the length correlation coefficient of object i, δ2 is the width correlation coefficient of object i, α is the speed correlation coefficient of object i, lenth is the length of object i, wid c is the width of object i, (x i ,y i ) is the coordinate of the center of mass of object i, (x j ,y j ) is the center of mass coordinate of other object j.

[0030] The beneficial effects of the present invention are:

[0031] Based on the driving risk field model, the present invention adds the influence of road boundaries on driving, adds the consideration of the influence of object speed and distance, that is, adds the consideration of vehicle headway, and uses the improved vehicle headway d ij The formula is more adaptive to different directional spacing. At the same time, according to the Gaussian formula (such as exp(-d ij 2 ) / 2σ i 2 ) to obtain a more accurate risk probability distribution. By further improving the driving risk field strength formula, more effective support is provided for driving warning and path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the kinetic energy field strength of a single object from a bird's-eye view.

[0033] Figure 2 This is a field strength distribution map from an overhead view of the driving risk field. DETAILED DESCRIPTION

[0034] The specific embodiments of the present invention are further described below according to the accompanying drawings:

[0035] In road traffic, whether it is a moving object such as a vehicle and a pedestrian, or a stationary object such as a median and a roadblock, there is a possibility of colliding with a moving vehicle and causing safety risks. For a moving object, the severity of the collision depends on factors such as its equivalent mass, driving direction and distance. The possibility of a vehicle colliding with a moving object increases as the distance decreases, and this increase presents a power function relationship rather than a linear relationship. In addition, the degree of danger of a collision also depends on the direction in which the vehicle approaches and the speed of the moving object; the degree of danger is greatest when the vehicle approaches from the front, and on the contrary, it is the least when it approaches from the rear. For a stationary object, the severity of the collision is also affected by the equivalent mass. In addition, the danger of a vehicle approaching a stationary object increases exponentially as the distance decreases. Unlike a moving object, when a vehicle approaches a stationary object from any direction, the law of change in the degree of danger is the same and nonlinear. In general, whether it is a moving or stationary obstacle, when they approach a vehicle, they will increase the degree of danger of driving, and this increase presents a nonlinear characteristic.

[0036] Based on the above analysis, this embodiment provides a method for establishing a driving risk field model, including:

[0037] Step (1), establishing an x-axis along the direction of the road line, establishing a y-axis perpendicular to the direction of the road line, constructing a plane environment field and establishing a z-axis representing the field strength perpendicular to the plane environment field, so as to construct a three-dimensional environment field;

[0038] Step (2): For the moving object i, let its center of mass coordinates be (x i,y i ), if the driver's driving style coefficient does not need to be considered, the field strength of the moving object i in the environmental field is:

[0039]

[0040] Among them, G, p are both unknown constants greater than 0; R i is (x i ,y i ) Road conditions at M affect the model; i is the equivalent mass of object i; TI i is the influence model of the speed and distance of object i; θ i is the deflection angle of object i and the x-axis, d ij is the distance between object i and other objects j, σ i Represents the convergence factor that determines the area affected by the obstacle, which can be obtained by performing Gaussian fitting on the result calculated by Line; r ij represents the distance vector between the center of mass of object i and the center of mass of other objects j;

[0041] Step (3), for the moving object i, let its center of mass coordinates be (x i ,y i ), if the moving object i is a vehicle, and the driving style coefficient of the driver corresponding to the vehicle needs to be considered, then the field strength of the moving object i in the environmental field is:

[0042]

[0043] Among them, Style i represents the driving style coefficient of the driver corresponding to the object i as a vehicle;

[0044] Step (4): For a stationary object i, let its centroid coordinates be (x i ,y i ), then the field strength of the stationary object i in the environmental field is:

[0045]

[0046] Step (5), the unified model of driving risk field is:

[0047]

[0048] Among them, n is the number of field strengths of objects of the type in step (2) in the environment field, l is the number of field strengths of objects of the type in step (4) in the environment field, and m is the number of field strengths of objects of the type in step (3) in the environment field.

[0049] Furthermore, in the driving environment, the road has the most direct impact on driving. The road is determined by its transverse and longitudinal section design, roadbed and pavement engineering, maintenance and many other aspects. In this design, the basic properties of the road environment mainly focus on factors such as road adhesion coefficient, slope, visibility, etc. in road conditions, and add road boundary factors that can reflect the range of the driving environment. Based on the above factors, they are integrated into a comprehensive road impact model, namely (x i ,y i ) Road conditions affect model R i :

[0050]

[0051] Where: i is the road visibility, δ * is the standard road visibility; μ i is the road adhesion coefficient, μ * is the standard road adhesion coefficient; ρ i is the road curvature, ρ * is the standard road curvature; τ i is the road slope, τ * is the standard road slope; γ1, γ2, γ3, γ4, γ5 are all unknown constants, which can be calibrated by analytic hierarchy process or principal component analysis; Line represents the impact model of road dividing line on driving.

[0052] Line represents the impact model of the road dividing line on driving. The potential energy generated by the lane boundary has different effects on vehicle handling, and has potential energy obstacles to the vehicle's driving position and lane change intention. For highways, the impact of the road boundary line and the lane dividing line is mainly considered. The potential energy expression of the lane boundary is:

[0053]

[0054] Among them, A boundary Represents the potential coefficient of the lane boundary line, y boundary,c Represents the c th Lane boundary line; A divide Represents the potential energy coefficient of the lane dividing line, y divide,s Represents the s th Lane dividing line, Represents the speed of potential energy change, which is proportional to the lane width. For two-lane road sections, when exploring the driving risk characteristics on the road, for the lower risk state of the vehicle near the center of the lane, a trigonometric function with a small amplitude and relatively stable change can be used as the modeling basis. When the vehicle gradually approaches the road boundary, an exponential function can be used for modeling. The "repulsive force" effect generated by the road boundary helps to "guide" the vehicle from the edge of the road back to the center of the lane, thereby ensuring driving safety. The specific example formula is as follows:

[0055]

[0056] In formula (7): y i is the y coordinate of a certain point on the road, that is, the y coordinate of object i, y i =0 is the straight line where the outer boundary of the lane is located; wid r is the total width of the road; y l They respectively represent the lateral position of the center line of the left lane. This expression is applicable to dual lanes.

[0057] For moving objects on the road, their mass, type, state, etc. will cause certain risks to the vehicle. Wang Jianqiang et al. proposed the "equivalent mass" expression to describe the safety impact caused by the properties of objects in the environment. The expression is as follows:

[0058]

[0059] In the formula, m i is the mass of object i; T i is the type of object i; α k , β k are all unknown constants; k is the number of terms in the speed polynomial. In 2004, the World Bank and the World Health Organization pointed out in the report "Road Traffic Safety Countermeasures in Developing Countries" that the number of traffic accidents, the number of injured people and the number of deaths are respectively related to the square, the cubic and the quartic of the average road speed. Therefore, the speed polynomial is used To represent the effect of speed on equivalent mass.

[0060] In order to better reflect the safety of vehicle driving, this paper adds the consideration of headway distance. The consideration of headway distance also reflects the impact of vehicle speed and vehicle distance on vehicle driving safety. According to the different driving speeds of vehicles, the following influence model can be obtained, that is, the influence model TI of object i speed and distance i for:

[0061]

[0062] Among them, q1 and q2 are unknown constants, v iis the real-time velocity of object i, model TI i The above formula is applicable to high-speed object i, that is, object i with a speed greater than or equal to 80 km / h; model TI i The following formula in is applicable to objects i traveling at medium and low speeds, i.e., objects i traveling at a speed less than 80 km / h. In order to more adaptably express the distance between the vehicle and the target in different directions, the basic distance formula is improved to obtain the following d ij expression:

[0063]

[0064] In formula (10), δ1 is the length correlation coefficient of object i, δ2 is the width correlation coefficient of object i, α is the speed correlation coefficient of object i, lenth is the length of object i, wid c is the width of object i (such as a vehicle), (x i ,y i ) is the coordinate of the center of mass of object i, and (xj, yj) is the coordinate of the center of mass of other object j. δ1, δ2 and α can be calculated using the existing formula.

[0065] The movement and position change of an object at different angles can be expressed by the function exp(kcos(θ i )). When the object's motion direction is consistent with the reference direction, the function value is maximum, otherwise it is minimum. Added Gaussian form: exp(-d ij 2 ) / 2σ i 2 , used to determine the probability distribution.

[0066] The environmental field formula of formula (1) in this paper is applicable to moving objects. When used for moving objects, it can be directly used for other moving vehicles in the environment; if the self-vehicle is considered, the applicable formula can be obtained by combining the driver's style coefficient:

[0067] Among them, Style i represents the driving style coefficient corresponding to object i as a vehicle; it can be obtained by comprehensively considering the driver's psychology, cognitive response, technology, error rate, etc.; when used for a stationary object i, its corresponding v i =0, the corresponding field strength formula can be simplified to:

[0068] Based on the above, formula (4) can be obtained.

[0069] All the unknown constants in this article can be set to a constant value through experience.

[0070] Experimental verification:

[0071] The schematic diagram of the kinetic energy field generated by a moving object is shown in the figure below. The center of the kinetic energy field is the location of the moving object. The field strength at the center of the kinetic energy field is infinite, which means that if other objects coincide with this point, a traffic accident will inevitably occur.

[0072] Depend on Figure 1 It can be seen that the kinetic energy field is more densely distributed in the direction of the object's movement, that is, at the same distance, the closer to the object's movement direction, the greater the kinetic energy field strength.

[0073] For the establishment of the overall environmental field, this paper simulates a two-lane driving traffic scene, in which the driving risk field is composed of the common field strength and of the moving vehicles, stationary obstacles, roads and environmental facilities in the scene. In this setting, the calibration of each parameter is as follows: p1 = 1, p2 = 0.05, M1 = M2 = 5500kg, Style1 = 0.4, Style2 = 0.6, R1 = R2 = 1. In the above symbols, subscript 1 is the parameter corresponding to moving vehicle 1 (referred to as vehicle 1), and subscript 2 is the parameter corresponding to moving vehicle 2 (referred to as vehicle 2); thus, the driving risk map corresponding to the traffic scene can be obtained, as shown in Figure 2 , Figure 2 The car 1 indicated in the figure represents the field strength distribution near car 1, and the car 2 indicates the field strength distribution near car 2.

[0074] Figure 1 and Figure 2 In the field strength distribution diagram, different colors are used to represent different field strengths, |Ev| represents the field strength, the redder the color, the stronger the field strength and the greater the degree of danger.

[0075] The intensity distribution map of the driving risk field can be used to intuitively determine the degree of driving danger at each location, providing support for vehicle driving warning and path planning.

[0076] The protection scope of the present invention includes but is not limited to the above embodiments. The protection scope of the present invention shall be based on the claims. Any replacement, deformation, and improvement of the technology that can be easily thought of by technicians in this field shall fall within the protection scope of the present invention.

Claims

1. A method for establishing a driving risk field model, characterized in that: include: Step (1), establishing an x-axis along the direction of the road line, establishing a y-axis perpendicular to the direction of the road line, constructing a plane environment field and establishing a z-axis representing the field strength perpendicular to the plane environment field, so as to construct a three-dimensional environment field; Step (2): For the moving object i, let its center of mass coordinates be (x i ,y i ), if the driving style coefficient is not required, the field strength of the moving object i in the environmental field is: Among them, G, p are both unknown constants greater than 0; R i is (x i ,y i ) Road conditions at M affect the model; i is the equivalent mass of object i; TI i is the influence model of the speed and distance of object i; θ i is the deflection angle of object i and the x-axis, d ij is the distance between object i and other objects j, σ i Represents the convergence factor that determines the area affected by the obstacle; r ij Represents the distance vector between object i and other objects j; Step (3), for the moving object i, let its center of mass coordinates be (x i ,y i ), if the moving object i is a vehicle, and the driving style coefficient corresponding to the vehicle needs to be considered, then the field strength of the moving object i in the environmental field is: Among them, Style i represents the driving style coefficient corresponding to object i as a vehicle; Step (4): For a stationary object i, let its centroid coordinates be (x i ,y i ), then the field strength of the stationary object i in the environmental field is: Step (5), the unified model of driving risk field is: Wherein, n is the number of field strengths of objects of the type in step (2) in the environment field, l is the number of field strengths of objects of the type in step (4) in the environment field, and m is the number of field strengths of objects of the type in step (3) in the environment field; (x i ,y i ) Road conditions affect model R i Specifically: Where: i is the road visibility, δ * is the standard road visibility; μ i is the road adhesion coefficient, μ * is the standard road adhesion coefficient; ρ i is the road curvature, ρ * is the standard road curvature; τ i is the road slope, τ * is the standard road slope; γ1, γ2, γ3, γ4, γ5 are all unknown constants; Line represents the impact model of the road dividing line on driving.

2. The method for establishing a driving risk field model according to claim 1, characterized in that: The equivalent mass M of the object i is i for: Among them, m i is the mass of object i; T i is the type of object i; α k , β k are all unknown constants; k is the number of terms in the velocity polynomial.

3. The method for establishing a driving risk field model according to claim 2, characterized in that: The influence model TI of the speed and distance of object i i for: Among them, q1 and q2 are unknown constants, v i is the real-time velocity of object i, model TI i The above formula is applicable to objects i with a speed greater than or equal to 80 km / h; model TI i The following formula is applicable to objects i with a speed less than 80 km / h; Among them, d ij for: Among them, δ1 is the length correlation coefficient of object i, δ2 is the width correlation coefficient of object i, α is the speed correlation coefficient of object i, lenth is the length of object i, wid c is the width of object i, (x i ,y i ) is the coordinate of the center of mass of object i, (x j ,y j ) is the center of mass coordinate of other object j.

Images