Automatic driving vehicle sharing remote driving control method, system, equipment and medium

Abstract

The invention discloses an automatic driving vehicle sharing remote driving control method, system and device and a medium, and relates to the technical field of automatic driving vehicle control. The method comprises the following steps: constructing a vehicle-road augmentation system model, constructing a two-point driver preview model by introducing a near preview point and a far preview point, and constructing a driver-in-the-loop human-vehicle-road system model based on the vehicle-road augmentation system model and the two-point driver preview model; on the basis of a human-vehicle-road system model, driver activeness is introduced to represent the driving intention and driving ability of a driver. According to the invention, a brand-new automatic driving vehicle remote sharing control architecture is constructed, the technical difficulties of large time delay of a wireless transmission network and reduced scene perception in remote driving are broken through, the remote driving performance is effectively improved, the development of a feasible region of the design of the automatic driving vehicle is realized, and the landing and popularization of an automatic driving technology are promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous vehicle control, and specifically to a method, system, device and medium for shared remote driving control of autonomous vehicles. Background Art

[0002] Considering the diverse problems in reality, further technological breakthroughs are still needed to achieve fully autonomous driving without human takeover. Therefore, in the context of the current development of the autonomous driving industry, vehicle remote driving technology, as a backup function for autonomous driving, has received increasing attention. Generally speaking, remote driving technology involving human participation can help autonomous vehicles pass through corner scenarios or extreme working conditions, as well as extreme environments beyond the Operational Design Domain (ODD). At the same time, when the system fails and causes functional degradation, humans can take over the vehicle through remote driving and bring it to a safe area. In addition, vehicle remote driving can also help improve the management of autonomous vehicle fleets. For example, in the RoboTaxi business. However, although remote driving is highly expected as an autonomous driving assistance technology, due to problems such as network latency and reduced scene perception, the current performance of vehicle remote driving is still far from expectations. For this reason, the present invention proposes a method, system, device and medium for shared remote driving control of autonomous vehicles. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, system, device and medium for shared remote driving control of autonomous vehicles, which can effectively handle problems such as network latency and reduced scene perception, improve lane keeping performance in a dynamic environment, and reduce the mental fatigue and driving load of remote drivers.

[0004] To achieve the above purpose, the present invention provides the following technical solutions: A method for shared remote driving control of autonomous vehicles, including the following steps:

[0005] Construct a vehicle-road augmented system model, and construct a two-point driver preview model by introducing a near preview point and a far preview point. Based on the vehicle-road augmented system model and the two-point driver preview model, construct a human-vehicle-road system model with the driver in the loop;

[0006] Based on the human-vehicle-road system model, introduce driver activity to characterize the driving intention and driving ability of the driver, and adaptively adjust the automated assistance torque according to the driver activity to obtain a human-vehicle-road control system adopting a tactile shared steering strategy;

[0007] According to the haptic shared steering strategy, the longitudinal vehicle speed and driver activity are selected as time-varying system parameters. The T-S fuzzy control method is used to construct a fuzzy control system model, and the robust H∞ control algorithm is adopted to design a parallel distributed compensation controller;

[0008] An approximate delay estimation method is adopted, and the gradient descent algorithm is used to design a real-time online network delay estimator with minimizing the change of control input as the performance index;

[0009] The real-time network delay is equivalent to the input delay, the system delay dynamic equation is constructed, and the forward state prediction feedback technology is used to compensate the network time delay, and a new zero-delay augmented human-vehicle-road system model is constructed. Based on this augmented system model, a real-time online synthesized control law is obtained.

[0010] Furthermore, a vehicle-road augmented system model is constructed, and a two-point driver preview model is constructed by introducing a near preview point and a far preview point. Based on the vehicle-road augmented system model and the two-point driver preview model, a driver-in-the-loop human-vehicle-road system model is constructed as follows:

[0011] (21) Vehicle-road augmented system modeling:

[0012] The vehicle-road augmented system model including the vehicle dynamics, EPS dynamics, and path tracking kinematic model can be expressed as:

[0013]

[0014] The state vector in the formula is defined as The system disturbance is defined as w = [f w ρ r T , the system input is the sum of the driver's steering torque T h and the auxiliary steering torque T a ; where υ y represents the vehicle lateral speed, γ represents the yaw rate, ψ R represents the heading angle deviation, y R represents the lateral position deviation, δ sw and respectively represent the steering wheel angle and the change rate, f w represents the lateral wind force, ρ r represents the road curvature;

[0015] The vehicle-road augmented system matrix is as follows:

[0016]

[0017]

[0018] ​where m is the total vehicle mass, R s is the steering ratio, υ y is the longitudinal vehicle speed, l p is the preview distance, l f and l r are the distances from the center of mass to the front and rear axles respectively, l w is the equivalent lateral wind arm, λ t is the tire contact length, J z is the vehicle yaw moment of inertia, C f and C r are the cornering stiffnesses of the front and rear wheels respectively, D s is the steering system damping, I s is the steering system moment of inertia.

[0019] (22) Driver-in-the-loop system modeling:

[0020] By introducing the near preview point and the far preview point, the compensation and anticipation characteristics of the driver can be effectively described. Based on this anthropomorphic driver model, the driver output torque is expressed as:

[0021] T h = K p θ np + K c θ fp (Equation 2)

[0022] where K p and K c are the control gains obtained by fitting the real driver cab data;

[0023] Among them, based on the two-point driver preview model, the near-point view angle θ np and the far-point view angle θ fp can be expressed as:

[0024] θ fp = σ1υ y + σ2γ + σ3δ sw

[0025]

[0026] where θ np is the near-point view angle, θ fp is the far-point view angle, t p is the preview time, τ d is the driver anticipation time, and σ1, σ2 and σ3 are all intermediate variables;

[0027] Combining Equation 2 and Equation 3, the driver output torque can be expressed as:

[0028] Th = T h1 υ y + T h2 γ + K p ψ R + T h3 y R + T h4 δ sw (Equation 4)

[0029] Where T h1 , T h2 , T h3 and T h4 are all intermediate variables, and T h1 = K c σ1, T h2 = K c σ2, T h3 = K p / (υ x t p ), T h4 = K c σ3;

[0030] After combining Equation 4 and Equation 1, the driver-in-the-loop human-vehicle-road system model can be described as:

[0031]

[0032] Among them,

[0033] Where,

[0034] Furthermore, based on the human-vehicle-road system model, driver activity is introduced to characterize the driver's driving intention and driving ability, and the automated assistance torque is adaptively adjusted according to the driver activity, resulting in a human-vehicle-road control system adopting a haptic shared steering strategy, specifically as follows:

[0035] (31) To alleviate the man-machine conflict and reduce the driver's workload, the driver activity can be modeled as:

[0036]

[0037] Where DS ∈ [0, 1] represents the driver's driving intention, and the driver's steering torque T hN = |T h / T h max | directly affects the driving performance and is used to represent the driving ability to complete a specific driving task; T h max is the maximum steering torque output by the driver, and μ1 = 2, μ2 = μ3 = 3 are weight factors;

[0038] (32) Adjust the automated assistance torque adaptively according to the driver's activity, and model the assisted steering torque as:

[0039]

[0040] Wherein, is the introduced virtual torque to be designed, and η(ε) is a weight function formulated to adapt to the change of the driver's driving load, and is specifically designed as:

[0041] η(ε) = k1(ε - k2) 2 + η min (Formula 8)

[0042] Wherein, the relevant parameters are k1 = 3.2, k2 = 0.5, η min = 0.2. Thus, η(ε) can represent the degree of need for automated assistance;

[0043] Combining Formulas 5 to 8, the human-vehicle-road control system adopting the tactile shared steering strategy can be expressed as:

[0044]

[0045] Wherein, B u = [0 0 0 0 0 η(ε) / I s T .

[0046] Furthermore, according to the tactile shared steering strategy, select the longitudinal vehicle speed and the driver's activity as time-varying system parameters, construct a fuzzy control system model using the T-S fuzzy control method, and use the robust H∞ control algorithm to design a parallel distributed compensation controller, specifically as follows:

[0047] (41) To achieve shared steering control, the control output of the human-vehicle-road control system in Formula 9 can be defined as:

[0048]

[0049] Wherein, the lateral acceleration a y and the yaw rate are related to driving comfort, the tracking accuracy is represented by the near-point view angle θ np and the far-point view angle θ fp , the steering wheel steering rate is a driving comfort index related to the response of the steering system, and introducing T h - T a can reflect the human-machine interaction conflict; representing the defined control output vector with the relevant vectors in Formula 9, we can obtain:

[0050] z = Cυr x υr +D u u+E w w (Formula 11)

[0051] In the formula,

[0052] In the formula, υ x and η(ε) are time-varying system parameters, and the actual variation ranges of these two time-varying parameters are bounded, that is, υ x min ≤υ x ≤υ x max , η max ≤η(ε)≤η max ;

[0053] (42) The T-S fuzzy control method is adopted to transform the nonlinear problem into a linear problem for processing:

[0054] (42.1) Define So 1 / υ x and υ x can be expressed by Taylor expansion as:

[0055]

[0056] In the formula, is the adjustment factor;

[0057] (42.2) Select the fuzzy variable as Adopt the sector nonlinear method, and the human-vehicle-road control system in Formula 9 can be described by the following four subsystems:

[0058]

[0059] Its corresponding membership function can be defined as:

[0060]

[0061] Among them,

[0062] (42.3) According to the T-S fuzzy control method, combining Formula 9 and Formula 11, the following control system is obtained:

[0063]

[0064] The control law of this fuzzy control system can be given in the form of parallel distributed compensation, that is

[0065] (42.4) According to the H ∞ control theory, define The performance index ||z||2 ≤ Υ||w||2, where Υ represents the degree of disturbance rejection. Then, the parallel distributed compensation controller can be designed according to the following theorem:

[0066] For the fuzzy control system described by Equation 12, given a constant positive definite coefficient Υ > 0, if there exist positive definite symmetric matrices P and Q i , i = 1, 2, 3, 4, such that the following inequality holds

[0067]

[0068] Then, the control law of each subsystem is calculated according to K i = Q i P -1 .

[0069] Furthermore, an approximate delay estimation method is adopted, and a gradient descent algorithm is used to design a real-time online network delay estimator with the performance index of minimizing the change of control input, as follows:

[0070] (51) Assume that the delay is slowly varying. Then, the delay estimation is achieved by minimizing the following performance index:

[0071]

[0072] where τ L (t) and are the true and estimated delay times, respectively, and τ L (t) ∈ [τ min , τ max ;

[0073] According to the gradient descent algorithm, the delay estimation dynamic equation can be expressed as:

[0074]

[0075] where is the adjustment factor;

[0076] Considering that:

[0077]

[0078] Then, it can be deduced that:

[0079]

[0080] where τ P (t) is the projection value of the delay estimation, and this projection ensures that τ L (t) ∈ [τ min , τ max ;

[0081]

[0082] Among them,

[0083] In the formula, τ min and τ max represent the minimum and maximum values of the time delay respectively;

[0084] (52) The designed delay estimator above can ensure convergence, and for general noise signals, the alternative delay estimation dynamic equation can be expressed as:

[0085]

[0086] In the formula, t w > 0 is the time window used to adjust the approximation performance;

[0087] In addition, if the control input is rapidly changing, then the following normalization formula can be given:

[0088]

[0089] In the formula, and ν > 0 is the adjustment factor.

[0090] Furthermore, the real-time network delay is equivalent to the input delay, the system delay dynamic equation is constructed, and the forward state prediction feedback technology is used to compensate for the network delay, and a new augmented human-vehicle-road system model with zero time delay is constructed. Based on this augmented system model, the real-time online synthesized control rate is obtained as follows:

[0091] (61) Considering the network delay as the equivalent input delay, the fuzzy control system formula 12 with the input delay introduced can be expressed as:

[0092]

[0093] u(t) = u0(t) for all t ∈ [—τ, 0]

[0094] x(0) = x0 (Formula 21)

[0095] In the formula u(t), w(t) are bounded and measurable, and τ is the known and constant network delay;

[0096] (62) According to Formula 21, for all t ≥ τ L At the moment, the state prediction at time t for time t + τ L At the moment can be modeled as:

[0097] χ(t) = ζ(t) + x(t) - ζ(t - τ L) (Formula 22)

[0098] Wherein,

[0099] In the formula, ζ(t) is the predicted state, and χ(t) is the system composite state vector after time-delay compensation;

[0100] (63) To simplify the solution difficulty, according to the Austin order reduction method, the system described by Formula 21 can be transformed into:

[0101]

[0102] Meanwhile, the system controller is Thus, stabilizing the composite state χ(t) can stabilize the actual state x(t).

[0103] According to the second aspect of the present invention, an autonomous vehicle sharing remote driving control system is used to implement the above-mentioned autonomous vehicle sharing remote driving control method, including:

[0104] A shared remote driving system construction module, configured to construct a vehicle-road augmented system model, and construct a two-point driver preview model by introducing a near preview point and a far preview point, and construct a driver-in-the-loop human-vehicle-road system model based on the vehicle-road augmented system model and the two-point driver preview model;

[0105] A tactile shared steering control module, configured to introduce driver activity to characterize the driving intention and driving ability of the driver based on the human-vehicle-road system model, and adaptively adjust the automated assistance torque according to the driver activity to obtain a human-vehicle-road control system adopting a tactile shared steering strategy;

[0106] A fuzzy shared controller design module, configured to select the longitudinal vehicle speed and driver activity as time-varying system parameters according to the tactile shared steering strategy, construct a fuzzy control system model by using the T-S fuzzy control method, and design a parallel distributed compensation controller by using the robust H∞ control algorithm;

[0107] A time-varying network delay estimation module, configured to design a real-time online network delay estimator by using an approximate delay estimation method and the gradient descent algorithm with the minimization of the control input change as the performance index;

[0108] A constant delay predictive control module, configured to equivalently convert the real-time network delay into an input delay, construct a system delay dynamics equation, and compensate for the network time delay by using the forward state prediction feedback technology to construct a new zero-delay human-vehicle-road augmented system model, and obtain a real-time online synthesized control law based on this augmented system model.

[0109] According to a third aspect of the present invention, the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The computer program stored in the memory can run on the processor. When the processor loads and executes the computer program, the above-mentioned remote driving control method for shared autonomous vehicles is adopted.

[0110] According to a fourth aspect of the present invention, the present invention provides a storage medium containing computer-executable instructions, and the computer-executable instructions are used to execute the above-mentioned remote driving control method for shared autonomous vehicles when executed by a computer processor.

[0111] According to a fifth aspect of the present invention, the present invention provides a computer program product, and the computer program product includes a computer program. When the computer program is executed by a processor, it is used to load and execute the above-mentioned remote driving control method for shared autonomous vehicles.

[0112] The present invention has at least the following beneficial effects:

[0113] 1. The present invention constructs a brand-new remote sharing control architecture for autonomous vehicles, breaks through the technical difficulties of large wireless transmission network delay and decreased scene perception in remote driving, effectively improves the remote driving performance, realizes the expansion of the feasible design domain of autonomous vehicles, and promotes the implementation and popularization of autonomous driving technology.

[0114] 2. The present invention fully considers the randomness, non-linearity, personalization and other situations of the driver's manipulation behavior, and designs a tactile sharing control strategy based on the dynamic interaction between the driver and the automation system, avoiding the problems of both competing for driving authority and even man-machine confrontation, and improving the safety of vehicle driving.

[0115] Of course, it is not necessary for any product implementing the present invention to achieve all the above-mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0116] Figure 1 It is a schematic flowchart of the control method described in the present invention;

[0117] Figure 2 It is a schematic diagram of the control principle of the control method described in the present invention;

[0118] Figure 3 It is a schematic diagram of the control architecture of the control system described in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0119] Next, the technical solutions in the embodiments of the present disclosure will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without creative efforts belong to the scope of protection of the present disclosure.

[0120] Please refer to Figure 1 and Figure 2 , the present invention provides a technical solution: an autonomous vehicle sharing remote driving control method, including the following steps:

[0121] Step S1. Modeling of the shared remote driving system: Construct a vehicle-road augmented system model, and construct a two-point driver preview model by introducing a near preview point and a far preview point. Based on the vehicle-road augmented system model and the two-point driver preview model, a driver-in-the-loop human-vehicle-road system model is constructed;

[0122] As Figure 2 shown, the driver operates the vehicle at the console. The output control commands (acceleration, deceleration, and steering) after interacting with the shared controller are wirelessly transmitted to the remote vehicle to achieve vehicle motion control. The system state of the vehicle is then wirelessly transmitted into the predictor for compensation. One end is input to the display to provide the driver with scene perception, and the other end is input to the shared controller for assisted driving control;

[0123] Step S11: Modeling of the vehicle-road system

[0124] The vehicle-road augmented system model including vehicle dynamics, EPS dynamics, and path tracking kinematics models can be expressed as:

[0125]

[0126] In the formula, the state vector is defined as The system disturbance is defined as w = [f w ρ r T , the system input is the sum of the driver's steering torque T h and the auxiliary steering torque T a ; where υ y represents the lateral velocity of the vehicle, γ represents the yaw angular velocity, ψ R represents the heading angle deviation, y R represents the lateral position deviation, δ sw and respectively represent the steering wheel angle and the change rate, f w represents the lateral wind force, ρ r represents the road curvature;

[0127] ​The vehicle-road augmentation system matrix is as follows:

[0128]

[0129] In the formula, m is the total vehicle mass, R s is the steering ratio, υ y is the longitudinal vehicle speed, l p is the preview distance, l f and l r are the distances from the center of mass to the front and rear axles respectively, l w is the equivalent lateral wind arm, λ t is the tire contact length, J z is the vehicle yaw moment of inertia, C f and C r are the cornering stiffnesses of the front and rear wheels respectively, D s is the steering system damping, I s is the moment of inertia of the steering system;

[0130] Step S12: In-the-loop driver system modeling

[0131] The driving behavior of a human driver when performing a path-tracking driving task can be described by a two-point preview driver model. By introducing a near preview point and a far preview point, the compensation and anticipation characteristics of the driver can be effectively described. Based on this anthropomorphic driver model, the driver output torque can be expressed as:

[0132] T h = K p θ np + K c θ fp (Formula 2)

[0133] In the formula, K p and K c are control gains fitted according to the real driver cab data;

[0134] Among them, based on the two-point driver preview model, the near-point view angle θ np and the far-point view angle θ fp can be expressed as:

[0135] θ fp = σ1υ y + σ2γ + σ3δ sw

[0136]

[0137] In the formula, θ np is the near-point view angle, θ fp is the far-point view angle, t p is the preview time, τd is the driver's expected time, and σ1, σ2, and σ3 are all intermediate variables;

[0138] Combining Equation 2 and Equation 3, the driver's output torque can be expressed as:

[0139] T h = T h 1υ y + T h2 γ + K p ψ R + T h3 y R + T h4 δ sw (Equation 4)

[0140] where T h1 , T h2 , T h3 and T h4 are all intermediate variables, and T h1 = K c σ1, T h2 = K c σ2, T h3 = K p / (υ x t p ), T h4 = K c σ3;

[0141] After combining Equation 4 and Equation 1, the in-loop driver-vehicle-road system model can be described as:

[0142]

[0143] where,

[0144] In the formula,

[0145] Step S2. Tactile shared steering control strategy: Based on the driver-vehicle-road system model, the driver's activity is introduced to characterize the driver's driving intention and driving ability, and the automated assistance torque is adaptively adjusted according to the driver's activity to obtain a driver-vehicle-road control system adopting the tactile shared steering strategy, as follows:

[0146] S21. To alleviate the human-machine conflict and reduce the driver's workload, the driver's activity can be modeled as:

[0147]

[0148] where \(DS\in[0,1]\) represents the driving intention of the driver, and its real-time data can be obtained from the driver monitoring system. The normalized driver's steering torque \(T\) hN \( = |T\) h \( / T\) h max \(|\) directly affects the driving performance, which can represent the driving ability to complete a specific driving task. \(T\) h max is the maximum steering torque output by the driver, and \(\mu_1 = 2\), \(\mu_2=\mu_3 = 3\) are weighting factors;

[0149] S22. Adaptively adjust the automated assistance torque according to the driver's activity, and model the assisted steering torque as:

[0150]

[0151] where, is the introduced virtual torque to be designed, and \(\eta(\varepsilon)\) is a weighting function formulated to adapt to the change of the driver's driving load, and is specifically designed as:

[0152] \(\eta(\varepsilon)=k_1(\varepsilon - k_2)\) 2 \(+\eta\) min (Equation 8)

[0153] where the relevant parameters \(k_1 = 3.2\), \(k_2 = 0.5\), \(\eta\) min \( = 0.2\). Thus, \(\eta(\varepsilon)\) can represent the degree of need for automated assistance;

[0154] Combining Equation 5 to Equation 8, the human-vehicle-road control system adopting the tactile shared steering strategy can be expressed as:

[0155]

[0156] where, \(B\) u \(=[0\ 0\ 0\ 0\ 0\ \eta(\varepsilon) / I\) s \) T ;

[0157] Step S3. Fuzzy shared controller design: According to the tactile shared steering strategy, select the longitudinal vehicle speed and the driver's activity as time-varying system parameters, adopt the T-S fuzzy control method to construct a fuzzy control system model, and adopt the robust \(H_{\infty}\) control algorithm to design a parallel distributed compensation controller, specifically as follows:

[0158] S31. To achieve shared steering control, the control output of the system shown in Equation 9 can be defined as:

[0159]

[0160] where the lateral acceleration \(a\) y and the yaw rate Related to driving comfort, the tracking accuracy is represented by the near-point view angle θ np and the far-point view angle θ fp The steering wheel steering rate is a driving comfort index related to the response of the steering system. Introducing T h -T a can reflect the human-machine interaction conflict; the defined control output vector can be represented by the relevant vector in Equation 9, and we can get:

[0161] z = C υr x υr + D u u + E w w (Equation 11)

[0162] In the formula,

[0163] It should be noted that υ x and η(ε) are time-varying system parameters, which will make the control problems described by Equations 9 and 11 become nonlinear problems. Notice that the actual variation ranges of these two time-varying parameters are bounded, that is, υ x min ≤ υ x ≤ υ x max , η max ≤ η(ε) ≤ η max ;

[0164] S31. Use the T-S fuzzy technology to transform the nonlinear problem into a linear problem for processing:

[0165] S31.1 Define So 1 / υ x and υ x can be expressed by Taylor expansion as:

[0166]

[0167] In the formula, is the adjustment factor;

[0168] S31.2 Select the fuzzy variable as Using the sector nonlinear method, the human-vehicle-road control system in Equation 9 can be described by the following four subsystems:

[0169]

[0170] Its corresponding membership function can be defined as:

[0171]

[0172] Among them,

[0173] S31.3 According to the T-S fuzzy control method, combining Formula 9 and Formula 11, the following control system is obtained:

[0174]

[0175] The control law of this fuzzy control system can be given in the form of parallel distributed compensation, that is

[0176] S31.4 According to the H∞ control theory, define The performance index ||z||2 ≤ Υ||w||2, where Υ represents the degree of disturbance suppression. Then the parallel distributed compensation controller can be designed according to the following theorem:

[0177] For the fuzzy control system described by Formula 12, given a constant positive definite coefficient Υ > 0, if there exist positive definite symmetric matrices P and Q i , i = 1, 2, 3, 4, such that the following inequality holds

[0178]

[0179] Then the control law of each subsystem is calculated according to K i = Q i P -1 ;

[0180] Step S4. Time-varying network delay estimation: Adopt the approximate delay estimation method, use the gradient descent algorithm, and design a real-time online network delay estimator with minimizing the change of control input as the performance index, specifically as follows:

[0181] S41. Assume that the delay is slowly changing, then the delay estimation is achieved by minimizing the following performance index:

[0182]

[0183] where τ L (t) and are the true and estimated delay times respectively, and;

[0184] According to the gradient descent algorithm, the delay estimation dynamic equation can be expressed as:

[0185]

[0186] where is the adjustment factor;

[0187] Considering:

[0188]

[0189] Then it can be deduced that:

[0190]

[0191] where τ P (t) is the projected value of the delay estimate, and this projection ensures that τ L (t) ∈ [τ min , τ max ;

[0192]

[0193] where

[0194] in the formula, τ min and τ max represent the minimum and maximum values of the time delay respectively, z(t) is the predicted state, and c(t) is the system composite state vector after time delay compensation;

[0195] S4.2 The delay estimator designed above can ensure convergence, and for general noise signals, the alternative delay estimation dynamic equation can be expressed as:

[0196]

[0197] where t w > 0 is the time window for adjusting the approximation performance;

[0198] In addition, if the control input is fast-changing, then the following normalization formula can be given:

[0199]

[0200] where and ν > 0 is the adjustment factor; Step S5. Steady-state delay predictive control: Equivalent the real-time network delay to an input delay, construct the system delay dynamic equation, and use the forward state prediction feedback technology to compensate for the network time delay, construct a new augmented human-vehicle-road system model with zero time delay, and obtain the real-time online synthesized control law based on this augmented system model, so as to make the actual state of the vehicle stable, specifically as follows:

[0201] S51. Consider the network delay as an equivalent input delay, then the fuzzy system formula introducing the input delay

[0202] Equation 12 can be expressed as:

[0203]

[0204] u(t) = u0(t) for all t ∈ [-τ, 0]

[0205] x(0) = x0 (Formula 21)

[0206] wherein u(t) and w(t) are bounded and measurable, and τ is a known constant network delay;

[0207] S52. Then according to Equation 21, for all t≥τ L moments, at time t, the state prediction for t+τ L moment can be modeled as:

[0208] χ(t) = ζ(t) + x(t) - ζ(t - τ L ) (Equation 22)

[0209] where

[0210] wherein, ζ(t) is the predicted state, and χ(t) is the system composite state vector after time-delay compensation;

[0211] To simplify the solution difficulty, according to the Austin order reduction method, the system described by Equation 21 can be transformed into:

[0212]

[0213] Meanwhile, the system controller is so that the stability of the composite state χ(t) can make the actual state x(t) stable.

[0214] See Figure 2 and Figure 3 It can be seen that the present invention designs a shared remote control architecture based on real-time network delay estimation and compensation, introduces driver activity to characterize the driving behavior of the driver, thereby modeling real-time human-machine interaction, and using this as the basis for sharing control authority allocation; adopts the T-S fuzzy method to process the vehicle speed and driver activity that change with time; constructs an online delay estimator to estimate the unknown time-varying network delay, and develops a predictor to compensate for the state delay, and then designs a predictor-based shared remote fuzzy control method, which can effectively improve the lane keeping performance in a dynamic environment and reduce the mental fatigue and driving load of the remote driver.

[0215] In summary, the present invention constructs a brand-new remote shared control architecture for autonomous driving vehicles, breaks through the technical difficulties of large wireless transmission network delay and decreased scene perception in remote driving, effectively improves the remote driving performance, realizes the expansion of the feasible design domain of autonomous driving vehicles, and promotes the implementation and popularization of autonomous driving technology; in addition, this method fully considers the randomness, non-linearity, personalization and other situations of the driver's operation behavior, designs a tactile shared control strategy based on the dynamic interaction between the driver and the automation system, avoids the problem of both competing for driving authority or even human-machine confrontation, and improves the driving safety of the vehicle.

[0216] Embodiment 2:

[0217] This embodiment provides an autonomous vehicle sharing remote driving control system for implementing the above-mentioned autonomous vehicle sharing remote driving control method, including:

[0218] A shared remote driving system construction module, configured to construct a vehicle-road augmented system model, and construct a two-point driver preview model by introducing a near preview point and a far preview point, and construct a driver-in-the-loop human-vehicle-road system model based on the vehicle-road augmented system model and the two-point driver preview model;

[0219] A tactile shared steering control module, configured to introduce driver activity to characterize the driving intention and driving ability of the driver based on the human-vehicle-road system model, and adaptively adjust the automated assistance torque according to the driver activity to obtain a human-vehicle-road control system adopting a tactile shared steering strategy;

[0220] A fuzzy shared controller design module, configured to select the longitudinal vehicle speed and driver activity as time-varying system parameters according to the tactile shared steering strategy, construct a fuzzy control system model using the T-S fuzzy control method, and design a parallel distributed compensation controller using the robust H∞ control algorithm;

[0221] A time-varying network delay estimation module, configured to design a real-time online network delay estimator using an approximate delay estimation method and the gradient descent algorithm with minimizing the change of control input as the performance index; a constant delay predictive control module, configured to equivalently convert the real-time network delay into an input delay, construct a system delay dynamics equation, and compensate for the network time delay using the forward state prediction feedback technique to construct a new zero-delay human-vehicle-road augmented system model, and obtain a real-time online synthesized control law based on this augmented system model.

[0222] Specifically, the above-mentioned shared remote driving system construction module, tactile shared steering control module, fuzzy shared controller design module, time-varying network delay estimation module, and constant delay predictive control module can be embedded in a computer processing system. The computer calls the above-mentioned modules to complete the task of stably controlling the shared remote driving system according to the above-provided autonomous vehicle sharing remote driving control method; the above-mentioned shared remote driving system construction module, tactile shared steering control module, fuzzy shared controller design module, time-varying network delay estimation module, and constant delay predictive control module can perform operations according to the specific steps given by the autonomous vehicle sharing remote driving control method.

[0223] It should be noted that the division of each module of the above system is only a division of logical functions. In actual implementation, it can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also all be implemented in the form of hardware; or some modules can be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the shared remote driving system construction module can be a separately established processing element, or can be integrated in a certain chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and the function of the above signal processing module can be called and executed by a certain processing element of the above device. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together or can be independently implemented. The processing element mentioned here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the integrated logic circuit of the hardware in the processor element or the instruction in the form of software.

[0224] For example, the above modules can be one or more integrated circuits configured to implement the above method, such as: one or more Application Specific Integrated Circuits (ASICs), or, one or more Digital Signal Processors (DSPs), or, one or more Field Programmable Gate Arrays (FPGAs), etc. Again, when a certain module above is implemented in the form of a processing element scheduling program code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processors that can call program code. Again, these modules can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0225] Embodiment 3:

[0226] The present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores a computer program capable of running on the processor. When the processor loads and executes the computer program, the above-mentioned shared remote driving control method for autonomous driving vehicles is adopted.

[0227] It should be noted that the terminal device can be a computer device such as a desktop computer, a laptop computer, or a cloud server, and the terminal device includes, but is not limited to, a processor and a memory. For example, the terminal device may further include input / output devices, network access devices, and a bus, etc.

[0228] Furthermore, the processor can be a central processing unit (CPU). Of course, according to the actual usage situation, other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be used. The general-purpose processor can be a microprocessor or any conventional processor, etc. This application does not make any restrictions in this regard.

[0229] Embodiment Four:

[0230] The present invention provides a storage medium containing computer-executable instructions, and the computer-executable instructions are used to execute the above-mentioned shared remote driving control method for autonomous vehicles when executed by a computer processor.

[0231] Among them, the computer program can be stored in a computer-readable medium. The computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some middleware form, etc. The computer-readable medium includes any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the computer-readable medium includes, but is not limited to, the above-mentioned components.

[0232] Embodiment Five:

[0233] The present invention provides a computer program product, and the computer program product includes a computer program. When the computer program is executed by a processor, it is used to load and execute the above-mentioned shared remote driving control method for autonomous vehicles.

[0234] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device.

[0235] For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. When an element is referred to as being "assembled on", "mounted on", "fixed to" or "disposed on" another element, it can be directly on the other element or there may also be an intermediate element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be an intermediate element at the same time. The terms "vertical", "horizontal", "upper", "lower", "left", "right" and similar expressions used herein are for illustrative purposes only and do not represent the only embodiments.

[0236] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

[0237] In the description of this specification, the description with reference to terms such as "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

Claims

1. A remote driving control method for sharing autonomous driving vehicles, characterized in that: The following steps are involved: A vehicle-road augmentation system model is constructed, and a two-point driver preview model is constructed by introducing near preview points and far preview points. Based on the vehicle-road augmentation system model and the two-point driver preview model, a human-vehicle-road system model with a driver in the loop is constructed. Based on the human-vehicle-road system model, the driver's activity is introduced to characterize the driver's driving intention and driving ability, and the automated auxiliary torque is adaptively adjusted according to the driver's activity, thus obtaining a human-vehicle-road control system with a tactile shared steering strategy. According to the tactile shared steering strategy, the longitudinal vehicle speed and driver activity are selected as time-varying system parameters, the TS fuzzy control method is used to build the fuzzy control system model, and the robust H∞ control algorithm is used to design the parallel distributed compensation controller. By adopting the approximate delay estimation method and using the gradient descent algorithm, a real-time online network delay estimator is designed with minimizing the control input change as the performance indicator. The real-time network delay is equated to the input delay, the system delay dynamics equation is constructed, and the forward state prediction feedback technology is used to compensate for the network delay. A new human-vehicle-road augmented system model with zero delay is constructed, and the real-time online synthetic control rate is obtained based on the augmented system model.

2. The autonomous driving vehicle shared remote driving control method according to claim 1, characterized in that: A vehicle-road augmentation system model is constructed, and a two-point driver preview model is constructed by introducing near preview points and far preview points. Based on the vehicle-road augmentation system model and the two-point driver preview model, a human-vehicle-road system model with a driver in the loop is constructed, as follows: (21) Vehicle-road augmentation system modeling: The vehicle-road augmentation system model including vehicle dynamics, EPS dynamics and path tracking kinematics models can be expressed as: The state vector is defined as System disturbance is defined as w = [f w ρ r ] T , the system input is the driver's steering torque T h and auxiliary steering torque T a The sum of y represents the lateral velocity of the vehicle, γ represents the yaw rate, ψ R Represents the heading angle deviation, y R represents the lateral position deviation, δ sw and Respectively represent the steering wheel angle and rate of change, f w represents the lateral wind force, ρ r represents the road curvature; The vehicle-road augmentation system matrix is ​​as follows: Where m is the total mass of the vehicle, R s is the steering ratio, v y is the longitudinal speed, l p is the preview distance, l f and l r are the distances from the center of mass to the front and rear axles, l w is the equivalent lateral wind arm, λ t is the tire contact length, J z is the vehicle yaw moment of inertia, C f and C r are the front and rear wheel cornering stiffness, D s is the steering system damping, I s is the moment of inertia of the steering system; (22) Driver-in-the-loop system modeling: By introducing near and far preview points, the driver's compensation and expectation characteristics can be effectively described. Based on this anthropomorphic driver model, the driver's output torque is expressed as: T h = K p θ np + K c θ fp (Formula 2) Where K p and K c It is the control gain obtained by fitting the real driver's cab data; Among them, based on the two-point driver preview model, the near point viewing angle θ np and the far point viewing angle θ fp can be expressed as: Where θ np is the near-point viewing angle, θ fp is the far point perspective, t p is the preview time, τ d is the driver's expected time, σ1, σ2 and σ3 are intermediate variables; Combining Formula 2 and Formula 3, the driver output torque can be expressed as: T h = T hl v y + T h2 γ + K p ψ R + T h3 y R + T h4 δ sw (Formula 4) Where T h1 ,T h2 ,T h3 and T h4 are all intermediate variables, and T h1 =K c σ1,T h2 =K c σ2, T h3 =K p / (v x t p ), T h4 =K c σ3; Combining Formula 4 with Formula 1, the human-vehicle-road system model with a driver in the loop can be described as: in, In the formula, 3. The autonomous driving vehicle shared remote driving control method according to claim 2, characterized in that: Based on the human-vehicle-road system model, the driver's activity is introduced to characterize the driver's driving intention and driving ability, and the automated auxiliary torque is adaptively adjusted according to the driver's activity, resulting in a human-vehicle-road control system using a tactile shared steering strategy, as follows: (31) To alleviate human-machine conflict and reduce the driver’s workload, the driver’s activity can be modeled as: Where DS∈[0,1] represents the driver's driving intention, and the driver's steering torque T hN =|T h / T hmax |Directly affects driving performance and is used to represent the driving ability to complete a specific driving task; T hmax is the maximum steering torque output by the driver, μ1=2, μ2=μ3=3 are weight factors; (32) The automated assist torque is adaptively adjusted according to the driver's activity, and the assist steering torque is modeled as: In the formula, is the virtual torque to be designed, and η(ε) is a weight function formulated to adapt to the change of the driver's driving load, which is specifically designed as: η(ε) = k1(ε - κ2) 2 + η min (Equation 8) In the formula, the relevant parameters κ1=3.2, κ2=0.5, η min = 0.2, thus, η(ε) can represent the degree of need for automated assistance; Combining Formula 5 to Formula 8, the human-vehicle-road control system using a tactile shared steering strategy can be expressed as: Among them, B u =[0 0 0 0 0 η(ε) / I s ] T 。 4. The autonomous driving vehicle shared remote driving control method according to claim 3, characterized in that: According to the tactile shared steering strategy, the longitudinal vehicle speed and driver activity are selected as time-varying system parameters, the TS fuzzy control method is used to build a fuzzy control system model, and the robust H∞ control algorithm is used to design a parallel distributed compensation controller, as follows: (41) To achieve shared steering control, the control output of the human-vehicle-road control system in Formula 9 can be defined as: Where lateral acceleration a y and yaw rate Related to driving comfort, tracking accuracy is determined by the near-point viewing angle θ np and the far point viewing angle θ fp Indicates that the steering wheel turning rate It is a driving comfort index related to the steering system response. h -T a It can reflect the human-computer interaction conflict; the control output vector defined by this definition is expressed by the related vector in formula 9, and we can get: z = C vr x vr + D u u + E w w (Formula 11) In the formula, Where v x and η(ε) are time-varying system parameters. The actual variation range of these two time-varying parameters is bounded, that is, v xmin ≤v x ≤v xmax , η max ≤η(ε)≤η max ; (42) The TS fuzzy control method is used to transform the nonlinear problem into a linear problem for processing: (42.1) Definition So 1 / v x and v x Taylor expansion can be applied to express it as: In the formula, is the regulating factor; (42.2) Select the fuzzy variables as Using the sector nonlinear method, the human-vehicle-road control system in Equation 9 can be described by the following four subsystems: The corresponding membership function can be defined as: in, (42.3) According to the TS fuzzy control method, combined with formula 9 and formula 11, the following control system is obtained: The control rate of the fuzzy control system can be given by the parallel distributed compensation form, that is, (42.4) According to H∞ control theory, we define The performance index ||z||2≤γ||w||2, where Υ represents the degree of disturbance suppression, the parallel distributed compensation controller can be designed according to the following theorem: For the fuzzy control system described by formula 12, given a constant positive definite coefficient γ>, if there exist positive definite symmetric matrices P and Q i , i = 1, 2, 3, 4, so that the following inequality holds Then the control rate of each subsystem is based on K i =Q i P -1 Calculated.

5. The autonomous driving vehicle shared remote driving control method according to claim 4, characterized in that: By adopting the approximate delay estimation method and using the gradient descent algorithm, the real-time online network delay estimator is designed with minimizing the control input change as the performance indicator, as follows: (51) Assuming that the delay varies slowly, the delay estimation is achieved by minimizing the following performance metric: Where τ L (t) and are the real and estimated delay times, and τ L (t)∈[τ min , τ max ]; According to the gradient descent algorithm, the delay estimation dynamics equation can be expressed as: In the formula is the regulating factor; Considering: Then it can be deduced that: Among them, τ P (t) is the projection value of the delay estimate, and the projection guarantees τ L (t)∈[τ min , τ max ]; in, In the formula, τ min and τ max Respectively represent the minimum and maximum delay; (52) The delay estimator designed above can guarantee convergence, and for general noise signals, the alternative delay estimation dynamic equation can be expressed as: Where t w >0 is the time window, used to adjust the approximation performance; In addition, if the control input is changing rapidly, the following normalized formula can be given: In the formula, And ν>0 is the adjustment factor.

6. The autonomous driving vehicle shared remote driving control method according to claim 5, characterized in that: The real-time network delay is equivalent to the input delay, the system delay dynamic equation is constructed, and the forward state prediction feedback technology is used to compensate for the network delay. A new zero-delay human-vehicle-road augmented system model is constructed. Based on the augmented system model, the real-time online synthetic control rate is obtained, as follows: (61) Considering the network delay as an equivalent input delay, the fuzzy control system formula (12) with input delay can be expressed as: In the formula u(t), w(t) are bounded and measurable, and τ is a known and constant network delay; (62) According to formula 21, for all t ≥ τ L moment, at time t, for t+τ L The state prediction at each moment can be modeled as: χ(t) = ζ(t) + x(t) - ζ(t - τ L ) (Equation 22) in, Where ζ(t) is the predicted state, χ(t) is the system composite state vector after delay compensation; (63) To simplify the solution difficulty, according to the Austin reduction method, the system described by formula 21 can be transformed into: At the same time, the system controller is Therefore, stabilizing the composite state χ(t) can make the actual state x(t) stable.

7. An autonomous driving vehicle shared remote driving control system, used to implement the autonomous driving vehicle shared remote driving control method according to any one of claims 1 to 6, characterized in that: include: The shared remote driving system building module is used to build a vehicle-road augmentation system model, and to build a two-point driver preview model by introducing a near preview point and a far preview point. Based on the vehicle-road augmentation system model and the two-point driver preview model, a human-vehicle-road system model with a driver in the loop is constructed. A tactile shared steering control module is used to introduce the driver's activity level to characterize the driver's driving intention and driving ability based on the human-vehicle-road system model, and adaptively adjust the automatic auxiliary torque according to the driver's activity level to obtain a human-vehicle-road control system using a tactile shared steering strategy; The fuzzy shared controller design module is used to select the longitudinal vehicle speed and driver activity as time-varying system parameters according to the tactile shared steering strategy, build the fuzzy control system model using the TS fuzzy control method, and design the parallel distributed compensation controller using the robust H∞ control algorithm; The time-varying network delay estimation module is used to adopt an approximate delay estimation method, use a gradient descent algorithm, and take minimizing the control input change as the performance indicator to design a real-time online network delay estimator; The steady-state delay predictive control module is used to equate the real-time network delay to the input delay, construct the system delay dynamics equation, and use the forward state prediction feedback technology to compensate for the network delay, build a new zero-delay human-vehicle-road augmented system model, and obtain the real-time online synthetic control rate based on the augmented system model.

8. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: The memory stores a computer program that can be run on a processor. When the processor loads and executes the computer program, the autonomous driving vehicle shared remote driving control method described in any one of claims 1 to 6 is adopted.

9. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions, when executed by a computer processor, are used to execute the autonomous driving vehicle shared remote driving control method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that The computer program product includes a computer program, which, when executed by a processor, is used to load and execute the autonomous driving vehicle shared remote driving control method according to any one of claims 1 to 6.

Images