A roadside sensor deployment optimization method and device for dynamic and static node collaborative sensing, and a storage medium

Abstract

The application relates to a roadside sensor layout optimization method and device facing dynamic and static node cooperative sensing, and a storage medium, wherein the method comprises the following steps: gridizing an intersection area, respectively establishing a roadside annular field of view and a vehicle-mounted circular field of view model, extracting an effective coverage range through polygon clipping and point-in-determination, and laying out sampling points in the coverage grid to obtain grid-level sensing accuracy by averaging according to distance attenuation; the comprehensive sensing accuracy of each vehicle is obtained by accumulating the actual occupied grid of the vehicle; under the condition that the total communication bandwidth is limited, a vehicle-sensor and vehicle-vehicle cooperative link is constructed, the sensor layout position, attitude parameter, bandwidth allocation and data transmission amount are jointly optimized, and the maximum vehicle sensing accuracy and the minimum sensor quantity are realized. Compared with the prior art, under the condition that the communication resource is limited, the overall sensing performance is maximized, and the system resource utilization efficiency is improved through a reasonable sensor site selection and bandwidth allocation strategy.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation systems and vehicle perception technology, and in particular to an optimized method for the deployment of roadside sensors for collaborative perception of dynamic and static nodes. Background Technology

[0002] With the rapid development of Intelligent Connected Vehicles (ICVs) and Vehicle-Infrastructure Cooperation (VIC) systems, environmental perception has become one of the core technologies for ensuring vehicle driving safety and intelligent decision-making. Currently, vehicles mainly rely on onboard sensors (such as cameras, LiDAR, and millimeter-wave radar) to achieve autonomous perception. However, due to limitations such as sensor installation height, interference from obstructions, blind spots, and environmental changes, the reliability and coverage of single-vehicle perception in complex traffic environments remain significantly insufficient.

[0003] To overcome the limitations of single-vehicle perception, cooperative perception technology has emerged. By deploying roadside sensing devices at key road nodes (such as intersections and ramp entrances) and combining them with V2X communication technology, perception information can be shared between vehicles and infrastructure, significantly improving the accuracy and safety of environmental perception. Especially in areas with frequent traffic conflicts, such as urban intersections, the rational deployment of roadside sensors is crucial for improving perception redundancy and ensuring traffic safety.

[0004] The existing technology CN119475962A establishes a probabilistic occupancy grid through Vissim-Carla co-simulation and optimizes RPU deployment using the Bayesian method with cross-entropy as the indicator. However, it only focuses on the location of the roadside equipment itself, does not introduce vehicle unit collaboration, and does not consider communication bandwidth constraints, making the results difficult to implement in resource-constrained scenarios.

[0005] The existing technology CN116170749A proposes to fuse vehicle-road perception information and compensate for time delay errors on the MEC server side to improve vehicle positioning accuracy. However, this method only solves the data fusion of existing equipment, does not solve the problem of how to deploy it in the early stage, and does not quantify the impact of dynamic occlusion on system performance.

[0006] Furthermore, roadside sensing equipment is costly to deploy, and its sensing performance is significantly affected by parameters such as installation location, sensor field of view, installation height, and pitch angle. Simultaneously, achieving efficient transmission and collaborative processing of sensed information under limited communication bandwidth constraints is a major challenge. Therefore, there is an urgent need for an optimization method that simultaneously considers multiple factors such as sensing accuracy and communication resource allocation to maximize the benefits of sensor deployment.

[0007] Current research has attempted to optimize sensor placement from single perspectives such as sensing coverage, communication efficiency, or deployment cost. However, these studies generally lack systematic modeling of the vehicle-infrastructure cooperative sensing mechanism and fail to fully integrate key elements such as the advantages of cooperative sensing and communication constraints. This results in deployment schemes exhibiting insufficient performance or deployment redundancy in practical applications. Therefore, proposing a roadside sensor deployment optimization method that integrates cooperative sensing requirements and communication constraints has significant theoretical and practical value. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a roadside sensor deployment optimization method for collaborative perception of dynamic and static nodes. It aims to maximize overall perception performance and improve system resource utilization efficiency by adopting reasonable sensor location and bandwidth allocation strategies when communication resources are limited.

[0009] The objective of this invention can be achieved through the following technical solutions:

[0010] The first aspect of this invention provides a method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes, comprising the following steps:

[0011] S1. The intersection area is gridded, and a circular field-of-view model of the roadside sensor and a circular field-of-view model of the vehicle sensor are established in a unified coordinate system. The effective coverage range of each sensor in the intersection is extracted by polygon clipping and point determination. Then, sampling points are evenly distributed in the covered grid cells. The attenuated perception accuracy is calculated based on the distance between the sampling points and the sensors and the average is taken to obtain the grid-level perception accuracy.

[0012] S2. Based on the grid cells actually occupied by the vehicle, the corresponding grid-level accuracy is accumulated to obtain the comprehensive perception accuracy of each vehicle by the roadside or vehicle-mounted sensors.

[0013] S3. Under the premise of limited total communication bandwidth, construct communication mechanisms for vehicle-sensor links and vehicle-vehicle collaborative perception links, jointly optimize sensor deployment locations, attitude parameters, bandwidth allocation and data transmission volume of the two types of links, so as to maximize the comprehensive perception accuracy of each vehicle while minimizing the number of sensors deployed.

[0014] Furthermore, in S1, the specific process of meshing the intersection area and establishing the roadside sensor ring field-of-view model and the vehicle-mounted sensor circular field-of-view model under a unified coordinate system includes:

[0015] Establish a global coordinate system with the lower left corner of the intersection as the origin, and discretize the study area into a grid using equilateral square units;

[0016] Based on the installation height, pitch angle, horizontal rotation angle, and field of view parameters of the roadside sensors, a ring-shaped field of view model centered on the sensor projection position is constructed. Based on the installation position and maximum sensing radius of the vehicle-mounted sensors, a circular field of view model centered on the vehicle center is constructed.

[0017] Furthermore, in S1, the specific process of extracting the effective coverage area of ​​each sensor within the intersection through polygon clipping and point-to-point determination includes:

[0018] Perform polygon clipping operations on the original ring or circular coverage area and the intersection boundary polygon to obtain the effective sensing area of ​​the polygon formed by their intersection.

[0019] Using a method for determining whether a point is inside a polygon, the center point of each grid is sequentially determined to be within the effective sensing area of ​​the polygon. Only grid cells that are completely within the effective sensing area of ​​the polygon are retained, thereby extracting the actual effective coverage area of ​​the sensor within the intersection.

[0020] Furthermore, in S1, the specific process of uniformly distributing sampling points within the covered grid cells, calculating the attenuated sensing accuracy based on the distance between the sampling points and the sensor, and averaging the results to obtain the grid-level sensing accuracy includes:

[0021] Within each grid cell that has been confirmed to be covered by the sensor, several sampling points are generated evenly in a row and column manner;

[0022] For each sampling point, first calculate its distance to the center of the sensor, and then obtain the sensing accuracy attenuation value of that point based on the radial attenuation law of the roadside circular field of view and the vehicle-mounted circular field of view respectively.

[0023] The arithmetic mean of the accuracy attenuation values ​​of all sampling points within the grid is taken as the grid-level sensing accuracy of that grid cell.

[0024] Furthermore, in S2, the specific process of obtaining the comprehensive perception accuracy of each vehicle by the roadside or vehicle-mounted sensors by accumulating the corresponding grid-level accuracy based on the actual grid cells occupied by the vehicle includes:

[0025] Based on the intersection of the vehicle outline and the mesh, determine all the mesh cells occupied by the vehicle;

[0026] The precision values ​​of each grid level obtained in S1 are called up one by one and the precision values ​​of the grid occupied by the vehicle are accumulated.

[0027] The sum of these values ​​represents the overall perception accuracy of the sensor for that vehicle.

[0028] Furthermore, in S3, the specific process of constructing the communication mechanism for the vehicle-sensor link and the vehicle-vehicle cooperative perception link includes:

[0029] A candidate communication object set is established for each vehicle, the communication object set including deployed roadside sensors and interconnectable nearby vehicle-mounted sensors;

[0030] Based on the relative positions of vehicles and sensors, and between vehicles and the channel state, the channel power gain of each communication pair is pre-calculated, and a shared bandwidth constraint is established with the total bandwidth as the upper limit.

[0031] Based on Shannon's theorem, the maximum amount of information that can be transmitted for each communication pair is estimated, forming capacity models for two types of cooperative sensing links: vehicle-sensor and vehicle-to-vehicle.

[0032] Furthermore, the specific process of estimating the maximum transmittable information for each communication pair according to Shannon's theorem, and forming capacity models for both vehicle-sensor and vehicle-to-vehicle cooperative sensing links, includes:

[0033] Based on the measured channel power gain and preset transmit power, and combined with the ambient noise level, calculate the maximum error-free transmission rate for each pair of communication links.

[0034] Multiply the maximum error-free transmission rate by the allowed link duration to obtain the maximum amount of sensed data that the link can carry in a single scheduling cycle;

[0035] The maximum data volume of the vehicle-sensor link and the maximum data volume of the vehicle-to-vehicle link are respectively written into the capacity model as upper limits of bandwidth allocation and data volume constraints in joint optimization.

[0036] Furthermore, in S3, the specific process of jointly optimizing sensor deployment locations, attitude parameters, bandwidth allocation, and data transmission volume of the two types of links to maximize the overall perception accuracy of each vehicle while minimizing the number of sensors deployed includes:

[0037] Binary decision variables are used to represent whether to deploy a sensor at each candidate location, and the sensor installation height, pitch angle, horizontal rotation angle, bandwidth and data volume of each communication link are set as continuous decision variables.

[0038] With the primary objective of minimizing the number of activated sensors and the primary constraint that the cumulative perception accuracy obtained by each vehicle after collaborative transmission is not lower than a set threshold, a hybrid integer optimization model is constructed by adding bandwidth sharing, link capacity, and attitude feasible domain constraints.

[0039] By solving the mixed-integer optimization model, the system simultaneously outputs the optimal sensor deployment scheme, attitude configuration, and bandwidth and data volume allocation results for the two types of links, thereby achieving simultaneous optimization of sensing accuracy and deployment cost.

[0040] A second aspect of the present invention provides an electronic device, including a memory and a processor, wherein the processor is used to execute a program in the memory to realize the roadside sensor deployment optimization method for collaborative sensing of dynamic and static nodes as described above.

[0041] A third aspect of the present invention provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, is used to execute the roadside sensor deployment optimization method for collaborative sensing of dynamic and static nodes as described above.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] By constructing a communication and perception fusion model based on collaborative perception between dynamic and static nodes, and fully considering communication bandwidth limitations and information transmission constraints, efficient information sharing and perception collaboration between roadside sensors and vehicles are achieved, thereby significantly improving overall perception coverage and accuracy. Building upon this, the invention introduces collaborative perception capabilities into sensor deployment decisions, enabling the system to reduce reliance on high-density roadside equipment while meeting perception accuracy requirements, significantly lowering sensor deployment costs. Simultaneously, by optimizing sensor configuration strategies, the invention effectively balances perception benefits with resource allocation fairness, improving system resource utilization efficiency and enhancing the intelligent perception and control capabilities of the traffic environment. Attached Figure Description

[0044] Figure 1 This is the overall flowchart corresponding to the embodiments of the present invention;

[0045] Figure 2 This is a diagram showing the sensor sensing accuracy under different sensor deployment schemes in the verification examples of this invention;

[0046] Figure 3 This is a diagram showing the sensing accuracy of a single sensor in a verification example of the present invention. Detailed Implementation

[0047] Overall, this invention proposes an optimized method for roadside sensor deployment based on collaborative perception between dynamic and static nodes. The aim is to significantly reduce sensor deployment costs and improve resource utilization efficiency while ensuring traffic safety and perception accuracy. The method first constructs a vehicle target detection priority model based on traffic scenarios, and then establishes a collaborative perception model between vehicles and sensors to quantify perception performance. Next, under limited communication bandwidth, communication resources are allocated through a V2V / V2I communication model to optimize information transmission efficiency between vehicles and sensors. Based on this, a deployment optimization model is constructed with the objective of minimizing the number of sensors deployed, and constraints on perception accuracy, sensor deployment parameters, and collaborative perception communication are set to achieve global optimization of the sensor deployment scheme. This method effectively integrates collaborative perception capabilities, balancing perception benefits and deployment costs, and has good practicality and prospects for widespread application.

[0048] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Component models, material names, connection structures, circuit structures, control methods, algorithms, and other features not explicitly described in this technical solution are considered common technical features disclosed in the prior art.

[0049] Example 1

[0050] This embodiment presents a method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes. The process is described in [link to documentation]. Figure 1 It includes the following steps:

[0051] Step 1: First, the intersection area is divided into equilateral grids, and the actual coverage area in the global coordinate system is calculated based on the sensor parameters (circular shape for roadside and circular shape for vehicle-mounted sensors). Then, a polygon clipping and point-based determination method is used to extract the effective coverage area of ​​the sensor within the intersection and identify the covered grid cells. By uniformly distributing sampling points in each covered grid cell, the sensing accuracy attenuation value of each point based on its distance from the sensor is calculated, and the average value is taken to obtain the overall sensing accuracy of the cell.

[0052] Step 2: Based on the sensor perception accuracy model established in Step 1, and combined with the space occupancy of vehicles in the grid, the overall perception accuracy of the roadside sensors for each vehicle is obtained by accumulating the perception accuracy of the grids covered by the vehicles. At the same time, the same method can be used to evaluate the perception accuracy of the on-board sensors installed on connected vehicles for surrounding target vehicles.

[0053] Step 3: Considering that roadside sensors can communicate with connected vehicles and that connected vehicles can also interconnect, a collaborative perception mechanism between vehicles and between vehicles and roads can be further realized. Based on this, a roadside sensor deployment optimization model based on the collaborative perception communication mechanism is established. That is, under the premise of limited communication bandwidth, the sensor deployment scheme, communication links and data transmission volume of each link are optimized to maximize the perception accuracy of vehicles and minimize the overall deployment cost of sensors.

[0054] According to claim 1, the method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes, step 1 includes the following steps:

[0055] Step 11: Discretize the intersection mesh. Define a global coordinate system (x, y) with the bottom left corner of the intersection as the origin (0, 0), the x-axis parallel to the horizontal direction of the road, and the y-axis parallel to the vertical direction of the road. The mesh edge length d... g Divide the study area (intersection) into M×N square units, then the center coordinates of the (m,n)th unit are:

[0056]

[0057] Step 12: Calculate the coverage area of ​​sensor s∈S, including roadside sensor S r and vehicle-mounted sensors S o That is, S = S r ∪S o In this invention, the coverage area of ​​the roadside sensor is assumed to be a ring-shaped region, and the coverage area of ​​the vehicle-mounted sensor is assumed to be a circle with radius R centered at the vehicle coordinate origin. i A circular area.

[0058] (1) For roadside sensors s∈S r Its sensing range is modeled as a ring-shaped area. This is based on installation parameters (sensor location (x...)). s ,y s Installation height h s Pitch angle a s Horizontal rotation angle b s The inner radius, along with the sensor's vertical field of view α and horizontal field of view β, can be calculated using the following formulas. and outer radius

[0059]

[0060] Inner arc and outer arc The arc lengths are respectively:

[0061]

[0062] The equations of the inner and outer circles are as follows:

[0063]

[0064]

[0065] Based on the horizontal rotation angle b s The two radial boundaries of the annular region are determined by the following equations:

[0066]

[0067] By combining equations (26)-(29), the coverage area Q of the roadside sensor s can be determined. s .

[0068] (2) For vehicle-mounted sensors s∈S o Its coverage area Q s Modeling is based on the vehicle coordinate origin (x) s ,y s ( ) is the center and radius R s Circular area:

[0069] (xx s ) 2 +(yy s ) 2 ≤R s 2 (10)

[0070] Step 13: Extract the intersection coverage area: Calculate the sensor coverage area Q using a polygon clipping algorithm (such as the Sutherland-Hodgman algorithm). s The intersection with the polygon of the intersection boundary yields the effective sensing area P within the actual intersection. s .

[0071] Step 14: Identify the covered grid cells. For each grid center point C m,n The method of determining whether a point lies within a polygon (such as the even-odd ray casting algorithm) is used to determine whether it falls within the perceptual region P. s Inside:

[0072]

[0073] Among them, C m,n,s =1 indicates that the grid cell (m,n) is covered by the sensor s.

[0074] Step 15: For each covered grid cell (C m,n,s=1), and multiple sampling points are evenly distributed within it. For each sampling point, two sensing accuracy attenuation coefficients are calculated based on its normalized longitudinal and lateral positions, and their product is used to define the total sensing accuracy of that point. Then, the average value of all sampling points is taken as the overall sensing accuracy of the grid cell. Within the covered grid cell (m,n), K points are divided in the global coordinate system {(x m,n,k ,y m,n,k The sampling points are uniformly distributed in the interval (k=1 to K), and the coordinates of the sampling point k are (x, y, k). m,n,k ,y m,n,k ).

[0075] Step 16: Calculate the sensing accuracy attenuation coefficient for each sampling point. Since roadside sensors and vehicle-mounted sensors have different sensing characteristics, their sensing accuracy is modeled separately.

[0076] (1) For roadside sensors s, s∈S r Taking LiDAR as an example, the sensing area is defined by (x... s ,y s The ring-shaped region centered on () represents the sensing accuracy of sampling point k. The attenuation coefficient of sensing accuracy decreases from the inner radius to the outer radius. Defined by formulas (12)-(13). The sensing accuracy is at its maximum when the distance between sampling points equals the inner radius; and at its minimum when the distance equals the outer radius.

[0077]

[0078] In formulas (12) and (13), and These represent the inner and outer radii of the annular region, respectively. This represents the distance between sensor s and sampling point k in the grid (m,n).

[0079] (2) For the vehicle-mounted sensor s, s∈S o The perception range is modeled as a radius R centered on the vehicle. s The circular region. Sensing accuracy decreases radially outward from the sensor center, and its sensing accuracy attenuation coefficient... As shown in formula (14):

[0080]

[0081] Step 17: Calculate the grid sensing accuracy based on sampling points. Let... This represents the maximum radial sensing accuracy of the sensor. For roadside sensors, sensing capability is quantified by their maximum data point density (dps) in the radial direction. The sensing accuracy of sampling point k in the grid (m,n) is... It can be represented as:

[0082]

[0083] in, This indicates the maximum sensing accuracy of the roadside sensor. This represents the sensing accuracy of sensor s at sampling point k in the grid (m,n).

[0084] The overall sensing accuracy of sensor s for grid cells (m,n) Take the arithmetic mean of the sensing accuracy of all sampling points within the unit, as shown in formula (16):

[0085]

[0086] In practice, step 2 includes the following steps:

[0087] Step 21: Using the sensor sensing accuracy model established in Step 1, the sensing accuracy of sensor s on the grid (m,n) can be obtained. In a given sensor deployment scheme (h s ,a s ,b s Under the condition that the sensing accuracy attenuation coefficient of sampling point k in grid (m,n) is calculated, the value of the attenuation coefficient is calculated. This is expressed as formula (17). Where d min (h s ,b s ) and d max (h s ,b s ) is determined by the deployment parameter h s and b s The determined sensing range and outer radius (which are pre-calculated and used as model input):

[0088]

[0089] Step 22: Calculate the perception accuracy p of sensor s∈S for target vehicle j within its range of interest. j,s The vehicle set is I, i∈I. Based on the grid occupied by vehicle j, the perception accuracy of sensor s for vehicle i is given by formula (18), where, It is a binary parameter, which is 1 if vehicle i occupies grid (m,n), and 0 otherwise.

[0090]

[0091] In specific implementation, step 3 includes the following steps:

[0092] Step 31: This method considers a candidate sensor set S, where each sensor s has known position coordinates (x, y, s). s ,y s The decision variables for each sensor include its installation height h. s Pitch angle a s and horizontal rotation angle b s That is, {h s ,a s ,b s Let E be a binary variable, s∈S. s Indicates whether the sensor is installed: If the sensor is installed, then E s =1; otherwise E s =0. Due to the high cost of sensor deployment, the objective function is to minimize the number of sensors installed while meeting the sensing accuracy requirements:

[0093] max∑ i∈I Ω i (20)

[0094] Step 32: Based on the deployment decision of each sensor s, constrain its configuration parameters (height, horizontal rotation angle, vertical rotation angle):

[0095] h min ≤h s ≤h max ,if E s =1 (21)

[0096] a min ≤a s ≤a max ,if E s =1 (22)

[0097] b min ≤b s ≤b max ,if E s =1 (23)

[0098] Among them, h min and h max These are the maximum and minimum height limits, a min and a max These are the maximum and minimum horizontal rotation angle limits, b min and b max These are the maximum and minimum vertical rotation angle limits, respectively.

[0099] To improve the field-of-view coverage of the sensors in the intersection area, the sensing accuracy p of each sensor at the intersection is... s A minimum coverage accuracy requirement must be met. Right now:

[0100]

[0101] Step 33: Maximum transmission bandwidth B i,s Based on the deployment of roadside sensors E s The bandwidth allocated to the vehicle-sensor communication pair is shown in Equation (26). Let the total available bandwidth at the intersection be B, and the communication bandwidth allocated between vehicle i and roadside / onboard sensor s be denoted as Bi. i,s All communication pairs share the total bandwidth B, as shown in equation (27):

[0102] ∑ i∈I ∑ s∈S B i,s ≤B (27)

[0103] Next, the amount of sensing information transmitted by each communication pair is calculated using equations (28)–(30). Equation (28) limits the amount of information transmitted between vehicle i and roadside sensor s. The maximum transmission rate must not be exceeded. According to Shannon's theorem, the maximum transmission rate for V2I (vehicle-to-infrastructure) and V2V (vehicle-to-vehicle) communication is given by the right side of the equation. Equation (29) indicates that the amount of information transmitted cannot exceed the maximum amount of sensed information that can be obtained after communicating with all paired sensors.

[0104]

[0105]

[0106] In equation (28), Indicates the sensor's transmit power. This represents the channel power gain for V2I or V2V communication pairs. The channel gain for all communication pairs can be pre-calculated based on the vehicle and sensor locations and used as input parameters to reduce the real-time computational burden on the model. σ 2 This indicates the noise power at the receiving end.

[0107] Finally, the perceived gain Ω obtained by each vehicle i i As given by equation (31):

[0108]

[0109] Application Example 1

[0110] This application example selects a six-lane, two-way intersection in Nanjing, Jiangsu Province, as a case study to verify the effectiveness of the proposed model. To reflect real traffic flow characteristics, 20 frames of vehicle distribution maps inside the intersection were extracted from surveillance videos during peak and off-peak hours, serving as the basis for sensor placement. The intersection was modeled in a Cartesian coordinate system with the lower left corner as the origin, the west entrance straight-ahead direction as the x-axis, and the south entrance straight-ahead direction as the y-axis. The intersection area was discretized into 144 square grids with sides of 1.75m. Twenty candidate sensor locations were selected at the four corners of the intersection, and their coordinates are shown in Table 1. The intersection sensing accuracy under each deployment scheme is as follows: Figure 2 As shown. Figure 2 The results show that as the number of sensors increases, the sensing accuracy gradually increases, and after deploying four sensors, the overlap of their sensing ranges increases significantly, leading to diminishing marginal benefits.

[0111] Table 1 Coordinates of Candidate Sensors

[0112]

[0113] To further illustrate the vehicle's communication and information transmission process, a detailed description is provided for the first frame of a vehicle distribution scene under three sensor deployment schemes. The specific deployment schemes of the three sensors are shown in Table 2, and the corresponding sensor sensing ranges are as follows: Figure 3 As shown in Table 2 and Figure 3 It can be seen that by carefully deploying each sensor at different positions and heights, their field of view can be maximized within the intersection, thereby significantly improving area coverage efficiency and data acquisition quality. For example, increasing the installation height and elevation angle can significantly widen the sensing range; while the horizontal rotation angle is mostly aligned along the diagonal of the intersection to ensure that the field of view covers the internal area as much as possible and reduces blind spots.

[0114] Table 2 Sensor Layout Scheme

[0115]

[0116] Example 2

[0117] This embodiment provides an electronic device, including a memory and a processor. The processor executes a program in the memory to implement the roadside sensor deployment optimization method for collaborative sensing of dynamic and static nodes as described above. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The memory may include Random Access Memory (RAM) and may also include Non-Volatile Memory, such as at least one disk storage device. The memory can be an internal memory of the Random Access Memory (RAM) type. The processor and memory can be integrated into one or more independent circuits or hardware, such as an Application Specific Integrated Circuit (ASIC). It should be noted that when the computer program in the aforementioned memory is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention.

[0118] Example 3

[0119] This embodiment provides a storage medium containing computer-executable instructions. When executed by a computer processor, the storage medium is used to perform the roadside sensor deployment optimization method for collaborative sensing of dynamic and static nodes as described above. The storage medium can be an electronic medium, magnetic medium, optical medium, electromagnetic medium, infrared medium, or semiconductor system or propagation medium. The storage medium may also include semiconductor or solid-state memory, magnetic tape, removable computer disk, random access memory (RAM), read-only memory (ROM), hard disk, and optical disk. Optical disks may include optical disc-read-only memory (CD-ROM), optical disc-read / write (CD-RW), and DVD.

[0120] The above description of the embodiments is provided to enable those skilled in the art to understand and use the invention. It will be apparent to those skilled in the art that various modifications can be made to these embodiments, and the general principles described herein can be applied to other embodiments without inventive effort. Therefore, the present invention is not limited to the above embodiments, and any improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the invention should be within the protection scope of the present invention.

Claims

1. A method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes, characterized in that, Includes the following steps: S1. The intersection area is gridded, and a circular field-of-view model of the roadside sensor and a circular field-of-view model of the vehicle sensor are established in a unified coordinate system. The effective coverage range of each sensor in the intersection is extracted by polygon clipping and point determination. Then, sampling points are evenly distributed in the covered grid cells. The attenuated perception accuracy is calculated based on the distance between the sampling points and the sensors and the average is taken to obtain the grid-level perception accuracy. S2. Based on the grid cells actually occupied by the vehicle, the corresponding grid-level accuracy is accumulated to obtain the comprehensive perception accuracy of each vehicle by the roadside or vehicle-mounted sensors. S3. Under the premise of limited total communication bandwidth, construct communication mechanisms for vehicle-sensor links and vehicle-vehicle collaborative perception links, jointly optimize sensor deployment locations, attitude parameters, bandwidth allocation and data transmission volume of the two types of links, so as to maximize the comprehensive perception accuracy of each vehicle while minimizing the number of sensors deployed.

2. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 1, characterized in that, In S1, the specific process of meshing the intersection area and establishing the roadside sensor ring field-of-view model and the vehicle-mounted sensor circular field-of-view model under a unified coordinate system includes: Establish a global coordinate system with the lower left corner of the intersection as the origin, and discretize the study area into a grid using equilateral square units; Based on the installation height, pitch angle, horizontal rotation angle, and field of view parameters of the roadside sensors, a ring-shaped field of view model centered on the sensor projection position is constructed. Based on the installation position and maximum sensing radius of the vehicle-mounted sensors, a circular field of view model centered on the vehicle center is constructed.

3. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 1, characterized in that, In S1, the specific process of extracting the effective coverage area of ​​each sensor within the intersection through polygon clipping and point-based determination includes: Perform polygon clipping operations on the original ring or circular coverage area and the intersection boundary polygon to obtain the effective sensing area of ​​the polygon formed by their intersection. Using a method for determining whether a point is inside a polygon, the center point of each grid is sequentially determined to be within the effective sensing area of ​​the polygon. Only grid cells that are completely within the effective sensing area of ​​the polygon are retained, thereby extracting the actual effective coverage area of ​​the sensor within the intersection.

4. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 1, characterized in that, In S1, sampling points are uniformly distributed within the covered grid cells. The attenuated sensing accuracy is calculated based on the distance between the sampling points and the sensor, and the average is taken to obtain the grid-level sensing accuracy. The specific process includes: Within each grid cell that has been confirmed to be covered by the sensor, several sampling points are generated evenly in a row and column manner; For each sampling point, first calculate its distance to the center of the sensor, and then obtain the sensing accuracy attenuation value of that point based on the radial attenuation law of the roadside circular field of view and the vehicle-mounted circular field of view respectively. The arithmetic mean of the accuracy attenuation values ​​of all sampling points within the grid is taken as the grid-level sensing accuracy of that grid cell.

5. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 1, characterized in that, In S2, the specific process of obtaining the comprehensive perception accuracy of each vehicle by the roadside or vehicle-mounted sensors by accumulating the corresponding grid-level accuracy based on the actual grid cells occupied by the vehicle includes: Based on the intersection of the vehicle outline and the mesh, determine all the mesh cells occupied by the vehicle; The precision values ​​of each grid level obtained in S1 are called up one by one and the precision values ​​of the grid occupied by the vehicle are accumulated. The sum of these values ​​represents the overall perception accuracy of the sensor for that vehicle.

6. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 1, characterized in that, In S3, the specific process of constructing the communication mechanism for the vehicle-sensor link and the vehicle-vehicle cooperative perception link includes: A candidate communication object set is established for each vehicle, the communication object set including deployed roadside sensors and interconnectable nearby vehicle-mounted sensors; Based on the relative positions of vehicles and sensors, and between vehicles and the channel state, the channel power gain of each communication pair is pre-calculated, and a shared bandwidth constraint is established with the total bandwidth as the upper limit. Based on Shannon's theorem, the maximum amount of information that can be transmitted for each communication pair is estimated, forming capacity models for two types of cooperative sensing links: vehicle-sensor and vehicle-to-vehicle.

7. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 6, characterized in that, The specific process of estimating the maximum transmittable information for each communication pair according to Shannon's theorem, and forming capacity models for two types of cooperative sensing links—vehicle-sensor and vehicle-to-vehicle—includes the following: Based on the measured channel power gain and preset transmit power, and combined with the ambient noise level, calculate the maximum error-free transmission rate for each pair of communication links. Multiply the maximum error-free transmission rate by the allowed link duration to obtain the maximum amount of sensed data that the link can carry in a single scheduling cycle; The maximum data volume of the vehicle-sensor link and the maximum data volume of the vehicle-to-vehicle link are respectively written into the capacity model as upper limits of bandwidth allocation and data volume constraints in joint optimization.

8. The method for optimizing the deployment of roadside sensors for collaborative sensing of dynamic and static nodes according to claim 1, characterized in that, In S3, the specific process of jointly optimizing sensor deployment locations, attitude parameters, bandwidth allocation and data transmission volume of the two types of links to maximize the overall perception accuracy of each vehicle while minimizing the number of sensors deployed includes: Binary decision variables are used to represent whether to deploy a sensor at each candidate location, and the sensor installation height, pitch angle, horizontal rotation angle, bandwidth and data volume of each communication link are set as continuous decision variables. With the primary objective of minimizing the number of activated sensors and the primary constraint that the cumulative perception accuracy obtained by each vehicle after collaborative transmission is not lower than a set threshold, a hybrid integer optimization model is constructed by adding bandwidth sharing, link capacity, and attitude feasible domain constraints. By solving the mixed-integer optimization model, the system simultaneously outputs the optimal sensor deployment scheme, attitude configuration, and bandwidth and data volume allocation results for the two types of links, thereby achieving simultaneous optimization of sensing accuracy and deployment cost.

9. An electronic device, comprising a memory and a processor, characterized in that, The processor is used to execute the program in the memory to implement the roadside sensor deployment optimization method for collaborative perception of dynamic and static nodes as described in any one of claims 1 to 8.

10. A storage medium containing computer-executable instructions, characterized in that, When executed by a computer processor, the storage medium of the computer-executable instructions is used to execute the roadside sensor deployment optimization method for collaborative perception of dynamic and static nodes as described in any one of claims 1 to 8.

Images