Speed control method and device of unmanned vehicle and unmanned vehicle

Abstract

The application relates to the technical field of unmanned vehicle control, and discloses a speed control method and device of an unmanned vehicle and the unmanned vehicle, which comprises the following steps: acquiring multi-modal sensor data of the surrounding environment of the vehicle; constructing a dynamic vortex field model and a Brownian motion parameter mapping relationship based on the sensor data; dynamically adjusting a sensor acquisition strategy and compressing a prediction time domain according to a vortex field intensity; constructing a joint control equation coupling the vortex field and the Brownian motion parameter; and generating a speed control instruction by solving through variational optimization. Through the technical scheme of fusing the vortex field theory of fluid mechanics and the Brownian motion random mapping, the application realizes unified modeling of traffic flow micro-disturbance and environmental uncertainty, effectively solves the defect that a deterministic model is difficult to adapt to dynamic changes of a complex traffic scene, and enables the speed control of the unmanned vehicle to have a physically explainable response capability to sudden disturbance.

Description

Technical Field

[0001] This invention relates to the field of unmanned vehicle control technology, and in particular to a speed control method, device, and unmanned vehicle for unmanned vehicles. Background Technology

[0002] With the rapid development of artificial intelligence and autonomous driving technologies, driverless vehicles have gradually become a research hotspot in the transportation field. Currently, speed control technology for driverless vehicles mainly relies on environmental data collected by sensors and preset control algorithms to achieve autonomous driving. However, complex and ever-changing road environments, real-time data processing requirements, and obstacle avoidance remain challenges for the technology. Speed ​​control of driverless vehicles not only affects driving efficiency but also directly impacts driving safety.

[0003] Current autonomous vehicle speed control technologies generally employ discretized traffic element modeling methods, treating parameters such as road obstacles, traffic density, and moving targets as independent variables. This modeling approach struggles to characterize the dynamic coupling characteristics of multiple targets in traffic flow. Especially in complex interaction scenarios such as vehicle intrusion and merging, the linear superposition models established by traditional methods cannot accurately represent the nonlinear interaction mechanism between vehicles and the flow field, leading to frequent control command lags or overshoots, severely restricting driving safety in complex urban scenarios.

[0004] At the environmental perception and data processing level, most technical solutions adopt a rigid strategy of fixed scanning frequency and full-time domain prediction, failing to dynamically allocate perception resources according to traffic conditions. This data acquisition mode results in both insufficient data acquisition accuracy in key areas and waste of resources in non-critical areas. At the same time, the excessively long prediction time domain introduces a large amount of uncertainty interference, which significantly reduces the real-time performance of the control model in sudden scenarios. Summary of the Invention

[0005] The purpose of this invention is to provide a speed control method, device, and unmanned vehicle for autonomous vehicles, which solves the problem that deterministic models are difficult to adapt to the dynamic changes in complex traffic scenarios.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A speed control method for an autonomous vehicle includes: Acquire multimodal sensor data of the vehicle's surrounding environment; Based on the sensor data, a dynamic vortex field model and Brownian motion parameter mapping relationship are constructed. The sensor acquisition strategy is dynamically adjusted based on the intensity of the vortex field, and the prediction time domain is compressed. Construct the joint control equations for the coupled vortex field and Brownian motion parameters; Speed ​​control commands are generated by solving a variational optimization problem. Control commands are executed based on dynamic parameter adjustment rules, and model parameters are updated in a closed loop. Furthermore, the core of the dynamic vortex field model is to treat traffic flow as a viscous fluid, quantifying the local vortex intensity distribution through modified fluid dynamics equations. Unlike traditional methods, this scheme embeds dynamic parameters such as vehicle mass and moment of inertia into the fluid characteristic calculations, enabling the model to reflect the active impact of vehicle motion on the surrounding traffic flow in real time. Brownian motion parameter mapping, through thermodynamic analogy, transforms traffic density and speed fluctuations into equivalent temperature parameters, giving physical meaning to random disturbance terms. The variational optimization process, through the principle of energy optimality, jointly optimizes the vehicle's kinetic energy and the vortex field potential energy, generating speed commands that balance smoothness and energy efficiency.

[0007] Preferably, the construction of the dynamic vortex field model includes: The local vorticity distribution is calculated based on the modified Navier-Stokes equations, where the viscosity coefficient is related to the vehicle mass and moment of inertia. Traffic density is mapped to an equivalent temperature parameter to construct vehicle dynamics equations that include random disturbance terms; Navier-Stokes equation calculation formula: in: The instantaneous acceleration of the traffic flow velocity field reflects the dynamic changes in the environment surrounding the vehicle; Local traffic flow velocity vector; t: time; v: Kinematic viscosity coefficient, calculated from fluid density and dynamic viscosity; k: Dimensionless proportionality coefficient; m: Vehicle weight; I: Moment of inertia of the vehicle about its center of mass.

[0008] Furthermore, traditional traffic flow models typically use a fixed viscosity coefficient, which fails to reflect the disturbance effect of vehicle inertia on the surrounding environment. This scheme, through the functional relationship between dynamic viscosity and vehicle parameters, automatically enables heavy vehicles or vehicles with high rotational inertia to exhibit stronger flow field damping characteristics in the model. The mapping logic of equivalent temperature parameters transforms the product of traffic density and average speed into a thermodynamic temperature analogy, allowing the intensity of random disturbance terms to be quantified by the ratio of vehicle mass to flow field parameters, thus achieving a coupled characterization of traffic flow randomness and vehicle inertia.

[0009] Preferably, the dynamic adjustment sensor acquisition strategy includes: When the vortex intensity exceeds a preset threshold, the infrared thermal imaging channel is activated and the high-temperature gradient region is focused. The lidar scanning area is dynamically reduced based on the temperature gradient distribution, and the time domain range of the model prediction control is simultaneously compressed.

[0010] Furthermore, the activation condition of the infrared thermal imaging channel is related to the vortex field intensity threshold. When strong vortex disturbances (such as traffic conflicts at intersections) are detected, the system prioritizes capturing areas with significant temperature gradients (such as vehicle engine heat sources), thereby quickly locating potential risk targets. The dynamic adjustment of the lidar scanning range adopts a nonlinear compression strategy, focusing on the core area in front of the vehicle; the prediction time domain is dynamically shortened according to the vortex intensity to ensure optimized real-time computation in high-disturbance scenarios.

[0011] Preferably, the construction of the joint control equations includes: A partial differential coupling relationship between vortex field intensity and Brownian motion parameters is established, wherein the coupling coefficient is dynamically related to the vehicle yaw rate. Define a multi-objective optimization function that includes acceleration smoothing term and energy consumption gradient term, with its weight coefficients adaptively adjusted according to the intensity of vortex field; Partial differential coupling formula: in: The rate of change of vortex field intensity per unit time; ω: Scalar value of vortex field intensity; Real-time yaw rate of this vehicle; kT: composed of traffic density ρ and average speed The generated equivalent temperature parameters; m: Vehicle weight; Local traffic flow velocity vector; x: Vehicle longitudinal coordinate; Instantaneous acceleration of the traffic flow velocity field.

[0012] Furthermore, the physical quantity calculations on the right-hand side of the equation reflect the impact of random traffic flow disturbances on vehicle acceleration. When the vehicle is turning, the coupling coefficient automatically adjusts the evolution rate of the vortex field intensity, causing the turning action to actively change the characteristics of the surrounding flow field. In the multi-objective optimization function, the weight of the acceleration smoothing term is inversely proportional to the vortex intensity, while the weight of the energy consumption gradient term is directly proportional, achieving an adaptive balance of the control objective with environmental disturbances.

[0013] Preferably, the variational optimization solution includes: Embedding stochastic differential equations into the Hamiltonian action principle: in: δ: Variational operator, representing a small change in the functional of the action, used to solve for the extreme path of the vehicle's trajectory; ∫: Time integration operator; m: Vehicle weight; The vehicle's velocity vector; Integral potential energy of a vortex field; Spatial location The intensity of the vortex field at that location; The volume of a tiny spatial region in the environment surrounding a vehicle; dt: Discrete time step of the control system; Three-dimensional spatial vector; t: time; An adaptive step-size numerical discretization method is used to solve the optimization problem, with the step size dynamically switched according to the modified Reynolds number.

[0014] Furthermore, the construction of the energy-optimal principle breaks through the limitations of traditional single-point velocity optimization, calculating the total potential energy of the global vortex field around the vehicle through three-dimensional spatial integration. The adaptive step size mechanism dynamically adjusts according to the flow field state: a fine step size is used to improve resolution when turbulent characteristics are significant, and a coarse-grained step size is switched to accelerate convergence during the steady flow field stage, achieving synergistic optimization of computational accuracy and efficiency.

[0015] Preferably, the execution control instructions include: The gain coefficient of the proportional-integral-derivative controller is adjusted in real time based on the vortex field intensity and equivalent temperature parameters. The original control command is subjected to an exponential smoothing filter based on the relaxation time. The proportional gain coefficient is adjusted in real time based on the vortex field intensity. in: K p : Current proportional gain coefficient; K p0 The reference proportional gain coefficient is determined through vehicle calibration tests. ω: Real-time vortex field intensity; ω0: Reference vortex intensity threshold.

[0016] Furthermore, when the vortex intensity exceeds the reference threshold, the linear amplification of the proportional gain enhances the control system's response sensitivity to sudden disturbances. The relaxation time of the exponential smoothing filter is dynamically adjusted according to the vortex field intensity, reducing the filter intensity in high-disturbance scenarios and avoiding control lag caused by over-smoothing.

[0017] Preferably, the closed-loop update model parameters include: The vortex field viscosity coefficient is corrected in reverse based on the deviation between the actual acceleration and the control target; The mapping relationship of equivalent temperature parameters is dynamically adjusted based on energy consumption error.

[0018] Furthermore, the flow field characteristic parameters are corrected using a gradient optimization method, which propagates the deviation between the actual acceleration and the target value back to the model parameters. The dynamic adjustment of the equivalent temperature parameter mapping relationship uses energy consumption error as the optimization objective, enabling the model to continuously adapt to changes in macroscopic density and microscopic velocity fluctuations in traffic flow.

[0019] Preferably, the multimodal sensor data includes: The three-dimensional point cloud sequence generated by lidar, the surface temperature field distribution captured by infrared camera, and the relative velocity vector measured by millimeter-wave radar; The data undergoes spatiotemporal alignment and noise suppression preprocessing.

[0020] Furthermore, spatiotemporal alignment employs a three-dimensional coordinate system transformation model to unify the data from various sensors to the vehicle's reference coordinate system. Noise suppression is achieved through an adaptive filtering algorithm, which corrects motion distortion in the lidar point cloud and applies non-uniform noise reduction processing based on flow field intensity to the infrared thermal imager data.

[0021] An unmanned vehicle speed control device includes: The sensor module is used to collect multimodal environmental data and execute dynamically optimized acquisition strategies; The processing module is used to construct a joint control model of the coupled vortex field and Brownian motion parameters and generate optimization instructions; The control module is used to adjust the execution speed of the rules based on dynamic parameters and to achieve closed-loop updates of model parameters.

[0022] Furthermore, the processing module integrates a parallel computing unit for rapid solution of fluid dynamics equations. The control module embeds programmable logic devices to achieve microsecond-level adjustment of control parameters, ensuring real-time response in highly dynamic scenarios.

[0023] An autonomous vehicle includes: Speed ​​control device; Drive actuators are used to receive control commands and adjust the vehicle's drive torque; The on-board computing unit is configured to run the processing module and control module of the speed control device; When using the aforementioned unmanned vehicle, the speed control method for the unmanned vehicle is implemented.

[0024] Furthermore, the drive actuator employs a torque vectoring strategy, decomposing speed commands into independent control signals for each wheel. The onboard computing unit connects sensors and controllers via a high-speed bus, ensuring that data throughput and computational latency meet real-time requirements, achieving millisecond-level closed-loop control of perception-decision-execution.

[0025] In summary, the present invention has at least one of the following beneficial technical effects: 1. This invention achieves unified modeling of traffic flow micro-disturbances and environmental uncertainties by integrating fluid dynamics vortex field theory with Brownian motion stochastic mapping. Compared with the traditional method of using fixed-parameter traffic flow models and noise separation processing in existing technologies, it effectively solves the problem that deterministic models are difficult to adapt to the dynamic changes in complex traffic scenarios, enabling unmanned vehicle speed control to have physically interpretable response capabilities to sudden disturbances.

[0026] 2. This invention employs a dynamic sensing strategy driven by vortex field intensity and a predictive time-domain compression technique, achieving an intelligent balance between computational resource allocation and sensing accuracy. Compared to traditional full-range continuous scanning sensor control methods, it overcomes the bottleneck problems of data redundancy and computational latency in high-density traffic scenarios, significantly improving the real-time performance of the system in complex scenarios.

[0027] 3. This invention employs a variational optimization solution based on multi-physics coupling, constructing a joint optimization framework for vehicle kinematics and traffic flow potential energy field. Compared to existing single-objective speed tracking control methods, it breaks through the simplified handling mode of conflict relationships between control objectives, enabling speed command generation to possess multi-dimensional optimization features that combine energy efficiency, smoothness, and safety.

[0028] 4. This invention establishes a technical solution for a full-link closed-loop update mechanism, forming an adaptive control system through real-time data-driven reverse correction of model parameters and dynamic gain adjustment. Compared with traditional open-loop parameter-fixed control strategies, this completely solves the model mismatch problem caused by time-varying environmental characteristics, ensuring the control robustness and scenario generalization ability of autonomous vehicles during long-term operation. Attached Figure Description

[0029] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0030] The following is in conjunction with the appendix Figure 1 The present invention will be further described in detail below.

[0031] This invention provides a speed control method for an unmanned vehicle, comprising: S1. Acquire multimodal sensor data of the vehicle's surrounding environment; S2. Construct a dynamic vortex field model and Brownian motion parameter mapping relationship based on the sensor data; S3. Dynamically adjust the sensor acquisition strategy and compress the prediction time domain based on the vortex field intensity; S4. Construct the joint control equations for the coupled vortex field and Brownian motion parameters; S5. Generate speed control commands by solving variational optimization problems; S6. Execute control commands based on dynamic parameter adjustment rules and update model parameters in a closed loop.

[0032] Step S1: Acquire multimodal sensor data of the vehicle's surrounding environment In this embodiment, the acquisition of multimodal sensor data is achieved through heterogeneous sensor collaborative acquisition and spatiotemporal alignment fusion processing. Specifically, the lidar scans the environment around the vehicle by emitting a laser beam, generating a three-dimensional point cloud sequence containing spatial coordinate information. The scanning mode of the laser beam is dynamically adjusted according to the complexity of the environment, and the angular resolution is automatically improved in the high-interest area in front of the vehicle. The infrared thermal imaging camera simultaneously captures the surface temperature field distribution data of the target object, and identifies the heat source characteristics of dynamic obstacles through the difference in thermal radiation intensity. The millimeter-wave radar measures the relative velocity vector of surrounding targets based on the Doppler effect, and distinguishes between stationary and moving objects through multi-channel echo signal separation technology.

[0033] The data synchronization of the sensors is achieved through a combination of hardware trigger signals and software timestamp alignment, ensuring that data from each modality is acquired under the same time reference. For lidar point cloud data, an adaptive filtering algorithm based on motion compensation is used to eliminate point cloud distortion caused by the vehicle's own motion. Specifically, the point cloud coordinates are reversed using real-time pose data provided by the inertial measurement unit. Infrared thermal imaging data undergoes non-uniform noise reduction processing, preserving the original resolution in areas with significant high-temperature gradients and performing spatial mean filtering in areas with uniform temperature distribution to suppress noise. Millimeter-wave radar data uses a probability hypothesis density filter to separate effective target echoes from clutter interference, improving the detection stability of low-speed targets.

[0034] The spatiotemporal alignment of multimodal data is achieved using the Lie group SE(3) transformation model, which transforms the observation data in each sensor coordinate system to the body coordinate system with the vehicle's center of mass as the origin. Specifically, the extrinsic parameter calibration matrix of each sensor is established, and the extrinsic parameters are dynamically updated through a combination of offline calibration and online optimization to eliminate coordinate offset errors caused by mechanical vibration or temperature drift. For lidar point cloud data, the conversion formula is: Pvehicle =M lidar ·P lidar ; Among them, P lidar M represents the original lidar point cloud coordinates. lidar P is the extrinsic transformation matrix of the lidar. vehicle This is the point cloud in the transformed vehicle coordinate system. The coordinate system transformation of the velocity vector in millimeter-wave radar uses a similar principle, aligning the velocity direction with the vehicle's motion direction through matrix operations.

[0035] In the data fusion stage, a dynamic weighting function is constructed to adjust the fusion strategy based on sensor type, ambient lighting conditions, and target motion state. In strong backlight scenarios, the weighting coefficients of infrared thermal imaging data are automatically increased to compensate for the attenuation of lidar detection performance; in rainy or foggy weather, the confidence weight of millimeter-wave radar is increased to overcome signal attenuation from optical sensors. The fusion output is a spatiotemporally aligned multiphysics data cube, containing joint distribution information of spatial coordinates, temperature gradients, and velocity vectors, providing a highly consistent input data foundation for subsequent vortex field modeling.

[0036] The proposed method significantly improves the robustness of environmental perception in complex scenarios by leveraging the complementary advantages and synergistic optimization of heterogeneous sensors, overcomes the detection blind spots and noise interference problems of single sensors, and ensures the completeness and timeliness of the data required for constructing dynamic vortex field models.

[0037] Step S2: Construct a dynamic vortex field model and Brownian motion parameter mapping relationship based on the sensor data. In this embodiment, the dynamic vortex field model is constructed based on the modified Navier-Stokes equations. By introducing the coupling relationship between vehicle dynamics parameters and hydrodynamic properties, a quantitative characterization system for traffic flow micro-disturbances is established. Specifically, the velocity field distribution in lidar point cloud data is fused with millimeter-wave radar speed measurement results to calculate the local vortex intensity, and the effect of vehicle inertia on traffic flow is reflected by the dynamic adjustment of the viscosity coefficient.

[0038] Navier-Stokes equation calculation formula: in: The instantaneous acceleration of the traffic flow velocity field reflects the dynamic changes in the environment surrounding the vehicle; Local traffic flow velocity vector; t: time; v: Kinematic viscosity coefficient, calculated from fluid density and dynamic viscosity; k: Dimensionless proportionality coefficient; m: Vehicle weight; I: Moment of inertia of the vehicle about its center of mass.

[0039] Brownian motion parameter mapping is achieved through a thermodynamic analogy, relating traffic density ρ to average velocity. The product relationship is converted into equivalent temperature parameters: Wherein, kT characterizes the intensity of random disturbances in traffic flow, and k is the Boltzmann constant analogy, which is obtained through calibration in actual traffic scenarios; The average speed of local traffic flow is calculated by the moving average of velocity vectors measured by millimeter-wave radar, which transforms the macroscopic density fluctuations and microscopic speed randomness of traffic flow into a physical quantity that can be analyzed by vehicle dynamics.

[0040] In viscosity coefficient calculations, the introduction of vehicle mass *m* and moment of inertia *I* allows the model to dynamically reflect the damping effect of vehicle inertia on the surrounding flow field. For example, when the vehicle mass is large or the moment of inertia is high (e.g., under cargo load), the square root relationship of dynamic viscosity *μ* automatically increases the weight of the flow field drag term, suppressing abrupt changes in vortex intensity. Random disturbance term. The generation is simulated using the Wiener process, whose variance is proportional to the equivalent temperature parameter kT, thus mapping the random fluctuations of traffic flow into a quantifiable noise source in the control equation.

[0041] Preferably, the dynamic viscosity is calculated using an online learning mechanism, which continuously optimizes the model parameters based on the deviation between the actual acceleration and the control target to adapt to different road scenarios. Furthermore, the mapping relationship of the equivalent temperature parameters is verified through energy conservation constraints to ensure the consistency between the statistical characteristics of random traffic flow disturbances and the vehicle kinematic response.

[0042] Using the above method, this step achieves bidirectional coupled modeling of traffic flow dynamics and vehicle kinematics, providing a joint input with both deterministic trends and random disturbances for subsequent control strategies, and laying the physical foundation for vehicle autonomous decision-making.

[0043] Step S3: Dynamically Adjust the Sensor Acquisition Strategy and Compress the Prediction Time Domain Based on the Vortex Field Intensity. In this embodiment, the dynamic adjustment of the sensor acquisition strategy is based on the real-time calculated vortex field intensity distribution. Through a collaborative optimization mechanism of spatial domain focusing and temporal domain compression, the consumption of computing resources is reduced while ensuring the acquisition of key information. Specifically, when the vortex field intensity exceeds a preset threshold, the infrared thermal imaging channel is activated to preferentially scan the high-temperature gradient region. At the same time, the scanning range of the lidar is dynamically reduced, and the time domain window of the model prediction control is compressed according to a nonlinear relationship, forming a closed-loop optimization of perception and computation.

[0044] The threshold value for vortex field intensity is set adaptively based on the vehicle's motion state and the complexity of the scene, and its judgment condition is as follows: Where ω is the current scalar value of the vortex field intensity, which is obtained in real time from the model output in step S2; ω0 is the baseline threshold, which is initialized based on the road type (such as urban roads or highways); Let be the magnitude of the vehicle's velocity vector; v ref The reference speed is used. The physical meaning of this formula is that when the vehicle speed is high or the vortex intensity is significantly enhanced, the system automatically reduces the sensitivity of the trigger threshold and activates the sensor optimization strategy in advance.

[0045] Where ω is the current scalar value of vortex field intensity, which is obtained in real time from the model output in step S2; ω0 is the baseline threshold, which is initialized according to the road type (such as urban road, highway); v is the magnitude of the vehicle's velocity vector; ref The reference speed is set to 80% of the current road speed limit. The physical meaning of this formula is that when the vehicle speed is high or the vortex intensity is significantly enhanced, the system automatically reduces the sensitivity of the trigger threshold and activates the sensor optimization strategy in advance.

[0046] The activation of the infrared thermal imaging channel employs a spatial gradient filtering mechanism, prioritizing the scanning of temperature change rates. area ( (Using a gradient threshold), focusing on dynamic obstacles with significant thermal characteristics, such as engine heat sources and brake pads.

[0047] Preferably, the LiDAR scanning range is dynamically compressed according to the spatial distribution of the vortex field intensity: in the region where ω>ω0, the scanning angle is reduced from the default 120° to 60°, and a fan-shaped scanning window is generated with this region as the center, so that the point cloud data acquisition density in the key area is increased to twice the original level, while reducing redundant data acquisition in non-interested areas.

[0048] The compression of the prediction time domain is achieved through a time-varying window function, the time length T of which is inversely proportional to the vortex field intensity: Where T0 is the initial prediction time domain length and α is the compression coefficient, which is determined by vehicle dynamics stability constraints. This design enables the prediction time domain to be automatically shortened in high vortex intensity scenarios (such as traffic conflicts at intersections) to reduce the impact of environmental uncertainties on control commands, while increasing the control frequency to improve the system response speed.

[0049] Preferably, the coordinated execution of the sensor strategy adjustment and the prediction time-domain compression adopts an event-driven mechanism: when the vortex field intensity triggers the threshold condition, infrared thermal imaging commands, lidar scanning parameters, and prediction time-domain adjustment signals are generated simultaneously, and a microsecond-level response is achieved through a hardware interrupt channel. At the data processing level, the compressed prediction time-domain window uses a moving average filter to weight and fuse historical data, avoiding control command oscillations caused by abrupt changes in the time domain.

[0050] Using the above methods, this step achieves intelligent allocation of sensing resources and dynamic optimization of computational load in complex traffic scenarios. This not only ensures high-precision data acquisition in areas with strong disturbances, but also reduces the solution complexity of model predictive control through time-domain compression, providing stable input conditions for real-time control.

[0051] Step S4: Construct the joint governing equations of the coupled vortex field and Brownian motion parameters. In this embodiment, the joint control equations are constructed by establishing a partial differential coupling relationship between vortex field intensity and Brownian motion parameters. This involves fusing vehicle dynamics characteristics, random traffic flow disturbances, and sensor data through a multi-physics approach, forming a control model that combines deterministic trends with descriptions of random noise. Specifically, based on the dynamic vortex field intensity distribution and equivalent temperature parameters generated in step S2, a partial differential equation is established that incorporates the interaction between vehicle yaw motion and flow field evolution. A multi-objective optimization function is then used to coordinate the smoothness of control commands with energy efficiency.

[0052] The partial differential coupling relationship is expressed as follows: in: The rate of change of vortex field intensity per unit time; ω: Scalar value of vortex field intensity; Real-time yaw rate of this vehicle; kT: composed of traffic density ρ and average speed The generated equivalent temperature parameters; m: Vehicle weight; Local traffic flow velocity vector; x: Vehicle longitudinal coordinate; Instantaneous acceleration of the traffic flow velocity field.

[0053] The construction of the multi-objective optimization function includes an acceleration smoothing term and an energy consumption gradient term, whose weighting coefficients are adaptively adjusted according to the vortex field intensity. The specific expression is: Wherein, λ1 is the weight of the acceleration smoothing term, which is inversely proportional to the vortex intensity ω, allowing for more drastic acceleration changes in strong vortex disturbance scenarios to avoid risks. ∫ is the time integration operator; λ2 is the weight of the energy consumption gradient term, which is proportional to ω, prompting the optimization of energy efficiency in high-perturbation environments. This is the acceleration vector of the vehicle; Let be the derivative of the acceleration vector with respect to time, representing the rate of change of acceleration; dt is the discrete time step of the control system; The energy consumption gradient field is generated by jointly calculating the motor torque-speed characteristic curve and the battery discharge model.

[0054] Preferably, the discretization solution of the partial differential equations employs the finite volume method, dividing the space surrounding the vehicle into a dynamically updated non-uniform grid. In regions with significant vortex intensity gradients (such as the vehicle's side blind spot), the grid nodes are automatically refined to improve computational accuracy. Real-time calculation of the coupling coefficients is achieved through a hardware acceleration unit, ensuring that the equation iterative solution is completed within a millisecond-level time window.

[0055] This step incorporates vehicle kinematics, traffic flow hydrodynamics, and energy consumption characteristics into a unified control framework, breaking through the limitations of isolated treatment of physical quantities in traditional control models and providing a high-fidelity mathematical description foundation for variational optimization solutions.

[0056] Step S5: Generate speed control commands through variational optimization. In this embodiment, the variational optimization solution is based on the Hamiltonian action principle, transforming the joint control equations constructed in step S4 into an energy-optimal control problem. Speed ​​control commands satisfying multi-objective constraints are generated by dynamically adjusting optimization weights and using a discretization solution strategy. Specifically, by defining an action functional that includes vehicle kinetic energy, vortex field potential energy, and energy consumption gradient, and combining real-time updated sensor data and model parameters, the optimal speed trajectory satisfying boundary conditions and dynamic constraints is solved.

[0057] Embedding stochastic differential equations into the Hamiltonian action principle: in: δ: Variational operator, representing a small change in the functional of the action, used to solve for the extreme path of the vehicle's trajectory; ∫: Time integration operator; m: Vehicle weight; The vehicle's velocity vector; Integral potential energy of a vortex field; Spatial location The intensity of the vortex field at that location; The volume of a tiny spatial region in the environment surrounding a vehicle; dt: Discrete time step of the control system; Three-dimensional spatial vector; t: time.

[0058] The construction of the Hamiltonian action functional deeply integrates vehicle motion state with traffic flow environment characteristics. Its expression encompasses a vehicle kinetic energy term and a vortex field potential energy integral term. The kinetic energy term, composed of the square product of the vehicle's mass and velocity vector magnitude, reflects the energy consumption of the vehicle's own motion state. The potential energy integral term quantifies the potential impact of traffic flow disturbances on vehicle motion by calculating the volume integral of the vortex field intensity in the three-dimensional space surrounding the vehicle. The optimization objective is to find a velocity control sequence that minimizes the action functional while satisfying constraints on acceleration smoothness, energy efficiency, and safe distance.

[0059] An adaptive discretization strategy is employed during the solution process, dividing the continuous time domain into dynamically adjusted time step intervals. When a sudden change in vortex field intensity or a significant change in equivalent temperature parameters is detected, the time step is automatically reduced to improve the optimization accuracy of local regions; during the stable flow phase, the step size is increased to improve computational efficiency. The optimization variables at each discrete time node include longitudinal velocity, lateral velocity, and yaw rate, and constraints such as acceleration constraints, road boundary conditions, and following distance are introduced through the Lagrange product method.

[0060] The proportional gain coefficient is adjusted in real time based on the vortex field intensity. in: K p : Current proportional gain coefficient; K p0 The reference proportional gain coefficient is determined through vehicle calibration tests. ω: Real-time vortex field intensity; ω0: Reference vortex intensity threshold.

[0061] Preferably, the variational optimization problem is solved iteratively using the conjugate gradient method, and the initial value of the current iteration is initialized using the optimization result of the previous time window, reducing computational complexity. For the weight allocation of the acceleration smoothing term and the energy consumption gradient term in the multi-objective optimization function, a nonlinear mapping relationship with the vortex field intensity is established: when the vortex field intensity exceeds a threshold, the weight coefficient of the energy consumption gradient term increases exponentially, prompting the control system to prioritize energy efficiency under strong disturbance scenarios; while in the low vortex intensity region, the weight of the acceleration smoothing term dominates, ensuring smooth driving.

[0062] Subsequent processing of the optimization results includes exponential smoothing filtering with adaptive relaxation time, dynamically adjusting the filtering time constant based on the vortex field intensity. In high-frequency disturbance regions, a smaller time constant is used to preserve the fast response characteristics of the control command; in stable flow regions, the time constant is increased to suppress command jitter caused by sensor noise. The final generated speed control command includes speed amplitude, direction angle, and acceleration limit information, and is transmitted to the vehicle drive actuators via the CAN bus.

[0063] Using the above method, this step enables real-time solution of the multi-physics coupled control model in complex traffic environments, generating speed control commands that take into account safety, comfort, and energy efficiency, and providing a precise motion planning basis for vehicle autonomous decision-making.

[0064] Step S6: Execute control commands and update model parameters in a closed loop based on dynamic parameter adjustment rules. In this embodiment, the execution of control commands and the closed-loop update of model parameters are achieved through dynamic gain adjustment and data-driven reverse optimization mechanisms to ensure the real-time performance and environmental adaptability of the control system. Specifically, the speed control command generated in step S5 is input to the vehicle drive actuator, and the control parameters are dynamically adjusted according to the deviation between the actual acceleration and the target value. Furthermore, the key coefficients of the vortex field model and the Brownian motion mapping relationship are corrected in reverse using an online learning algorithm.

[0065] The core of the dynamic parameter adjustment rule lies in the vortex intensity correlation design of the proportional gain. When the vortex field intensity exceeds the reference threshold, the proportional gain coefficient linearly amplifies with the intensity value to enhance the system's response to sudden disturbances; while when the vortex field intensity is in the low fluctuation range, the gain coefficient automatically decays to avoid control oscillations caused by over-response. An inertial hysteresis element is introduced into the gain adjustment process to prevent frequent parameter jumps caused by sensor noise or instantaneous fluctuations in the vortex field, thereby improving control stability.

[0066] During the control command execution phase, an exponential smoothing filter with adaptive relaxation time is applied, and the filter time constant is dynamically adjusted according to the spatial gradient of the vortex field intensity. In high-intensity gradient regions (such as disturbances from nearby obstacles in front of the vehicle), a smaller time constant is used to preserve the high-frequency response characteristics of the command; in low-gradient regions, the time constant is increased to suppress command jitter caused by sensor noise and model errors. The filtered control command is decomposed to each drive motor through a torque distribution strategy to achieve coordinated control of longitudinal speed tracking and lateral attitude stabilization.

[0067] The closed-loop update of model parameters adopts a dual-channel optimization mechanism: for the dynamic vortex field model, based on the deviation between the actual acceleration and the target acceleration, the gradient direction of the viscosity coefficient and vehicle dynamic parameters is solved by the adjoint equation, and the proportional coefficient is iteratively updated by the momentum gradient descent method; for the Brownian motion mapping relationship, by monitoring the cumulative error between the actual energy consumption and the predicted value, the density-velocity coupling coefficient in the equivalent temperature parameter is adjusted in reverse, so that the statistical characteristics of the random disturbance of traffic flow tend to be consistent with the actual motion response of the vehicle.

[0068] Preferably, the closed-loop update process incorporates a sliding time window mechanism, retaining only sensor data and vehicle state information within the most recent time window for parameter optimization, thus avoiding model overfitting caused by the accumulation of historical data. The updated model parameters are written to the controller's non-volatile memory in real time via a hardware acceleration unit, ensuring that the latest calibration results are retained even after power failure. For key parameters (such as the dynamic viscosity proportionality coefficient), physical constraint boundaries are set to prevent non-physical numerical drift during the optimization process.

[0069] Through the above methods, this step achieves high-precision execution of control commands and continuous self-correction of model parameters, forming a fully closed-loop control link of "perception-decision-execution-verification", effectively overcoming the defects of model mismatch and parameter drift in traditional open-loop control systems.

[0070] The speed control device for an unmanned vehicle described below and the speed control method for an unmanned vehicle described above can be referred to in correspondence.

[0071] The present invention also provides a speed control device for an unmanned vehicle, comprising: The sensor module is used to collect multimodal environmental data and execute dynamically optimized acquisition strategies; The processing module is used to construct a joint control model of the coupled vortex field and Brownian motion parameters and generate optimization instructions; The control module is used to adjust the speed control according to the dynamic parameters and realize the closed-loop update of the model parameters. The device in this embodiment can be used to execute the above method embodiment, and its principle and technical effect are similar, so it will not be described again here.

[0072] The present invention also provides an unmanned vehicle, comprising: Speed ​​control device; Drive actuators are used to receive control commands and adjust the vehicle's drive torque; The on-board computing unit is configured to run the processing module and control module of the speed control device; When using the aforementioned unmanned vehicle, the speed control method for the unmanned vehicle is implemented.

[0073] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A speed control method for an unmanned vehicle, characterized in that, include: Acquire multimodal sensor data of the vehicle's surrounding environment; Based on the sensor data, a dynamic vortex field model and Brownian motion parameter mapping relationship are constructed. The sensor acquisition strategy is dynamically adjusted based on the intensity of the vortex field, and the prediction time domain is compressed. Construct the joint control equations for the coupled vortex field and Brownian motion parameters; Speed ​​control commands are generated by solving a variational optimization problem. Control commands are executed based on dynamic parameter adjustment rules, and model parameters are updated in a closed loop.

2. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The construction of the dynamic vortex field model includes: The local vorticity distribution is calculated based on the modified Navier-Stokes equations, where the viscosity coefficient is related to the vehicle mass and moment of inertia. Traffic density is mapped to an equivalent temperature parameter to construct vehicle dynamics equations that include random disturbance terms; Navier-Stokes equation calculation formula: ； in: The instantaneous acceleration of the traffic flow velocity field reflects the dynamic changes in the environment surrounding the vehicle. Local traffic flow velocity vector; :time; Kinematic viscosity coefficient, calculated from fluid density and dynamic viscosity; Dimensionless proportionality coefficient; The quality of this vehicle; : Moment of inertia of the vehicle about its center of mass.

3. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The dynamic adjustment strategy for sensor acquisition includes: When the vortex intensity exceeds a preset threshold, the infrared thermal imaging channel is activated and the high-temperature gradient region is focused. The lidar scanning area is dynamically reduced based on the temperature gradient distribution, and the time domain range of the model prediction control is simultaneously compressed.

4. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The construction of the joint control equations includes: A partial differential coupling relationship between vortex field intensity and Brownian motion parameters is established, wherein the coupling coefficient is dynamically related to the vehicle yaw rate. Define a multi-objective optimization function that includes acceleration smoothing term and energy consumption gradient term, with its weight coefficients adaptively adjusted according to the intensity of vortex field; Partial differential coupling formula: ； in: : The rate of change of vortex field intensity per unit time; : Scalar value of vortex field intensity; Real-time yaw rate of this vehicle; Traffic density and average speed The generated equivalent temperature parameters; The quality of this vehicle; Local traffic flow velocity vector; : Vehicle longitudinal coordinate; : Instantaneous acceleration of the traffic flow velocity field.

5. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The variational optimization solution includes: Embedding stochastic differential equations into the Hamiltonian action principle: ； in: Variational operators represent small changes in the functional of the action quantity and are used to solve for the extreme paths of vehicle trajectories. Time integration operator; The quality of this vehicle; : This vehicle's velocity vector; : Integral potential energy of vortex field; Spatial location The intensity of the vortex field at that location; The volume of a tiny spatial region in the environment surrounding a vehicle; Discrete time step of the control system; : Three-dimensional spatial vector; :time; An adaptive step-size numerical discretization method is used to solve the optimization problem, with the step size dynamically switched according to the modified Reynolds number.

6. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The execution control instructions include: The gain coefficient of the proportional-integral-derivative controller is adjusted in real time based on the vortex field intensity and equivalent temperature parameters. The original control command is subjected to an exponential smoothing filter based on the relaxation time. The proportional gain coefficient is adjusted in real time based on the vortex field intensity. ； in: : Current proportional gain coefficient; The reference proportional gain coefficient is determined through vehicle calibration tests. Real-time vortex field intensity; : Reference vortex intensity threshold.

7. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The closed-loop update model parameters include: The vortex field viscosity coefficient is corrected in reverse based on the deviation between the actual acceleration and the control target; The mapping relationship of equivalent temperature parameters is dynamically adjusted based on energy consumption error.

8. The speed control method for an unmanned vehicle according to claim 1, characterized in that, The multimodal sensor data includes: The three-dimensional point cloud sequence generated by lidar, the surface temperature field distribution captured by infrared camera, and the relative velocity vector measured by millimeter-wave radar; The data undergoes spatiotemporal alignment and noise suppression preprocessing.

9. A speed control device for an unmanned vehicle, applied to the speed control method for an unmanned vehicle as described in claims 1-8, characterized in that, include: The sensor module is used to collect multimodal environmental data and execute dynamically optimized acquisition strategies; The processing module is used to construct a joint control model of the coupled vortex field and Brownian motion parameters and generate optimization instructions; The control module is used to adjust the execution speed of the rules based on dynamic parameters and to achieve closed-loop updates of model parameters.

10. An unmanned vehicle, characterized in that, include: Speed ​​control device; Drive actuators are used to receive control commands and adjust the vehicle's drive torque; The on-board computing unit is configured to run the processing module and control module of the speed control device; When using the aforementioned unmanned vehicle, the speed control method for the unmanned vehicle as described in any one of claims 1-8 is implemented.

Images