Method and system for right-of-way preemption of autonomous sightseeing vehicle

Abstract

This invention belongs to the field of autonomous driving and intelligent vehicle control technology. Specifically, it discloses an autonomous sightseeing vehicle right-of-way acquisition method and system, including: acquiring physical distance data between the vehicle and the preceding vehicle, and perception data of a target crossing the road; calculating the estimated intersection time of the target reaching the lane boundary; combining the battery's maximum discharge power limit and the motor's torque response delay to deduce the maximum physical displacement of the vehicle within the intersection time; calculating the estimated minimum distance, comparing it with the sum of the target's geometric envelope width and safety margin to generate a right-of-way decision index; controlling the motor to output positive torque to reduce the distance to the vehicle or controlling the braking system to output negative torque to yield; and finally controlling a directional acoustic array to emit a directional sound beam matching the torque state towards the target. This invention achieves a closed-loop interaction of vehicle movement and acoustic linkage for right-of-way, balancing traffic efficiency and pedestrian safety.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving and vehicle intelligent control technology, and relates to a method and system for autonomous sightseeing vehicles to seize road rights. Background Technology

[0002] Autonomous sightseeing vehicles operating in enclosed or semi-enclosed environments such as scenic spots and parks face complex traffic situations where pedestrians or non-motorized vehicles may cross the road at any time. To balance traffic efficiency and the safety of vulnerable road users, vehicles must have the ability to actively determine right-of-way and communicate their decision-making intentions. This complete right-of-way management technology, which consists of perception, decision-making, vehicle motion control, and interpersonal acoustic interaction, is one of the key supports for autonomous driving on low-speed unstructured roads.

[0003] Current technologies for handling pedestrian crossing scenarios generally employ passive avoidance strategies based on collision risk ratings. These strategies either brake unconditionally to a stop upon detecting a pedestrian approaching the lane, or use a fixed-frequency horn to issue a warning. Some solutions introduce vehicle-to-infrastructure (V2I) architectures, broadcasting pedestrian information to vehicles via roadside sensing units and communication equipment, allowing the vehicles to then execute emergency stops or simple deceleration.

[0004] This existing solution ignores the actual need for autonomous sightseeing vehicles to actively reduce distance to block the passage during convoy driving. Its decision-making does not take into account the instantaneous power limit limited by battery charge and temperature state, as well as the delay time for the drive motor torque to build up. This means that the intention to seize right-of-way may not be achieved through the actual movement of the vehicle. At the same time, the non-directional horn sound field does not have the spatial resolution capability for specific individuals crossing the road, and cannot switch beam patterns according to different decisions of seizing or yielding right-of-way. This can easily cause misunderstanding or even panic among pedestrians, and may lead to misjudgment of pedestrians' intentions, reducing the efficiency and safety of human-vehicle interaction. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above objectives, the present invention proposes the following technical solution: a method for right-of-way seizure by an autonomous sightseeing vehicle, comprising: S1, obtaining the physical distance between the vehicle and the vehicle in front, and obtaining the coordinate data and velocity vector of the target traversing the vehicle relative to the vehicle, as environmental perception data.

[0006] S2. Using environmental perception data and the geometric boundary line of the lane where the vehicle is located, calculate the estimated intersection time of the plane formed by the geometric boundary line of the target lane and the gap between the convoy.

[0007] S3. Read the current instantaneous maximum discharge power limit of the vehicle's battery management system and the torque response delay time of the drive motor. Convert the instantaneous maximum discharge power limit into the maximum allowable acceleration. Combine the torque response delay time with the expected intersection time to deduce the maximum physical displacement that the vehicle can move forward within the expected intersection time.

[0008] S4. Calculate the predicted displacement of the preceding vehicle within the expected intersection time by combining the longitudinal relative speed of the preceding vehicle. Add the predicted displacement of the preceding vehicle to the physical distance and then subtract the maximum physical displacement to obtain the estimated minimum distance. Compare the estimated minimum distance with the sum of the geometric envelope width of the target crossing and the preset longitudinal safety margin to generate the right-of-way decision index.

[0009] S5. Based on the right-of-way decision index, control the positive torque output of the vehicle's motor to match the instantaneous maximum discharge power limit in order to reduce the convoy spacing, or control the braking system to output negative torque in order to actively give way.

[0010] S6. Read the output torque status and control the directional acoustic array to emit a directional sound beam corresponding to the torque status to the position indicated by the coordinate data of the target crossing, so as to complete the information exchange of right-of-way and yielding.

[0011] The second aspect of the present invention provides an autonomous driving sightseeing vehicle right-of-way preemption system, comprising: an environmental perception module, used to obtain the physical distance between the vehicle and the vehicle in front, and to obtain the coordinate data and velocity vector of the target traversing the vehicle relative to the vehicle, as environmental perception data.

[0012] The intersection prediction module is used to calculate the estimated intersection time of the plane formed by the geometric boundary line of the target's approach lane and the gap between the convoy, using environmental perception data and the geometric boundary line of the lane where the vehicle is located.

[0013] The displacement estimation module is used to read the current instantaneous maximum discharge power limit of the vehicle's battery management system and the torque response delay time of the drive motor. It converts the instantaneous maximum discharge power limit into the maximum allowable acceleration, and combines the torque response delay time with the expected intersection time to estimate the maximum physical displacement that the vehicle can move forward within the expected intersection time.

[0014] The right-of-way decision module is used to calculate the predicted displacement of the preceding vehicle within the expected intersection time by combining the longitudinal relative speed of the preceding vehicle. The physical distance is added to the predicted displacement of the preceding vehicle and then the maximum physical displacement is subtracted to obtain the estimated minimum distance. The estimated minimum distance is compared with the sum of the geometric envelope width of the target and the preset longitudinal safety margin to generate the right-of-way decision index.

[0015] The torque control module is used to control the vehicle's motor output to match the instantaneous maximum discharge power limit with the right-of-way decision index to reduce the platoon spacing, or to control the braking system to output negative torque to actively give way.

[0016] The acoustic interaction module is used to read the output torque status and control the directional acoustic array to emit directional sound beams corresponding to the torque status to the position indicated by the coordinate data of the target, so as to complete the information exchange of right-of-way and yielding.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention can accurately determine whether the gap between the vehicles is sufficient for the target to pass through by calculating the expected intersection time of the target in real time and the maximum physical displacement of the vehicle constrained by the battery power limit and the motor torque response delay, thus avoiding decision failure caused by ignoring the dynamic constraints of the power system.

[0018] (2) The present invention uses the right-of-way decision index to control the motor to preempt acceleration or the braking system to give way and decelerate, and simultaneously drives the directional acoustic array to warn and drive away with a narrow beam or to make a friendly gesture with a wide beam, realizing the human-vehicle closed-loop interaction of vehicle movement and acoustic signal linkage, and improving the clarity and safety of intention expression in mixed traffic scenarios.

[0019] (3) The present invention can be mounted on an on-board computing platform without relying on roadside facilities, from the side-facing lidar to sense moving targets, to generating right-of-way decision index and executing torque control and directional acoustic beam emission, thereby enhancing the autonomous sightseeing vehicle's ability to independently navigate traffic in densely populated areas. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0022] Figure 2 This is a schematic diagram of the cross-target intersection prediction plane mapping of the present invention.

[0023] Figure 3 This is a schematic diagram illustrating the time history of the maximum physical displacement derivation of this invention.

[0024] Figure 4 This is a polar coordinate diagram of the directional acoustic array beam pattern of the present invention.

[0025] Figure 5 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Please see Figure 1 As shown, the autonomous driving sightseeing vehicle right-of-way acquisition method proposed in this invention includes: S1, obtaining the physical distance between the vehicle and the vehicle in front, and obtaining the coordinate data and velocity vector of the target traversing the vehicle relative to the vehicle, as environmental perception data.

[0028] In a preferred embodiment, acquiring the coordinate data and velocity vector of the traversing target relative to the vehicle includes detecting a moving target located to the side of the vehicle using the vehicle's side-facing lidar and acquiring point cloud data of the moving target; clustering and identifying the point cloud data to select pedestrian or non-motorized vehicle targets as traversing targets; extracting the orientation and distance of the geometric center point of the traversing target relative to the vehicle, generating coordinate data, and calculating the moving speed and direction of the traversing target based on the point cloud displacement of consecutive frames to form a velocity vector.

[0029] Specifically, environmental perception data consists of the physical distance between the vehicle and the vehicle in front, as well as the coordinates and velocity vector of the target traversing the vehicle relative to it. The physical distance is obtained through a forward-facing millimeter-wave radar mounted at the center of the front of the vehicle. This radar transmits a frequency-modulated continuous wave signal with a center frequency of 77 GHz and a bandwidth of B, with a modulation period of T. The electromagnetic wave is reflected from the rear of the vehicle in front and received by the radar antenna. The received signal is mixed with the local oscillator signal at the current transmission frequency, and after low-pass filtering, a difference frequency signal is output. The frequency of the difference frequency signal is denoted as... According to the principle of frequency-modulated continuous wave ranging, the radial distance from the radar antenna phase center to the reflection point of the preceding vehicle is... From the formula Given, where c represents the speed of electromagnetic wave propagation in air, taking a value of Meters per second. The longitudinal offset between the reference plane of the forward-facing millimeter-wave radar housing and the foremost plane of the vehicle's front bumper. This offset is obtained during the vehicle's factory calibration phase and remains constant; the physical distance... The calculation method is as follows Physical spacing represents the straight-line distance between the frontmost edge of this vehicle and the rearmost edge of the vehicle in front.

[0030] While measuring physical distances, the vehicle activates the side-mounted lidar sensors installed on both sides of the roof to detect the environment in the area to the side of the vehicle. The side-mounted lidars are multi-line mechanical rotating laser sensors with a vertical field of view extending from -15 degrees to +15 degrees, a horizontal field of view reaching 180 degrees, a maximum detection distance of 50 meters, an angular resolution of 0.2 degrees, and a data update frequency of 10 Hz. The side-mounted lidars establish a radar coordinate system with the optical center of the laser emitter as the origin. By emitting narrow-pulse laser beams into the surrounding space and receiving laser echoes reflected from object surfaces, the coordinates (x, y, z) and reflection intensity of each scan point relative to the radar coordinate system are calculated using the time-of-flight method. The set of three-dimensional spatial points output across multiple consecutive frames constitutes the point cloud data.

[0031] After acquiring point cloud data, clustering algorithms and target recognition models are used to separate cross-traversing targets. The clustering algorithm employs a region growing method based on Euclidean distance, calculating the spatial distance between any two points in the point cloud and identifying all points with a distance less than a threshold. Points are merged into the same cluster. The threshold was set to 0.3 meters. This threshold was based on statistical analysis of point clouds of people and non-motorized vehicles measured by over 200 sets of industrial sensors. The analysis showed that the spacing between points on the surface of similar objects is typically less than 0.3 meters, while the minimum gap between different objects is greater than this value. Clustering resulted in several independent clusters, each representing a candidate object. For each cluster, its geometric features were extracted, including the vertical height distribution of all points within the cluster, the difference H between the lowest and highest points, the length L and width W of the minimum bounding rectangle projected onto the horizontal plane, and the point cloud density. The target recognition process utilized pre-trained geometry-based discrimination rules, established through statistical learning from over 10,000 sets of real pedestrian and non-motorized vehicle point cloud samples. The specific discrimination rules are as follows: If the cluster height H is between 1.5 meters and 2.0 meters, and the height-to-width ratio H / W is greater than 1.5, and the point cloud density in the top region of the cluster is concentrated (i.e., the point cloud count in the upper third of the cluster's height is extracted and its proportion of the total number of points in the cluster is calculated), if this proportion is greater than a set density concentration threshold (e.g., 30%), then the cluster is determined to be a pedestrian target. If the cluster height H is less than 1.7 meters, the width W is greater than 0.6 meters, and the aspect ratio L / W of the projected bounding rectangle is greater than 2.1, and it exhibits a symmetrical wheel-shaped point cluster distribution (i.e., the absolute value of the difference in point cloud count on both sides of the cluster along its longitudinal axis of symmetry is less than a symmetry threshold (e.g., 10%), and there are two obvious point cloud density peaks in the lower part of the cluster), then the cluster is determined to be a non-motorized vehicle target. Pedestrians and non-motorized vehicles selected through the above discrimination are uniformly marked as crossing targets.

[0032] Extract the geometric center point that traverses the target. This geometric center point is defined as the arithmetic mean of the three-dimensional coordinates of all N points within the cluster. Let the radar coordinates of the i-th point in the cluster be... The coordinates of the geometric center point C are:

[0033] The geometric center point is transformed from the radar coordinate system to the vehicle's coordinate system to obtain the coordinate data of the target relative to the vehicle. The origin of the vehicle's coordinate system is defined as the projection point of the vehicle's rear axle center onto the ground. The X-axis points in the vehicle's forward direction, the Y-axis points horizontally to the left side of the vehicle, and the Z-axis points vertically upward. The lateral lidar is rigidly connected to the vehicle body, and its translation vector relative to the origin of the vehicle's coordinate system is... The 3×3 rotation matrix R representing the radar attitude is predetermined through multi-sensor joint calibration. This rotation matrix is ​​determined by the radar's roll, pitch, and yaw attitude angles relative to the vehicle coordinate system. The position vector of the geometric center point in the vehicle coordinate system... From the conversion formula Calculated. Take X-axis components and Y-axis components Ignoring the Z-axis height, coordinate data uses azimuth θ and distance. express:

[0034] The azimuth angle θ is zero degrees in the positive X-axis direction of the vehicle and increases counterclockwise. The arctan2 function is used to ensure that the azimuth angle covers the entire quadrant.

[0035] The velocity vector is calculated based on the positional change of the geometric center point in two consecutive frames of point cloud data. At time t in the current frame, the planar position of the geometric center point traversing the target in the vehicle coordinate system is... The position of the previous frame at time t-1 is The time interval Δt between two frames is determined by the fixed frame rate of the lateral lidar; a frame rate of 10 Hz corresponds to Δt equal to 0.1 seconds. To eliminate interference from the vehicle's own motion on the calculation of the target crossing speed, the longitudinal speed of the vehicle is obtained through the wheel speed sensors on the chassis, and the steering change is obtained in conjunction with the yaw rate sensor. The position of the previous frame is then used to calculate the vehicle's speed. Perform vehicle motion compensation, map it to the vehicle's coordinate system at time t in the current frame, and obtain the compensated position of the previous frame. Velocity vector X-axis component and components along the Y-axis Calculated separately using differences:

[0036] Movement speed The magnitude of the velocity vector:

[0037] The direction of movement is represented by the azimuth angle ψ of the velocity vector in the XY plane:

[0038] The absolute velocity vector traversing the target is composed of the movement speed and direction. At this point, the physical distance, the coordinates of the target traversal, and the velocity vector are combined to form environmental perception data for use in subsequent steps.

[0039] For example, in a mixed urban traffic scenario, the forward millimeter-wave radar is set with a sweep bandwidth B of 1 GHz and a modulation period T of 0.02 seconds. After receiving the echo and performing mixing processing, the frequency of the difference frequency signal is measured. The value is 15 kHz. Substituting the parameters into the radial distance formula: The radar longitudinal offset calibrated at the vehicle's factory settings. If it equals 0.3 meters, then the physical distance is... rice.

[0040] At the same time, the right-side lateral lidar outputs a frame of point cloud data, and the clustering threshold is set. The distance was 0.3 meters, resulting in three clusters. Cluster A contained 235 points. Based on geometric discrimination rules, this cluster had a height H of 1.7 meters, a minimum projected bounding rectangle length L of 1.7 meters, a width W of 0.8 meters, and an aspect ratio H / W of 2.125, exceeding the set threshold of 1.5. Furthermore, the upper third of the cluster contained 82 points (34.9%), exceeding the judgment threshold of 30%, thus meeting the pedestrian discrimination criteria and being identified as a crossing target. The geometric center point C of this pedestrian cluster has the following coordinates in the radar coordinate system: Translation vector of the right-side lateral lidar Set as Furthermore, no attitude rotation was performed during this calibration, and the rotation matrix R is the identity matrix. After transformation to the vehicle coordinate system, the position of the geometric center plane... rice, Meters. Distance corresponding to coordinate data: Azimuth: In the previous frame, the geometric center plane position of the same pedestrian, after being mapped to the current frame via vehicle motion compensation, has the following coordinates: With a frame interval Δt of 0.1 seconds, the absolute velocity vector components are: Movement speed: Direction of movement: This set of environmental perception data fully describes a pedestrian crossing the road at a distance of 5.449 meters, located at an orientation of 47.23 degrees, moving at an absolute speed of 1.803 meters per second along a direction of 56.31 degrees, 44.7 meters ahead of the vehicle in front, and on the right side of the pedestrian. This provides accurate input for the vehicle's subsequent decision-making.

[0041] S2. Using environmental perception data and the geometric boundary line of the lane where the vehicle is located, calculate the estimated intersection time of the plane formed by the geometric boundary line of the target lane and the gap between the convoy.

[0042] In a preferred embodiment, calculating the estimated intersection time of the traversing target reaching the plane formed by the lane geometric boundary line and the gap between the convoy includes: extracting the velocity and direction components of the velocity vector from environmental perception data; mapping the coordinate data and direction components of the traversing target to the lane projection plane in the direction of vehicle travel to obtain the lateral entry velocity of the traversing target; and using the lateral entry velocity and the lateral distance between the coordinate data of the traversing target and the lane geometric boundary line, calculating the time for the traversing target to enter the lane geometric boundary line through a uniform linear motion model, as the estimated intersection time.

[0043] Specifically, calculating the estimated intersection time of the target crossing the plane formed by the lane geometric boundary line and the gap between the convoy is based on the acquired environmental perception data, while also incorporating the geometric boundary line information of the vehicle's current lane. The lane geometric boundary line is the projection line of the physical edge of the vehicle's current lane onto the ground, and this information is retrieved from the onboard high-precision map. The onboard high-precision map pre-stores the absolute coordinate sequence of each lane edge line in the geodetic coordinate system. The vehicle's precise position and heading angle in the geodetic coordinate system are obtained through the vehicle's onboard real-time dynamic differential positioning system. The geodetic coordinates of the lane edge lines are then transformed into the vehicle's coordinate system, with the ground projection of the vehicle's rear axle center as the origin, the X-axis pointing in the vehicle's direction of travel, and the Y-axis pointing horizontally to the left. At the same time, the heading angle of the vehicle's direction of travel relative to the lane center line is obtained through the positioning system. If the heading angle exceeds a set threshold, the coordinate system is rotated to be parallel to the lane line to eliminate projection errors caused by lane changes or yaw. In a coordinate system aligned with lane lines, the geometric boundary line of the lane immediately adjacent to the vehicle's trajectory can be approximated as a straight line parallel to the vehicle's X-axis over a short distance. The near-side boundary line is represented by a Y-coordinate constant. express. The specific value is determined by the vehicle's lateral position within the lane and the lane width. The lane width is taken as 3.75 meters based on the standard lane width. If the vehicle is centered, since the Y-axis is positive to the left, the constant of the left boundary line is... meters, right boundary constant Meters. The plane formed by the gaps between vehicles refers to the lateral extension of the longitudinal gap between the vehicle and the vehicle in front, defined by the physical distance. Once a target crosses the lane geometric boundary line, it means entering the lateral projection area of ​​that gap. Therefore, the meeting time is determined by the instant the target laterally crosses the boundary line.

[0044] Environmental perception data includes coordinate data and velocity vectors for crossing the target, with the coordinate data expressed as distance. Expressed in terms of azimuth angle θ, the velocity vector is expressed in terms of speed. The direction ψ is expressed as the velocity vector. As a measure of the target's speed, the direction ψ is extracted as the angle between the target's absolute velocity vector and the positive X-axis direction of the vehicle. To obtain the target's precise planar position within the vehicle's coordinate system, the coordinate data is expanded using a polar-to-Cartesian coordinate conversion formula, and the X-axis coordinates traversing the target are... and Y-axis coordinates The formulas are given respectively:

[0045] Then the coordinate data traversing the target will be... Mapping is performed with the lane projection plane, which is oriented with the direction ψ towards the vehicle's direction of travel. The lane projection plane is defined as an abstract plane that compresses the three-dimensional space along the vehicle's X-axis, retaining only the lateral Y-axis dimension. This mapping aims to filter out the influence of longitudinal motion and purely analyze the lateral crossing behavior of the target. Lateral cut-in velocity of the target. Derived from the Y-axis component of the absolute velocity vector:

[0046] A positive value indicates velocity along the positive Y-axis, and a negative value indicates velocity along the negative Y-axis. This applies to a target currently outside the lane's geometric boundary (i.e.,...). or When determining the relative position of the target crossing the lane's geometric boundary line, firstly based on... The symbol is selected from the lane boundary line on the same side as it as the reference object. If the value is positive, then the left boundary line constant is taken. ,like If the value is negative, the right boundary line constant is taken. This reference constant value is uniformly denoted as The lateral distance between the target's coordinates and the lane's geometric boundary is a scalar, calculated by taking the target's current Y-coordinate and... The absolute value of the difference: Horizontal distance It represents the lateral distance a target needs to travel before crossing the lane boundary line.

[0047] To calculate the time it takes for a traversing target to enter the lane's geometric boundary line, a uniform linear motion model under the assumption of constant velocity is introduced. Since the traversing target moves from its current position with a constant lateral entry velocity, it will only touch the boundary line within a finite time if the direction of this lateral entry velocity is pointing inwards into the lane. This directional condition is determined through a product... The sign of the product is used to determine the direction of the velocity. If the product is less than or equal to zero, it indicates that the velocity direction is opposite to or parallel to the lane boundary line, and the target is moving away or has no lateral approach trend. In this case, there is no meaningful penetration time. When the product is greater than zero, the target is approaching the lane boundary, and the estimated intersection time is... The displacement is directly derived from the formula for uniform linear motion:

[0048] lateral distance The unit is meters, and the lateral cutting speed is... The unit is meters per second, calculated The unit is seconds. The value indicates the estimated time for a traversing target to reach the lane geometry boundary from its current position, i.e., to intrude into the convoy gap plane. This result is output as the estimated intersection time to subsequent displacement extrapolation steps.

[0049] For example, the deduction can be continued based on the environmental perception data provided in step S1. In the example of step S1, the physical distance detected by the forward millimeter-wave radar is 44.7 meters, and the coordinate data of the target crossing detected and calculated by the lateral lidar is the distance. Equal to 5.449 meters, azimuth angle θ equal to 47.23 degrees, absolute velocity vector rate The velocity is 1.803 meters per second, and the direction ψ is 56.31 degrees. This vehicle obtains the right-side boundary constant of its lane from the onboard high-precision map. meters, right and left boundary constants Meters, and based on real-time positioning, confirm that the vehicle is centered and its heading is parallel to the lane. First, perform a coordinate transformation, unfolding the coordinate data of the vehicle crossing the target into the vehicle coordinate system, and calculate its planar position: ; .because A positive value indicates that the target is actually located on the left outer side of the lane where the vehicle is located; therefore, the left boundary line constant is selected. Meters. Lateral distance: .

[0050] Based on the raw sensing data, the lateral cutting speed is calculated according to the direction ψ: Meters per second. At this point, determine the direction condition and calculate: A negative result indicates that the crossing target is veering to the left and away from the lane, and the system will... The value is directly assigned to 999 seconds to prevent the target from interfering with subsequent decisions.

[0051] To fully demonstrate the calculation logic for the target's lateral approach, assuming that in the aforementioned scenario the velocity direction ψ of the lateral target is -56.31 degrees towards the inside of the lane, the lateral cut-in velocity is: The direction is determined at this point by: A positive result indicates that the target is approaching the left boundary line, and the estimated intersection time is calculated using the formula: This example fully demonstrates the processing chain from environmental perception data to the estimated meeting time. Various technical features, such as target coordinate data, lateral cutting speed, and lane geometric boundary lines, are explicitly used in the calculation process. Furthermore, the estimated meeting time of 1.417 seconds accurately reflects the kinematic relationship between the lateral distance of 2.125 meters and the lateral cutting speed of 1.5 meters per second, confirming the applicability of the uniform linear motion model in this case.

[0052] Please see Figure 2 As shown, Figure 2 This is a schematic diagram of the cross-target intersection prediction plane mapping of the present invention. As shown in the figure, the intersection prediction module projects the initial perceived position of the cross-target onto the two-dimensional coordinate system of the vehicle, and, in conjunction with the calibrated left and right lane boundary lines, extrapolates the target prediction trajectory based on the target velocity vector; through the geometric intersection of this trajectory line and the lane boundary, the spatial position of the expected intersection point is intuitively established, thereby completing the geometric verification support for the expected intersection time.

[0053] S3. Read the current instantaneous maximum discharge power limit of the vehicle's battery management system and the torque response delay time of the drive motor. Convert the instantaneous maximum discharge power limit into the maximum allowable acceleration. Combine the torque response delay time with the expected intersection time to deduce the maximum physical displacement that the vehicle can move forward within the expected intersection time.

[0054] In a preferred embodiment, deriving the maximum physical displacement that the vehicle can move forward within the expected intersection time includes: inputting the instantaneous maximum discharge power limit into a preset vehicle dynamic response model to obtain the maximum wheel-end torque that the drive motor can generate under the discharge power constraint; calculating the theoretical maximum acceleration using the maximum wheel-end torque, the vehicle's mass, and the tire rolling radius, and treating the process of the motor torque rising from zero to the maximum wheel-end torque as an equivalent delay time based on the motor torque response delay time to obtain the effective acceleration time; and substituting the effective acceleration time and the theoretical maximum acceleration into the displacement integral formula for uniformly accelerated linear motion to obtain the maximum physical displacement, where the displacement integral formula is: the displacement is equal to the sum of the initial velocity multiplied by the expected intersection time multiplied by half the square of the effective acceleration time, and the initial velocity is the real-time vehicle speed.

[0055] Specifically, step S3 is initiated after obtaining the estimated intersection time output from step S2, aiming to quantify the extreme forward maneuvers the vehicle can take to seize right-of-way. The vehicle's internal state parameters required for this step are read from the vehicle network and control unit. The instantaneous maximum discharge power limit is broadcast by the battery management system via the controller area network bus at 100-millisecond intervals. This value represents the maximum DC power that the battery is allowed to release instantaneously under current state of charge, temperature, and aging conditions, measured in watts and denoted as . Torque response delay time of the drive motor Stored in the erasable read-only memory of the motor controller, this delay time is measured during the bench calibration phase before mass production of the vehicle. It is defined as the time it takes for the wheel-end torque to rise from zero to 90% of the maximum available wheel-end torque after the motor controller receives the full torque command. For a typical permanent magnet synchronous motor system, this delay time is taken as 0.05 seconds and remains constant in pure electric mode when the engine is not running.

[0056] The first step in deriving the maximum physical displacement is to determine the instantaneous maximum discharge power limit. The input is fed into a pre-defined vehicle dynamic response model. This model is a static analytical relationship that maps battery power constraints to the maximum wheel-end torque that the drive motor can generate at the current vehicle speed. The model considers the energy conversion path from the battery's DC bus to the drive wheels, where electrical energy is converted into AC power by an inverter to supply the drive motor, and the motor's output shaft is connected to the wheels via a single-stage reducer and differential. There are overall efficiency losses during power transfer; an efficiency coefficient is used. express, The average value of the measured values ​​of the same model electric drive assembly under multiple operating conditions on a dynamometer is set at 0.92. The angular velocity of the wheel rotation is... With real-time vehicle speed and tire rolling radius Directly related:

[0057] Among them, real-time vehicle speed Tire rolling radius is obtained from the wheel speed sensor signal of the anti-lock braking system. The value is 0.353 meters, which is the vehicle's factory specification parameter. Under this power constraint, the maximum wheel-end torque that the drive motor can produce is denoted as... Its calculation follows the physical relationship between power, torque, and angular velocity. To conform to the constant torque physical characteristics of the motor in the low-speed region and to avoid calculation divergence caused by the denominator approaching zero when the vehicle speed is extremely low or zero, an equivalent reference vehicle speed for the motor is introduced. For example, it can be set to 2.0 meters per second. The formula for calculating the maximum wheel-end torque is then revised as follows:

[0058] Substituting into the angular velocity expression, we get:

[0059] To obtain maximum wheel end torque Then, considering the quality of this vehicle With the same tire rolling radius Calculate the theoretical maximum acceleration. Vehicle mass. The curb weight is the sum of the curb weight and the currently estimated load capacity. The curb weight is obtained from the vehicle's factory certificate, while the load capacity is indirectly estimated by measuring the change in sprung mass using suspension height sensors, and is taken as 2100 kg. The tangential driving force of the drive axle to the ground is obtained by dividing the maximum wheel-end torque by the tire rolling radius.

[0060] Meanwhile, vehicles are inevitably affected by road resistance during operation, hence the definition of a constant rolling resistance coefficient. (Typical value of 0.015 for asphalt pavement), and gravitational acceleration g of 9.8 m / s², then the basic driving resistance experienced by the vehicle is... Therefore, the theoretical maximum acceleration Derived from Newton's second law:

[0061] At the expected meeting time Within the window, the drive motor cannot immediately build up the maximum wheel-end torque at moment zero; the torque reaches its maximum from zero. The process takes time The torque build-up process can be equivalent to a pure delay process, that is, assuming the preceding... The vehicle did not receive additional driving force during the period, and the driving force was... A single moment is instantly established and maintained. This effectively accelerates time. Defined as the net acceleration time after subtracting torque response delay from the expected convergence time, when Less than or equal to At that time, the effective acceleration time is zero:

[0062] This will effectively accelerate time. Compared with the theoretical maximum acceleration Substituting the displacement integral formula for uniformly accelerated linear motion, we calculate the maximum physical displacement that the vehicle can propel forward during the expected intersection time, denoted as . Based on the principle of superposition of independent motion, during the entire expected convergence time... Inside the window, the vehicle always maintains its initial speed. The contribution of the basic displacement component; while in the effective acceleration time Within the window, the vehicle additionally acquires the acceleration displacement component generated by the driving force. From this, the formula for calculating the total displacement is derived:

[0063] The first term on the right side of the formula represents the basic displacement caused by the initial velocity, and the second term represents the additional displacement generated by constant acceleration within the effective acceleration time. The unit is meters, which represents the physical positional propulsion limit that the front of the vehicle can achieve before crossing the target and intruding into the lane boundary under the dual strict constraints of the current battery discharge capacity and the motor torque build-up delay. This displacement directly determines the degree of compression of the minimum clearance with the vehicle in front.

[0064] For example, the scenario of steps S1 and S2 is continued for extrapolation. Step S1 has obtained data such as the physical distance of 44.7 meters, the distance across the target of 5.449 meters, and the absolute velocity vector rate of 1.803 meters per second. Step S2 uses this data to calculate the estimated intersection time for the target to enter the lane geometric boundary line. Seconds. The instantaneous maximum discharge power limit that this vehicle is currently reading from the battery management system. Watts, torque response delay time read from the motor controller's storage area Seconds. Real-time vehicle speed The wheel speed sensor recorded a speed of 8.2 meters per second, corresponding to 29.52 kilometers per hour. The vehicle's mass... Set to 2100 kg, tire rolling radius It is 0.353 meters, and the efficiency coefficient is 0.353 meters. The equivalent reference speed is 0.92. Set to 2.0 meters per second.

[0065] First, input the instantaneous maximum discharge power limit into the vehicle dynamic response model to calculate the maximum wheel-end torque (since 8.2 > 2.0, take 8.2): .

[0066] Subsequently, the physical parameters are derived sequentially based on the aforementioned formula: tangential driving force. Ox; Basic driving resistance Wow. Based on this, the theoretical maximum acceleration can be calculated: meters per second squared. Calculate the effective acceleration time: Seconds. Substitute the parameters into the displacement integral formula to calculate the maximum physical displacement: .

[0067] The step-by-step calculation, the uniform displacement part is: For the accelerated displacement portion, first calculate the square value: Multiply by the acceleration and half: Adding the two parts together gives: This value indicates that, within the projected intersection time window of 1.417 seconds, constrained by the 120 kW battery discharge limit and the 0.05-second motor response lag, and after overcoming normal driving resistance, the vehicle, starting from its current speed of 8.2 meters per second, can reduce the distance to the vehicle in front by a maximum of 17.473 meters. This rigorous calculation of the displacement limit provides a precise dynamic basis for subsequent steps to determine whether the convoy gap can be compressed to less than the geometric envelope width before crossing the target intrusion.

[0068] Please see Figure 3 As shown, Figure 3 This is a schematic diagram illustrating the time history of the maximum physical displacement projection of this invention. As shown in the figure, within a given expected intersection time window, the displacement projection module equates the motor torque establishment process to a response delay time where the front end is constant at zero. After the delay ends, the acceleration instantaneously jumps to the theoretical maximum acceleration constrained by power and remains constant, driving the speed to climb smoothly and linearly. By integrating the speed curve under this dynamic constraint over time, the dynamic evolution of the maximum projected displacement that the vehicle can advance within the intersection time is visually presented.

[0069] S4. Calculate the predicted displacement of the preceding vehicle within the expected intersection time by combining the longitudinal relative speed of the preceding vehicle. Add the predicted displacement of the preceding vehicle to the physical distance and then subtract the maximum physical displacement to obtain the estimated minimum distance. Compare the estimated minimum distance with the sum of the geometric envelope width of the target crossing and the preset longitudinal safety margin to generate the right-of-way decision index.

[0070] In a preferred embodiment, generating a right-of-way decision index includes: obtaining the geometric envelope width representing the physical space occupied by the target; comparing the estimated minimum distance with the sum of the geometric envelope width and a preset longitudinal safety margin; if the estimated minimum distance is less than the sum of the geometric envelope width and the longitudinal safety margin, generating a right-of-way preemption index representing compressing the vehicle distance to block the crossing path; if the estimated minimum distance is greater than or equal to the sum of the geometric envelope width and the longitudinal safety margin, generating a right-of-way yielding index representing allowing the target to cross.

[0071] Specifically, step S4 receives the physical distance output in step S1. The maximum physical displacement output in step S3 Simultaneously, the longitudinal relative velocity between the preceding vehicle and the current vehicle is obtained through the Doppler frequency shift data of the forward millimeter-wave radar in step S1. Combined with the vehicle's real-time speed Calculate the absolute speed of the vehicle in front. To ensure consistency in the logic of kinematic symbols, the system defines relative velocity. The value is positive when the vehicle in front is far away from the vehicle ahead, and negative when it is close to the vehicle ahead. Assume the vehicle in front is at the expected meeting time. While maintaining the current speed, calculate the predicted displacement of the vehicle in front. The estimated minimum distance is obtained through interpolation, denoted as . The estimated minimum clearance represents the remaining longitudinal clearance between the foremost point of the front bumper of this vehicle and the rear of the vehicle in front after this vehicle has performed a maximum acceleration forward thrust.

[0072] In the formula, the physical distance Predicted displacement of the vehicle in front With the maximum physical displacement All measurements are in meters, estimated minimum spacing. The unit is also meters.

[0073] Another key quantity used in conjunction with the estimated minimum clearance is the geometric envelope width of the target crossing and the longitudinal safety margin set by the system. The geometric envelope width is defined as the lateral span of the physical space occupied by the target crossing the horizontal plane, approximately perpendicular to the vehicle's direction of travel. This span directly determines whether the target can pass through a given width of clearance. The process of obtaining the geometric envelope width reuses the cluster geometric features extracted during point cloud data clustering and target recognition in step S1. After identifying the target crossing in step S1, the minimum bounding rectangle of all points within the target cluster projected onto the horizontal plane has been calculated, and the width of this rectangle is denoted as... The typical shoulder width of pedestrians ranges from 0.55 meters to 0.85 meters, while the handlebar width of non-motorized vehicles is typically between 0.60 meters and 1.00 meters. The real-time values ​​are dynamically determined from the actual scanning results of the lateral lidar, directly reflecting the geometric scale of the target at the current moment related to its path. The geometric envelope width is directly taken as... No additional coefficients need to be set. In degenerate scenarios where no complete clusters are identified, the system calls the preset default values: the default width for pedestrians is the statistical average of 0.70 meters, and for non-motorized vehicles it is 0.90 meters. This setting is based on the measurement statistics of more than 2,000 sets of real pedestrian and non-motorized vehicle laser point cloud samples. The longitudinal safety margin is denoted as... It is used to absorb vehicle dynamics control errors, sensor ranging noise, and to ensure the psychological safety of pedestrians crossing the road. Its value is set as a preset constant according to the autonomous driving safety specifications, such as 1.0 meter.

[0074] Complete the estimated minimum spacing With geometric envelope width With longitudinal safety margin After numerical preparation, the system performs a comparison and judgment to generate a right-of-way decision index. The right-of-way decision index is the core logical signal guiding the vehicle's subsequent actions of preempting or yielding right-of-way. The comparison relationship is defined as follows: if the estimated minimum distance is less than the sum of the geometric envelope width and the longitudinal safety margin, i.e. This means that if the vehicle utilizes its powertrain to its maximum capacity to reduce the distance to the vehicle in front, it can successfully compress the remaining gap to a point where it is insufficient to allow a target to pass laterally across its geometric shape. At this point, a right-of-way preemption index is generated. The right-of-way preemption index is a specific numerical identifier, for example, represented by the unsigned integer code 0x01. Its physical meaning is that the distance must be immediately compressed to physically close the gap in the convoy, thereby blocking the path of the crossing target. Conversely, if the estimated minimum distance is greater than or equal to the sum of the geometric envelope width and the longitudinal safety margin, i.e. This indicates that even if the vehicle accelerates forward, there is still enough space for the target to pass safely. In this case, attempting to seize the lane is not only ineffective but also increases the risk of collision, thus generating a right-of-way index. The right-of-way index is identified by different numbers, such as code 0x00, which indicates that the vehicle does not need to forcibly seize the lane but should instead actively give up the space to pass.

[0075] This judgment logic is strictly based on the feasibility of spatial geometry and kinematic prediction, ensuring that the vehicle will not ineffectively compress the distance and cause a collision when the target is already impassable, nor will it blindly rush through when the target is passable and cause danger.

[0076] For example, the complete calculation of this step is supported by a series of data instances from steps S1 to S3. Step 1: Measure the physical distance between this vehicle and the vehicle in front. Meters, while forward millimeter-wave radar measures the relative speed of the vehicle in front. Meters per second, simultaneously identifying pedestrians as cross-traversing targets, and the minimum bounding rectangle width of their clustered projection. The width of this width is taken as the geometric envelope width. Step 3, based on parameters such as the real-time vehicle speed of 8.2 meters per second, the instantaneous maximum discharge power limit of 120,000 watts, and the torque response delay time of 0.05 seconds, deduce the maximum physical displacement of the vehicle within the estimated intersection time of 1.417 seconds. meters. Absolute speed of the vehicle in front. Meters per second, predicted displacement of the preceding vehicle during the intersection time. Meters. Add the predicted displacement of the preceding vehicle to the physical distance, then subtract the maximum physical displacement to obtain the estimated minimum distance: The estimated minimum spacing of 38.563 meters is compared with the sum of the geometric envelope width of 0.8 meters and the safety margin of 1.0 meter (i.e., 1.8 meters). 38.563 meters is greater than 1.8 meters, thus satisfying the condition. Therefore, the generated right-of-way decision index is the right-of-way yielding index, identified by code 0x00. This result indicates that, under the current dynamic conditions, even if the vehicle accelerates forward at its maximum power to compress the lane distance, there is still a wide gap of 38.563 meters between it and the vehicle in front. Pedestrians have ample space to cross the lane gap comfortably, and the vehicle should implement a yielding strategy. This determination provides a clear and traceable decision source for subsequent steps controlling the motor or braking system and the directional acoustic array.

[0077] S5. Based on the right-of-way decision index, control the positive torque output of the vehicle's motor to match the instantaneous maximum discharge power limit in order to reduce the convoy spacing, or control the braking system to output negative torque in order to actively give way.

[0078] In a preferred embodiment, controlling the vehicle motor to output a positive torque that matches the instantaneous maximum discharge power limit to reduce platoon spacing, or controlling the braking system to output negative torque to actively yield, includes: analyzing a right-of-way decision index; when the analysis result is a right-of-way preemption index, generating a first torque request command and sending the first torque request command to the motor controller to make the motor output the maximum positive torque constrained by the instantaneous maximum discharge power limit; when the analysis result is a right-of-way yielding index, generating a second torque request command and sending the second torque request command to the braking system to make the braking system output a preset braking torque value to increase platoon spacing.

[0079] Specifically, step S5 receives the right-of-way decision index generated in step S4. This index is either a right-of-way preemption index with identifier 0x01 or a right-of-way yielding index with identifier 0x00. Parsing the right-of-way decision index involves reading the value of this identifier and using it as the basis to branch into different torque request generation paths. The transmission of the right-of-way decision index is completed through the shared memory between tasks on the vehicle's central computing platform. The transmission from the decision task to the control task has a delay of no more than 5 milliseconds, ensuring real-time connection between decision-making and execution.

[0080] When the analysis result shows that the right-of-way decision index equals the right-of-way preemption index, the system generates a first torque request command. The first torque request command is a data frame conforming to the CAN 2.0B format of the Controller Area Network Bus protocol. Its identifier is the conventional ID of the motor torque request message. The 8 bytes of the data field are sequentially filled with the motor mode word, the target torque value, and the checksum. The target torque value is precisely the maximum positive torque that the drive motor needs to output immediately, constrained by the instantaneous maximum discharge power limit. This torque is defined as the maximum wheel-end torque in step S3. To fully characterize the constraint transfer from battery power to wheel-end torque, and to maintain the low-speed constant torque physical characteristic constraint consistent with step S3, the maximum positive torque here adopts the output of the vehicle dynamic response model in step S3:

[0081] in Represents the instantaneous maximum discharge power limit, measured in watts; The overall efficiency of the electric drive system is taken as 0.92; The tire rolling radius is taken as 0.353 meters; Real-time vehicle speed, in meters per second; It serves as an equivalent reference speed to prevent torque divergence at extremely low speeds; The unit is Newton-meter (N·m). The target torque value in the first torque request command is set to... The torque direction sign is positive. This command is then sent to the motor controller, which is the power inverter control unit that drives the motor. Its internal high-speed current loop adjusts the duty cycle of the insulated gate bipolar transistor at a frequency of 10 kHz, so that the actual output torque of the motor quickly tracks the target value. Within the limit of the battery's discharge capacity, it quickly shortens the gap between the vehicle and the vehicle in front, realizing the physical action of seizing the right-of-way.

[0082] When the analysis results show that the right-of-way decision index equals the right-of-way yielding index, the system generates a second torque request command. This second torque request command is also a CAN message, but its target receiver is the electronic stability control unit (ESU) of the braking system. The ESU of the braking system has the function of driving the hydraulic pump to build braking pressure and adjusting the pressure of each wheel cylinder through solenoid valves. The target torque value included in the second torque request command is a preset braking torque value, which is pre-calibrated to a constant value based on the gentle deceleration intensity required for the yielding condition. . The setting is based on producing a smooth deceleration feel of 0.8 meters per second squared on a dry asphalt road surface. To accurately control deceleration and avoid over-braking, the braking torque calculation must deduct the contribution of the vehicle's natural rolling resistance. This is combined with the baseline rolling resistance defined in step S3. The target braking torque is determined by the formula:

[0083] in The expected magnitude of the gradual deceleration is 2100 kg. It is 0.8 meters per second squared. It is 308.7 Newtons. The length is 0.353 meters. The total deceleration force required by the vehicle is... After deducting natural resistance, the actual frictional braking force required from the braking system is: Awesome. The dynamic target torque value was calculated. Newton-meters (N·m). To match the control step size and command precision of the underlying actuator, it is rounded down to -480 N·m, with a negative sign. When the braking system receives this command, the electronic stability control unit distributes the hydraulic braking pressure to each wheel according to the total braking torque demand through feedforward and feedback algorithms, outputting a continuous and stable negative torque. This actively and smoothly decelerates the vehicle, gradually increasing the distance to the vehicle in front and creating clear yield space for crossing the target.

[0084] The entire process of step S5 involves calculating the right-of-way decision index and precisely sending the corresponding torque command to the actuator, enabling the switching of the vehicle's power output direction within tens of milliseconds. It serves as a key execution hub for preceding intelligent decision-making and vehicle physical response.

[0085] For example, the execution process is explained using the right-of-way yield index generated in step S4. Step S4 compares the physical distance of 44.7 meters, the predicted displacement of the preceding vehicle of 11.336 meters, the maximum physical displacement of 17.473 meters, the geometric envelope width of 0.8 meters, and the longitudinal safety margin of 1.0 meter. It finds that the estimated minimum distance of 38.563 meters is greater than the sum of the geometric envelope width and the safety margin of 1.8 meters, and determines to generate a right-of-way yield index with an identifier code of 0x00. Step S5 parses the identifier code and enters the yield branch. The system then generates a second torque request command. The target ID of the CAN data frame of this command is 0x1A2, and the target torque value field in the data field is filled with the decimal value -480, corresponding to the braking torque value of -480 N·m after calibration step size processing. The command is sent via the controller area network bus with a period of 2 milliseconds. The electronic stability control unit of the braking system successfully receives the message and immediately starts the hydraulic pressure build-up program, generating a total braking torque of -480 N·m at the four wheel cylinders. The braking torque, combined with the vehicle's own rolling resistance, causes the vehicle to gradually increase the distance to the vehicle in front at a deceleration of 0.8 meters per second squared. The convoy spacing gradually increases from 44.7 meters to clearly convey the intention to yield to pedestrians on the side. If it is necessary to display the preemptive path, the system will calculate the corresponding action under the opposite decision. If the instantaneous maximum discharge power limit is 120,000 watts and the real-time vehicle speed is 8.2 meters per second, then The target torque value entered in the first torque request command is +4752.59 N·m. Based on this, the motor controller will drive the motor to output a strong positive torque, quickly compressing the gaps between convoys and completing the physical blockade to seize right-of-way. The example completely closes the entire control chain from the decision index to the torque command and then to the vehicle's physical actions.

[0086] S6. Read the output torque status and control the directional acoustic array to emit a directional sound beam corresponding to the torque status to the position indicated by the coordinate data of the target crossing, so as to complete the information exchange of right-of-way and yielding.

[0087] In a preferred embodiment, reading the output torque status includes: monitoring the direction sign of the torque request command sent to the motor controller or braking system; when the direction sign of the torque request command is positive, determining that the current torque status is positive; when the direction sign of the torque request command is negative, determining that the current torque status is negative.

[0088] In a further preferred embodiment, controlling the directional acoustic array to emit a directional acoustic beam corresponding to the torque state towards the location indicated by the coordinate data of the target crossing the target includes: converting the coordinate data from the lidar coordinate system to a vehicle coordinate system centered on the directional acoustic array to generate acoustic beam aiming direction data; when in a positive torque state, calling a first audio signal and controlling the directional acoustic array to emit a directional acoustic beam containing the first audio signal in the direction indicated by the acoustic beam aiming direction data with a first beamwidth to warn and drive away the target crossing the target; when in a negative torque state, calling a second audio signal and controlling the directional acoustic array to emit a directional acoustic beam containing the second audio signal in the direction indicated by the acoustic beam aiming direction data with a second beamwidth greater than the first beamwidth to provide a wide-angle yield indication to the target crossing the target.

[0089] In a further preferred embodiment, the step of obtaining coordinate data of the target crossing further includes: detecting and obtaining the real-time relative position of the target crossing by the vehicle's lateral lidar; spatially registering the real-time relative position with the vehicle's driving planning path in the vehicle-mounted high-precision map to filter out interfering targets located outside the polygonal monitoring area constructed by extending a set distance along the driving planning path from the lane geometric boundary line, thereby obtaining the coordinate data of the effective target crossing.

[0090] Specifically, step S6 starts immediately after step S5 sends a torque request command to the actuator. Its function is to convert the vehicle's motion decision into a perceptible acoustic signal that crosses the target, completing the physical closed loop of right-of-way interaction. Step S6 first executes the step of reading the output torque status. The output torque status is not directly measured by the physical torque value on the drive shaft, but is determined by software monitoring the direction sign of the torque request command generated and sent to the motor controller or braking system in step S5. The direction sign is embedded in the data field of the torque request command CAN data frame. The highest bit of the target torque value field is the sign bit, where 0 represents a positive value and 1 represents a negative value. The signal monitoring task of the vehicle's central computing platform polls this signal bit at a frequency of 100 Hz. When the sign bit is 0, the direction sign of the torque request command is positive, indicating that the current torque state is positive, corresponding to the vehicle performing an acceleration action to seize right-of-way. When the sign bit is 1, the direction sign of the torque request command is negative, indicating that the current torque state is negative, corresponding to the vehicle performing an deceleration action to actively yield.

[0091] While reading the torque status, the system is responsible for acquiring the coordinate data of traversing targets. Multiple candidate point cloud clusters may be generated during continuous scanning by the lateral LiDAR. It is necessary to filter out the valid traversing targets that have a potential impact on the convoy's movement and lock their real-time relative positions. The lateral LiDAR continuously detects and outputs point clouds at a frequency of 10 Hz. After clustering and target recognition processing as described in step S1, each traversing target is assigned an identifier for tracking. To filter out interfering targets, the system spatially registers the real-time relative position of each traversing target with its absolute velocity vector and the vehicle's planned driving path in the onboard high-precision map. The onboard high-precision map not only includes lane geometric boundaries but also records the road topology ahead of the vehicle's current driving path. The spatial registration steps are as follows: Obtain the planar coordinates of the traversing target in the vehicle's coordinate system. And utilize its absolute velocity vector Extrapolate a preset time (e.g., 3 seconds) to obtain the predicted trajectory point sequence. Then, using the vehicle's geodetic coordinates and heading angle provided by the real-time dynamic differential positioning system, ... The predicted trajectory points are transformed to the geodetic coordinate system using coordinate translation and rotation matrices to obtain the geodetic coordinates of the points traversing the target. The system also extracts the predicted geodetic coordinates of the target. Simultaneously, it extracts geodetic coordinates of the lane boundary within a 50-meter radius along the planned driving path from a high-precision map, constructing a polygonal monitoring area. If the geodetic coordinates or predicted geodetic coordinates of a target falling within this polygon, the target is considered to be within the lane geometric boundary monitoring range, and its coordinate data is retained as valid target data. If the target is outside the monitoring range and its predicted trajectory does not intersect with the lane, such as a pedestrian on a sidewalk that is stationary or moving in the opposite direction away from the lane, it is filtered out. After registration and filtering, the final obtained coordinate data represents the distance of valid targets to the target in the vehicle's coordinate system. And the azimuth angle θ, whose value is consistent with the definition in step S1, and the distance Let θ be the straight line length of the projection of the geometric center point relative to the center of the rear axle of the vehicle onto the ground, and let θ be the angle between this line and the positive direction of the X-axis of the vehicle.

[0092] After obtaining the torque state determination result and the coordinate data of the effective target crossing, the step of controlling the directional acoustic array to emit a matching directional acoustic beam is executed. The directional acoustic array is a rectangular array panel composed of 8 ultrasonic transducer units, installed in the center of the front of the vehicle roof. The acoustic center of each transducer unit has known fixed coordinates in the array coordinate system. The directional acoustic array supports two radiation modes: narrow beam and wide beam. The first beam width is generated by adjusting the phase delay of each transducer unit using a beamforming algorithm to generate a precisely directional sound column, with its half-power beamwidth compressed to 5 degrees. The second beam width is expanded by reducing the amplitude of some units to increase the beam coverage, with the half-power beamwidth extended to 30 degrees.

[0093] The control process first requires transforming the coordinate data from the lidar coordinate system to a vehicle coordinate system centered on the directional acoustic array, generating acoustic beam aiming direction data. In step S1, the azimuth angle θ and distance of the coordinate data... Based on a vehicle coordinate system with the ground projection of the rear axle center of the vehicle as the origin, the installation position of the directional acoustic array in this vehicle coordinate system is determined by a three-dimensional translation vector. The description states that the translation vector is obtained through precise measurement and calibration at the vehicle's factory. rice, rice, Meters. Considering that the directional sound beam has a certain divergence coverage angle in the vertical direction, and that the hearing height of the main interactive objects (pedestrians / non-motorized vehicles) is similar to the array installation height, the system performs dimensionality reduction processing here, calculating only the aiming azimuth angle in the horizontal plane. The planar coordinates of the target traversing the original vehicle coordinate system are obtained through polar coordinate transformation:

[0094] The planar coordinates of the acoustic array center are:

[0095] Acoustic beam aiming direction data is derived from vector difference Determine its direction angle. From the arctangent formula:

[0096] Calculation, where and They are The Y and X components, and the angle, are used as the control target input for beam pointing to the beam controller of the directional acoustic array.

[0097] The audio content and beam pattern of the beam transmission are determined by the torque state. When in a positive torque state, the system retrieves the first audio signal stored in the audio memory. This first audio signal is a highly alert modulated tone, validated through extensive pedestrian behavior experiments, covering a frequency range of 1 kHz to 4 kHz, arranged in short, intermittent pulses to induce a conditioned reflex of stopping and retreating in pedestrians. Simultaneously, the directional acoustic array is controlled to operate at the first beamwidth, focusing the radiated energy of all eight transducer units onto... In terms of direction, the sound beam energy is highly concentrated, generating a peak sound pressure level of 95 decibels at a distance of 20 meters. The effective attenuation to 70 decibels occurs at a distance of only 0.8 meters from the beam center, achieving highly directional warning and deterrence against targets crossing the beam. When under negative torque conditions, the system activates a second audio signal, a slow, gentle, continuous bass warning with a center frequency around 500 Hz, to convey a mild indication of yielding. Simultaneously, the directional acoustic array is controlled to operate at the second beamwidth, with beam energy... The sound spreads evenly in the direction and within a 30-degree range, with the peak sound pressure level at 20 meters controlled at 78 decibels, and the sound field coverage width reaches more than 10 meters, indicating safe passage across the target in a wide-angle manner.

[0098] Through this acoustic interaction, the vehicle's intention to seize or yield the right-of-way is accurately transmitted to the crossing target. The right-of-way control link extends from perception, decision-making, and execution to human-vehicle interaction and collaboration, forming a closed physical control loop.

[0099] For example, in step S5, the right-of-way decision index is the right-of-way yield index. Therefore, the system generates a second torque request command and sends it to the braking system. The target torque value field in the CAN message is -480, and the highest sign bit is 1. The monitoring task for reading the torque status detects that the sign bit is 1, determining that the current torque state is negative. The registration process for acquiring coordinate data continuously tracks the pedestrian in the example of step S1. In the vehicle's coordinate system, its real-time relative position is 5.449 meters away and 47.23 degrees azimuth. Based on its velocity vector, it is predicted that the pedestrian will enter the lane. After geodetic coordinate transformation, its predicted trajectory falls within the lane's geometric boundary monitoring area, thus being confirmed as a valid crossing target. The coordinate data is maintained as follows: rice, Spend.

[0100] The next step is sound beam control. The target's planar coordinates in the original vehicle coordinate system are: The center position of the directional acoustic array is: Calculate the acoustic beam aiming vector: Direction angle: Because the torque state is negative, the system invokes the second audio signal, namely a soft, continuous bass with a center frequency of 500 Hz, and instructs the directional acoustic array to switch to the second beamwidth. The beam controller then... The phases of each transducer are adjusted so that the sound beam covers a wide area centered on the pedestrian's location with a 30-degree half-power width and a sound pressure level of 78 dB. While crossing, the pedestrian receives a gentle, wide-angle warning sound from the vehicle's direction, clearly perceiving that the vehicle is slowing down to yield, thus building confidence to cross safely. In the opposite scenario, where the torque is positive, the system will invoke the first audio signal and emit a highly directional warning sound within a 5-degree accuracy of the first beamwidth, directing the pedestrian to stop crossing. This process completely transforms the vehicle's internal right-of-way decisions into understandable physical signals for the external crossing target, achieving direct intentional interaction between the pedestrian and the vehicle.

[0101] Please see Figure 4 As shown, Figure 4 This is a polar coordinate schematic diagram of the beam pattern of the directional acoustic array of the present invention. As shown in the figure, the acoustic interaction module relies on the directional acoustic array and flexibly switches the radiation mode according to the torque state output by the system. When performing a strong right-of-way preemption, the array generates a high-gain warning narrow beam (solid line), which accurately focuses high sound pressure level energy on the target aiming direction to drive it away; when performing an active yielding, the array switches to a yielding wide beam (dashed line), which diffuses the sound field with a lower decibel value and a wider half-power beamwidth, conveying a soft, wide-angle yielding indication to the target area.

[0102] Please see Figure 5 As shown, the autonomous driving sightseeing vehicle right-of-way preemption system provided by the second aspect of the present invention includes: an environmental perception module, an intersection prediction module, a displacement inference module, a right-of-way decision module, a torque control module, and an acoustic interaction module.

[0103] The environmental perception module is used to obtain the physical distance between the vehicle and the vehicle in front, and to obtain the coordinate data and velocity vector of the target crossing the vehicle relative to the vehicle, as environmental perception data.

[0104] The intersection prediction module is used to calculate the estimated intersection time of the plane formed by the geometric boundary line of the target's arrival lane and the gap between the convoy, using environmental perception data and the geometric boundary line of the lane where the vehicle is located.

[0105] The displacement deduction module is used to read the current instantaneous maximum discharge power limit of the vehicle's battery management system and the torque response delay time of the drive motor, convert the instantaneous maximum discharge power limit into the maximum allowable acceleration, and combine the torque response delay time with the expected intersection time to deduce the maximum physical displacement that the vehicle can move forward within the expected intersection time.

[0106] The right-of-way decision module is used to calculate the predicted displacement of the preceding vehicle within the expected intersection time by combining the longitudinal relative speed of the preceding vehicle. The physical distance is added to the predicted displacement of the preceding vehicle and then the maximum physical displacement is subtracted to obtain the estimated minimum distance. The estimated minimum distance is compared with the sum of the geometric envelope width of the target and the preset longitudinal safety margin to generate the right-of-way decision index.

[0107] The torque control module is used to control the output of the vehicle's motor to match the instantaneous maximum discharge power limit with the right-of-way decision index to reduce the convoy spacing, or to control the output of the braking system to actively give way.

[0108] The acoustic interaction module is used to read the output torque state and control the directional acoustic array to emit directional sound beams corresponding to the torque state to the position indicated by the coordinate data of the target crossing, so as to complete the information interaction of right-of-way and yielding.

[0109] Each of the modules can be implemented in whole or in part through software, hardware, or a combination thereof. It supports hardware embedded in or independent of the processor in the computer device, and also supports software stored in the memory of the computer device, so that the processor can call and execute the operations corresponding to each of the above modules.

[0110] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for autonomous sightseeing vehicles to seize road rights, characterized in that: include: S1. Obtain the physical distance between the vehicle and the vehicle in front, and obtain the coordinate data and velocity vector of the target crossing the vehicle relative to the vehicle, as environmental perception data; S2. Using environmental perception data and the geometric boundary line of the lane where the vehicle is located, calculate the estimated intersection time of the plane formed by the geometric boundary line of the lane and the gap between the vehicles when the target crosses over. S3. Read the current instantaneous maximum discharge power limit of the vehicle's battery management system and the torque response delay time of the drive motor. Convert the instantaneous maximum discharge power limit into the maximum allowable acceleration. Combine the torque response delay time with the expected intersection time to deduce the maximum physical displacement that the vehicle can move forward within the expected intersection time. S4. Combine the longitudinal relative speed of the preceding vehicle to calculate the predicted displacement of the preceding vehicle within the expected intersection time. Add the predicted displacement of the preceding vehicle to the physical distance and then subtract the maximum physical displacement to obtain the estimated minimum distance. Compare the estimated minimum distance with the sum of the geometric envelope width of the target and the preset longitudinal safety margin to generate the right-of-way decision index. S5. Based on the right-of-way decision index, control the positive torque output of the vehicle's motor to match the instantaneous maximum discharge power limit in order to reduce the platoon spacing, or control the braking system to output negative torque in order to actively give way. S6. Read the output torque status and control the directional acoustic array to emit a directional sound beam corresponding to the torque status to the position indicated by the coordinate data of the target crossing, so as to complete the information exchange of right-of-way and yielding.

2. The method for autonomous sightseeing vehicle right-of-way preemption according to claim 1, characterized in that, Obtain the coordinate data and velocity vector of the target relative to the vehicle, including: The vehicle uses its side-mounted lidar to detect moving targets located to the side of the vehicle and acquire point cloud data of the moving targets. Clustering and target identification are performed on point cloud data to filter out pedestrians or non-motorized vehicles as cross-traffic targets; The position and distance of the geometric center point of the target being crossed relative to the vehicle are extracted to generate coordinate data. Based on the point cloud displacement of consecutive frames, the movement speed and direction of the target being crossed are calculated to form a velocity vector.

3. The method for autonomous sightseeing vehicles to seize right-of-way according to claim 1, characterized in that, Calculate the estimated intersection time of the plane formed by the target's approach lane geometry boundary and the convoy gap, including: Extract the velocity and direction components of the velocity vector from environmental perception data; The coordinate data and direction components of the target are mapped onto the lane projection plane of the vehicle's direction of travel to obtain the lateral cutting speed of the target. Using the lateral entry velocity and the lateral distance between the target and the lane geometric boundary line in the coordinate data of the target, the time for the target to enter the lane geometric boundary line is calculated using a uniform linear motion model, and this time is used as the expected intersection time.

4. The method for autonomous sightseeing vehicle right-of-way preemption according to claim 1, characterized in that, Determine the maximum physical displacement that this vehicle can move forward within the expected intersection time, including: The instantaneous maximum discharge power limit is input into the preset vehicle dynamic response model to obtain the maximum wheel end torque that the drive motor can generate under the discharge power constraint. Using the maximum wheel-end torque, the vehicle's mass, and the tire rolling radius, the theoretical maximum acceleration is calculated. Based on the motor torque response delay time, the process of the motor torque rising from zero to the maximum wheel-end torque is treated as a delay time to obtain the effective acceleration time. Substituting the effective acceleration time and the theoretical maximum acceleration into the displacement integral formula for uniformly accelerated linear motion, the maximum physical displacement is obtained. The displacement integral formula is: the displacement is equal to the sum of the initial velocity multiplied by the expected intersection time multiplied by half the square of the effective acceleration time, where the initial velocity is the real-time vehicle speed.

5. The method for autonomous sightseeing vehicles to seize right-of-way according to claim 1, characterized in that, Generate right-of-way decision indices, including: Obtain the geometric envelope width representing the physical space occupied by the target; The estimated minimum spacing is compared with the sum of the geometric envelope width and the preset longitudinal safety margin; If the estimated minimum spacing is less than the sum of the geometric envelope width and the longitudinal safety margin, then a right-of-way preemption index is generated to characterize the compression of vehicle spacing to block the through path. If the estimated minimum spacing is greater than or equal to the sum of the geometric envelope width and the longitudinal safety margin, then a right-of-way yield index is generated that represents the right of way that allows passage across the target.

6. The method for securing right-of-way for autonomous sightseeing vehicles according to claim 5, characterized in that, Controlling the vehicle's motor output to match the instantaneous maximum discharge power limit with positive torque to reduce platoon spacing, or controlling the braking system to output negative torque to actively yield, includes: The right-of-way decision index is analyzed. When the analysis result is the right-of-way preemption index, a first torque request command is generated and sent to the motor controller so that the motor outputs the maximum positive torque constrained by the instantaneous maximum discharge power limit. When the analysis result is the right-of-way index, a second torque request command is generated and sent to the braking system so that the braking system outputs a preset braking torque value to increase the platoon spacing.

7. The method for autonomous sightseeing vehicles to seize right-of-way according to claim 1, characterized in that, Read the output torque status, including: Monitor the direction sign of torque request commands sent to the motor controller or braking system; When the direction sign of the torque request command is positive, it is determined that the current state is in positive torque mode; When the direction sign of the torque request command is negative, it is determined that the current state is in negative torque.

8. The method for autonomous sightseeing vehicles to seize right-of-way according to claim 7, characterized in that, Controlling the directional acoustic array to emit a directional acoustic beam corresponding to the torque state at a position indicated by the coordinate data traversing the target includes: The coordinate data is transformed from the lidar coordinate system to the vehicle coordinate system centered on the directional acoustic array to generate acoustic beam aiming direction data. When in a positive torque state, the first audio signal is invoked, and the directional acoustic array is controlled to emit a directional acoustic beam containing the first audio signal in the direction indicated by the acoustic beam aiming direction data with a first beamwidth, so as to warn and drive away targets that cross the line. When in a negative torque state, the second audio signal is invoked, and the directional acoustic array is controlled to emit a directional acoustic beam containing the second audio signal in the direction indicated by the acoustic beam aiming direction data with a second beam width greater than the first beam width, so as to provide a wide-angle yield signal to the target crossing.

9. The method for autonomous sightseeing vehicles to seize right-of-way according to claim 1, characterized in that, The steps for obtaining coordinate data traversing the target further include: The vehicle uses its lateral lidar to detect and obtain the real-time relative position of targets crossing the road. Spatial registration is performed between the real-time relative position and the vehicle driving plan path in the vehicle-mounted high-precision map to filter out interfering targets located outside the polygonal monitoring area constructed by extending a set distance along the driving plan path from the lane geometric boundary line, and to obtain the coordinate data of the effective targets.

10. An autonomous driving sightseeing vehicle right-of-way preemption system, characterized in that: include: The environmental perception module is used to obtain the physical distance between the vehicle and the vehicle in front, and to obtain the coordinate data and velocity vector of the target crossing the vehicle relative to the vehicle, as environmental perception data. The intersection prediction module is used to calculate the estimated intersection time of the target crossing the lane geometric boundary line and the gap between the vehicle and the convoy by using environmental perception data and the geometric boundary line of the lane where the vehicle is located. The displacement estimation module is used to read the current instantaneous maximum discharge power limit of the vehicle's battery management system and the torque response delay time of the drive motor. It converts the instantaneous maximum discharge power limit into the maximum allowable acceleration and combines the torque response delay time with the expected intersection time to estimate the maximum physical displacement that the vehicle can move forward within the expected intersection time. The right-of-way decision module is used to calculate the predicted displacement of the preceding vehicle within the expected intersection time by combining the longitudinal relative speed of the preceding vehicle. The physical distance is added to the predicted displacement of the preceding vehicle and then the maximum physical displacement is subtracted to obtain the estimated minimum distance. The estimated minimum distance is compared with the sum of the geometric envelope width of the target and the preset longitudinal safety margin to generate the right-of-way decision index. The torque control module is used to control the output of the vehicle's motor to match the instantaneous maximum discharge power limit with the right-of-way decision index to reduce the platoon spacing, or to control the braking system to output negative torque to actively give way. The acoustic interaction module is used to read the output torque status and control the directional acoustic array to emit directional sound beams corresponding to the torque status to the position indicated by the coordinate data of the target, so as to complete the information exchange of right-of-way and yielding.

Images