AGV (Automatic Guided Vehicle)-based intelligent reflecting cone dynamic laying and recycling system

Abstract

The invention provides an AGV-based intelligent reflective cone dynamic laying and recycling system, and relates to the technical field of reflective cone dynamic laying, and the system comprises an AGV transportation platform and a cloud decision system. The width d, the road speed limit vmax and the real-time traffic flow rho of the target road section are obtained; determining a basic layout included angle theta and a layout interval delta d corresponding to the combination of d, vmax and rho; generating a comprehensive feature vector XL; inputting the XL into a preset layout included angle weight and layout spacing weight prediction model to obtain a layout included angle weight lambda1 corresponding to theta and a layout spacing weight lambda2 corresponding to delta d; determining a target laying included angle theta'according to the delta d and the lambda 2, determining a target laying spacing delta d 'according to the delta d and the lambda 2, and sending the theta'and the delta d' to the AGV transportation platform so as to lay the reflective heap according to the theta 'and the delta d' through the AGV transportation platform. The method can enable a laying strategy to continuously approach to an optimal solution, and improves the reliability and the economical efficiency of a protection system for a long time.

Description

Technical Field

[0001] This invention relates to the field of dynamic deployment technology of reflective cones, and in particular to an intelligent dynamic deployment and retrieval system for reflective cones based on AGVs. Background Technology

[0002] In the field of safety protection during urban road and bridge construction, the placement, retrieval, and adjustment of reflective cones (traffic cones) have long relied on manual operation. Existing technologies have several significant drawbacks: Manual handling and placement of reflective cones is not only time-consuming and labor-intensive, but also inefficient. Especially in areas with heavy traffic, such as urban bridges and main roads, workers face a high risk of being hit by vehicles, leading to low coordination efficiency between construction and traffic management. Once traditional reflective cones are deployed, they are difficult to adjust dynamically and quickly according to changes in traffic flow (such as differences in traffic volume during peak / off-peak hours), adjustments to construction progress, or emergencies (such as accident site relocation or special vehicle passage), failing to achieve "on-demand protection." Furthermore, manual placement is affected by factors such as experience and physical strength, often resulting in suboptimal spacing and arrangement of reflective cones. Safety regulations, especially in complex terrains such as bridge curves, ramps, and narrowing sections, are prone to issues such as uneven spacing and route deviations, weakening the protective effect. When construction is completed or adjustments are made, manual retrieval of reflective cones is also inefficient. Furthermore, there is a lack of automated retrieval methods for cones that have tipped over, overturned, been partially buried, or shifted in position, requiring manual intervention and further increasing safety risks. In addition, the few automated warning devices have problems such as low positioning accuracy (difficult to place accurately on uneven roads and slopes), weak retrieval capacity (only able to handle cones in ideal conditions), poor environmental adaptability (unable to cope with complex terrain or weather), and high cost. Moreover, they lack intelligent decision-making mechanisms that are deeply integrated with the shape and deployment strategy of reflective cones, making it difficult to meet the high-efficiency protection needs of urban bridge construction. Summary of the Invention

[0003] To address the aforementioned technical problems, the technical solution adopted by this invention is as follows: According to a first aspect of this application, an intelligent reflective cone dynamic deployment and retrieval system based on AGV is provided, the system comprising: an AGV transportation platform and a cloud-based decision-making system; the cloud-based decision-making system is used to perform the following steps: S100, obtain the width d of the target road segment and the road speed limit v. max and real-time traffic flow ρ; S200, according to d, v max Based on ρ and the preset mapping table QT for the base layout angle and spacing, determine d and v. maxThe combination of ρ corresponds to the basic layout angle θ and the layout spacing Δd; where QT includes several rows, each including a set of road width, road speed limit and traffic flow, as well as the corresponding basic layout angle and layout spacing; the basic layout angle is the angle between the straight line formed by the preset number of reflective cones and the center line of the target road segment; S300 generates a comprehensive feature vector XL based on the static parameters, dynamic parameters, and environmental parameters of the target road segment; S400, input XL into the preset layout angle weight and layout spacing weight prediction model to obtain the layout angle weight λ1 corresponding to θ and the layout spacing weight λ2 corresponding to Δd; S500, based on θ and λ1, determine the target placement angle θ'=λ1×θ, and based on Δd and λ2, determine the target placement spacing Δd'=λ2×Δd; S600 sends θ' and Δd' to the AGV transport platform so that the reflector stack can be deployed according to θ' and Δd' via the AGV transport platform.

[0004] The present invention has at least the following beneficial effects: The present invention relates to an intelligent reflective cone dynamic deployment and retrieval system based on AGVs, which first determines the target road segment's width d and speed limit v. max The system obtains the basic layout angle θ and spacing Δd from the mapping table QT based on real-time traffic flow ρ, ensuring that the initial plan conforms to the basic characteristics of the road segment. Then, by integrating the comprehensive feature vector XL, which combines static, dynamic, and environmental parameters, and generating weights through the prediction model, the basic parameters are dynamically corrected. This ensures that the target angle θ' and spacing Δd' accurately match specific road conditions (such as curves, weather changes, etc.), solving the problem of mismatch between traditional fixed patterns and complex scenarios. Simultaneously, the system incorporates real-time traffic flow and dynamic parameters as decision-making basis. Combined with the model prediction weight adjustment mechanism, it can optimize layout parameters in real time according to traffic conditions (such as peak / off-peak traffic flow, sudden congestion), achieving a dynamic balance between "safety and efficiency." Furthermore, the cloud-based decision-making system automates parameter calculation and optimization, enabling deployment via AGVs without human intervention, significantly improving operational efficiency and avoiding safety risks associated with personnel working in traffic flow, making it particularly suitable for areas with high traffic density. Moreover, the two-layer decision-making logic of "basic parameters + model correction" forms a reusable and iterative intelligent mechanism. As data accumulates, the predictive model can continuously optimize the accuracy of weight calculations, making the deployment strategy constantly approach the optimal solution, possessing adaptability and evolutionary capabilities, and improving the reliability and economy of the protection system in the long term. Attached Figure Description

[0005] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0006] Figure 1 The flowchart illustrates the steps executed by the cloud-based decision-making system of the AGV-based intelligent reflective cone dynamic deployment and recycling system provided in this embodiment of the invention. Detailed Implementation

[0007] 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.

[0008] It should be noted that, based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Furthermore, this device and / or practice the method can be implemented using other structures and / or functionalities besides one or more of the aspects set forth herein.

[0009] The following will refer to Figure 1 The flowchart shown is a cloud-based decision-making system execution step of an AGV-based intelligent reflective cone dynamic deployment and recycling system, which introduces an AGV-based intelligent reflective cone dynamic deployment and recycling system.

[0010] The AGV-based intelligent reflective cone dynamic deployment and retrieval system includes: an AGV transportation platform and a cloud-based decision-making system; the cloud-based decision-making system is used to execute the following steps: S100, obtain the width d of the target road segment and the road speed limit v. max And real-time traffic flow ρ.

[0011] In this embodiment, the target road segment width d is achieved through a 16-line LiDAR and a LiDAR SLAM module mounted on the AGV. The LiDAR scans the construction area in real time, constructs a centimeter-level point cloud map, and analyzes the roadside boundaries (such as curbs and guardrails) using the point cloud data to calculate the effective width (the shoulder width needs to be deducted for bridge sections).

[0012] Speed ​​limit v maxThe data is obtained through the city traffic data API connected to the cloud-based decision-making system. This data comes from the road sign database of the traffic management department, ensuring consistency with the legal speed limit of the target road section.

[0013] Real-time traffic flow ρ: Obtained in real time through traffic big data platforms (such as urban traffic monitoring systems and floating car data), the unit is "vehicles / km", reflecting the current traffic congestion level of the target road segment.

[0014] It replaces manual measurement and statistics, avoiding human error (e.g., the error in road width measurement can be controlled within ±5cm, far higher than the ±50cm of manual measuring tape); the data is highly real-time (traffic flow is updated every 30 seconds), providing an accurate and timely input basis for subsequent dynamic decision-making, and solving the problem of lag in traditional manual deployment that relies on experience judgment.

[0015] S200, according to d, v max Based on ρ and the preset mapping table QT for the base layout angle and spacing, determine d and v. max The combination of ρ corresponds to the basic layout angle θ and the layout spacing Δd; where QT includes several rows, each including a set of road width, road speed limit and traffic flow, as well as the corresponding basic layout angle and layout spacing; the basic layout angle is the angle between the straight line formed by the preset number of reflective cones and the center line of the target road segment.

[0016] The mapping table QT is constructed based on historical construction cases (500+ urban bridge and main road construction scenarios) and safety specifications. QT is shown in Table 1.

[0017] Table 1 Basic layout angle θ: defined as the angle between the straight line formed by the first 3-5 reflective cones and the center line of the road section, used to guide traffic flow to gradually deviate from the construction area and avoid the risk of sharp turns.

[0018] Parameter matching logic: The cloud system uses d and v max The numerical range of ρ is matched with the closest combination in QT, and the corresponding θ and Δd are quickly invoked.

[0019] Based on historical data and pre-set parameters, the initial deployment is ensured to meet the "safety baseline" and avoids the arbitrariness of manual placement based on experience; the matching process takes less than 0.5 seconds, leaving time for subsequent dynamic adjustments and improving system response efficiency.

[0020] Furthermore, the reflective cone is a deformable intelligent reflective cone with a three-stage telescopic structure; it has a built-in counterweight self-balancing chassis, a multispectral warning light at the top of the cone, a cone pressure sensor, and a 5G-V2X communication module; the bottom of the cone is equipped with a ring-shaped electromagnet array to achieve magnetic docking with the robotic arm; and the surface of the cone is coated with a thermochromic self-healing coating.

[0021] S300 generates a comprehensive feature vector XL based on the static parameters, dynamic parameters, and environmental parameters of the target road segment.

[0022] Static parameters include road type (such as continuous beam bridge, urban arterial road, obtained through electronic map), radius of curvature R (analyzed by laser SLAM point cloud), and length of construction area (calculated from electronic fence boundary).

[0023] Dynamic parameters include real-time traffic flow speed V (derived from traffic big data), construction progress (e.g., "30% of the repairs have been completed," synchronized by the construction management system), and the status of adjacent cones (e.g., "whether there is tilt / collision," uploaded through the cone group self-organizing network).

[0024] Environmental parameters include weather type (sunny / heavy rain / fog, obtained from meteorological API), road surface humidity (detected by humidity sensor on AGV), and wind speed (obtained by wind speed sensor integrated in cone, used to determine whether vibration suppression needs to be activated).

[0025] Generating the feature vector XL: After quantizing the above parameters, they are concatenated according to fixed dimensions (e.g., XL=[road type code, R, real-time V, humidity value, ...]) to form structured data that can be input into the model.

[0026] Breaking through the limitations of traditional "single-parameter decision-making", it achieves multi-dimensional integration of "static terrain + dynamic traffic + real-time environment", providing a comprehensive basis for subsequent weight prediction; it quantifies unstructured information (such as weather and construction progress), enabling AI models to accurately learn scene features and improve the adaptability of decision-making (such as automatically strengthening the safety factor in rainy weather).

[0027] S400, input XL into the preset layout angle weight and layout spacing weight prediction model to obtain the layout angle weight λ1 corresponding to θ and the layout spacing weight λ2 corresponding to Δd.

[0028] Model construction: A reinforcement learning model (such as a deep Q-network) is used. The training data consists of "feature vector-optimal weight" samples from 500+ typical construction cases (the optimal weights are determined by post-accident rate and traffic efficiency evaluation). The model input is XL, and the outputs are λ1 (0.8-1.2) and λ2 (0.7-1.3).

[0029] The physical meaning of weights: λ1: The coefficient for adjusting the included angle θ (e.g., when the curve R < 50m, λ1 is increased to 1.1 to make the included angle θ' more inclined and guide the traffic flow away from the construction area of ​​the curve).

[0030] λ2: The coefficient for adjusting the spacing Δd (e.g., when ρ>100 vehicles / km during the morning peak, λ2 is reduced to 0.8 to shrink Δd' and increase the density of the deployment to enhance protection).

[0031] Model inference process: The cloud system inputs XL into the model and outputs λ1 and λ2 within 0.3 seconds, and supports online iteration (automatically updates the model parameters with the actual effect after each construction is completed).

[0032] This step enables "data-driven optimization" of weights, replacing the traditional "fixed coefficients" and making the adjustment more in line with real-time scenarios (such as automatically reducing λ2 to 0.7 in foggy weather to compensate for insufficient visibility by increasing spacing); the model continues to evolve, and with the accumulation of construction cases, the accuracy of weight prediction improves (85% in the initial stage → 98% after 1000 iterations), solving the problem of "difficulty in human prediction when the scene is complex".

[0033] S500, based on θ and λ1, determine the target placement angle θ'=λ1×θ, and based on Δd and λ2, determine the target placement spacing Δd'=λ2×Δd.

[0034] Calculation logic: Based on the basic parameters of S200 and the weights of S400, dynamic correction is achieved through simple multiplication. For example: with a base θ=15°, in a curve scenario λ1=1.1→θ'=16.5° (more inclined, enhancing curve guidance); with a base Δd=6m, during the morning rush hour λ2=0.8→Δd'=4.8m (reinforced, improving safety during congestion).

[0035] Constraint verification: The system automatically checks whether θ' (10°-30°) and Δd' (3m-8m) are within the safe range (Δd'∈[3m,8m]"). If they are outside the range, the system will force correction to the boundary value.

[0036] By combining the "standardization of basic parameters" and the "scenario adaptability of weights", the final parameters θ' and Δd' not only meet the safety baseline, but also accurately respond to dynamic changes (such as Δd' automatically reducing by 30% during heavy rain, solving the problem of insufficient protection under extreme weather conditions with traditional fixed spacing); the calculation process is simple and efficient, ensuring that the AGV can respond in real time and avoiding decision delays that affect the construction progress.

[0037] S600 sends θ' and Δd' to the AGV transport platform so that the reflector stack can be deployed according to θ' and Δd' via the AGV transport platform.

[0038] Command transmission: The cloud sends θ' and Δd' to the AGV via 5G communication with a transmission delay of <0.1 seconds, and includes dynamic route planning (generated by laser SLAM map to avoid manhole covers and potholes).

[0039] The execution process of AGV: The AGV travels at 5km / h along the planned path. The RGB-D camera on the wrist of the robotic arm identifies the road surface in real time to ensure a positioning accuracy of ±2mm. The release angle of the robotic arm is adjusted by pressing θ' (e.g., when θ'=16.5°, the robotic arm rotates to the corresponding angle to release the cone). The release timing is controlled by pressing Δd' (e.g., when Δd'=4.8m, the robotic arm is triggered to release after the AGV has traveled 4.8m).

[0040] The system achieves closed-loop automation of "decision-execution", with a robotic arm positioning accuracy of ±2mm, which is much higher than the ±50cm of manual placement, thus solving the problem of "insufficient accuracy of manual placement". The AGV executes instructions precisely, avoiding deviations caused by human fatigue and experience differences, and does not require personnel to enter the traffic flow, reducing the accident rate in the construction area by 90%.

[0041] Through a closed loop of "data acquisition - basic decision-making - feature fusion - intelligent weighting - dynamic correction - precise execution", the system retains the basic framework of the standard, while achieving dynamic adaptation through AI and multi-parameter fusion. Ultimately, it achieves dual optimization of "security protection" and "traffic efficiency", improving efficiency by 300% compared to traditional manual deployment and achieving a recovery rate of 98%, perfectly solving the technical pain points of "low efficiency, high risk and poor flexibility".

[0042] Furthermore, after step S600, the method further includes the following steps: The S700 uses different types of sensors pre-set on the reflective cones to obtain the number of reflective cones NUM1 that were hit within a preset historical time period.

[0043] A pressure sensor is used to detect collisions with a force greater than 50N and lasting for 2 seconds. Simultaneously, the cone group uploads status information, including "whether it was hit" and "time of collision," to the cloud in real time via a Zigbee+LoRa dual-mode network.

[0044] Preset historical time periods: Set according to the real-time requirements of the construction scenario, such as 30 minutes (peak hours on urban main roads) or 1 hour (low traffic hours at night), to ensure that short-term collision trends can be captured while avoiding frequent adjustments that could cause system fluctuations.

[0045] NUM1's statistical logic: The cloud-based decision-making system filters out the reflective cone IDs that are "marked as true" within a preset time period from the status data uploaded by the cone group, and counts the total number after deduplication (to avoid the same cone being counted repeatedly for multiple collisions).

[0046] By utilizing the built-in sensing capabilities of reflective cones, collisions can be "passively monitored and then actively statistically analyzed," solving the problem of traditional reflective cones being "passively protected but unable to provide feedback on the protective effect." The setting of historical time periods balances real-time performance and stability, avoiding blind adjustments due to instantaneous collisions and preventing missed risk windows due to delayed responses.

[0047] S710, if NUM1=0, then do not adjust θ' and Δd'; otherwise, proceed to S720.

[0048] Judgment logic: The cloud system compares NUM1 with 0. If NUM1=0, it means that the protection scheme corresponding to θ' and Δd' is effective (the traffic flow has not collided with the cone), and the existing parameters are maintained; if NUM1≥1, it means that there are weak links in the protection (such as the angle being too small, causing the traffic flow to be close to the construction area, or the spacing being too large, causing the warning to be discontinuous), and dynamic adjustment needs to be initiated.

[0049] Execution mechanism: This judgment is a "lightweight trigger switch", which does not require complex calculations and takes less than 0.1 seconds, ensuring a fast system response.

[0050] Avoid "meaningless adjustments" and reduce the waste of system computing power; clarify the triggering conditions to ensure that adjustments are only initiated when there is a "real risk of collision" and improve the accuracy of system decision-making.

[0051] S720, adjust θ' and Δd' based on NUM1 and the total number of reflective cones NUM deployed.

[0052] Furthermore, step S720 includes the following steps: S721, based on NUM1 and NUM, determine the impact rate τ of the reflective cone τ = NUM1 / NUM.

[0053] NUM Acquisition: NUM is the total number of reflective cones actually deployed in the current construction area, which is synchronized to the cloud from the AGV transportation platform's operation records.

[0054] The calculation of τ: The collision rate is obtained through simple division (e.g., if 10 cones are deployed and 2 are hit within 30 minutes, then τ = 2 / 10 = 20%), which quantitatively reflects the risk level of the current protection plan.

[0055] Transforming "absolute collision count" into "relative collision rate" eliminates the interference of "differences in the total number of cones" on risk assessment (e.g., 2 out of 10 cones being hit versus 3 out of 20 cones being hit, τ=20% and 15% clearly indicate that the former has a higher risk); quantitative indicators provide an objective basis for subsequent adjustments and avoid the bias of subjective human judgment.

[0056] S722, based on τ and the preset dynamic adjustment coefficient mapping table QR, determine the dynamic adjustment coefficient γ1 of θ' and the dynamic adjustment coefficient γ2 corresponding to Δd'.

[0057] Construction of the Dynamic Adjustment Coefficient Mapping Table (QR): The QR is generated based on the correlation analysis between historical collision data and optimized protection parameters, as shown in Table 2. Table 2 Parameter matching logic: The cloud system matches the corresponding γ1 and γ2 in QR based on the calculated τ (e.g., when τ=20%, it calls γ1=1.15 and γ2=0.85).

[0058] Based on historical data, the adjustment coefficient is preset to ensure that the adjustment range is "both effective and not excessive" (e.g., only a slight adjustment is made when τ=5% to avoid excessive protection due to minor collisions affecting traffic efficiency); the mapping table structure is simple and the call speed is fast (<0.2 seconds), which can meet the real-time requirements of construction scenarios.

[0059] S723, adjust θ' to θ'×γ1 and Δd' to Δd'×γ2.

[0060] Adjustment logic: Angle adjustment: θ'×γ1 (e.g., original θ'=15°, γ1=1.15→new θ'=17.25°). By increasing the angle, the "protective boundary" formed by the reflective cones is moved further away from the construction area, guiding the traffic flow to deviate earlier.

[0061] Spacing adjustment: Δd'×γ2 (e.g., original Δd'=6m, γ2=0.85→new Δd'=5.1m), to enhance the continuity of warnings by reducing the spacing.

[0062] Secondary verification: After adjustment, the system automatically checks whether the new parameters are within the safe range (θ'∈[10°,30°], Δd'∈[3m,8m]). If they exceed the range, they are truncated to the boundary value.

[0063] A closed loop of "collision feedback → parameter optimization" is achieved, enabling the protection plan to evolve dynamically with actual risks. The adjusted parameters are more in line with the on-site risks, and actual tests have shown that the subsequent collision rate can be reduced by more than 60%, while also taking into account traffic efficiency (avoiding excessive adjustments that lead to waste of road resources).

[0064] By employing a mechanism of "collision monitoring - risk quantification - dynamic adjustment," the system breaks through the traditional "fixed and unchanging after deployment" model, achieving the "self-correction" capability of the protection scheme. This not only addresses the pain point of "human intervention being unable to respond to collision risks in real time," but also ensures a balance between protective effectiveness and traffic efficiency through quantitative adjustments, further enhancing the system's intelligence and reliability.

[0065] In this embodiment, the width d and speed limit v of the target road segment are first determined. maxThe system obtains the basic layout angle θ and spacing Δd from the mapping table QT based on real-time traffic flow ρ, ensuring that the initial plan conforms to the basic characteristics of the road segment. Then, by integrating the comprehensive feature vector XL, which combines static, dynamic, and environmental parameters, and generating weights through the prediction model, the basic parameters are dynamically corrected. This ensures that the target angle θ' and spacing Δd' accurately match specific road conditions (such as curves, weather changes, etc.), solving the problem of mismatch between traditional fixed patterns and complex scenarios. Simultaneously, the system incorporates real-time traffic flow and dynamic parameters as decision-making basis. Combined with the model prediction weight adjustment mechanism, it can optimize the layout parameters in real time according to traffic conditions (such as peak / off-peak traffic flow, sudden congestion), achieving a dynamic balance between "safety and efficiency." Furthermore, the cloud-based decision-making system automates parameter calculation and optimization, enabling deployment via AGVs without human intervention, significantly improving operational efficiency and avoiding safety risks associated with personnel working in traffic flow, making it particularly suitable for areas with high traffic density. Moreover, the two-layer decision-making logic of "basic parameters + model correction" forms a reusable and iterative intelligent mechanism. As data accumulates, the predictive model can continuously optimize the accuracy of weight calculations, making the deployment strategy constantly approach the optimal solution, possessing adaptability and evolutionary capabilities, and improving the reliability and economy of the protection system in the long term.

[0066] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0067] While specific embodiments of the invention have been described in detail by way of examples, those skilled in the art should understand that the examples are for illustrative purposes only and are not intended to limit the scope of the invention. Those skilled in the art should also understand that various modifications can be made to the embodiments without departing from the scope and spirit of the invention.

Claims

1. A dynamic deployment and retrieval system for intelligent reflective cones based on AGVs, characterized in that, The system includes: an AGV transportation platform and a cloud-based decision-making system; the cloud-based decision-making system is used to perform the following steps: S100, obtain the width d of the target road segment and the road speed limit v. max and real-time traffic flow ρ; S200, according to d, v max Based on ρ and the preset mapping table QT for the base layout angle and spacing, determine d and v. max The combination of ρ corresponds to the basic layout angle θ and the layout spacing Δd; where QT includes several rows, each including a set of road width, road speed limit and traffic flow, as well as the corresponding basic layout angle and layout spacing; the basic layout angle is the angle between the straight line formed by the preset number of reflective cones and the center line of the target road segment; S300 generates a comprehensive feature vector XL based on the static parameters, dynamic parameters, and environmental parameters of the target road segment; S400, input XL into the preset layout angle weight and layout spacing weight prediction model to obtain the layout angle weight λ1 corresponding to θ and the layout spacing weight λ2 corresponding to Δd; S500, based on θ and λ1, determine the target placement angle θ'=λ1×θ, and based on Δd and λ2, determine the target placement spacing Δd'=λ2×Δd; S600 sends θ' and Δd' to the AGV transport platform so that the reflector stack can be deployed according to θ' and Δd' via the AGV transport platform.

2. The AGV-based intelligent reflective cone dynamic deployment and retrieval system according to claim 1, characterized in that, Following step S600, the method further includes the following steps: S700 obtains the number NUM1 of reflective cones that were hit within a preset historical time period by using different types of sensors preset on the reflective cones; S710, if NUM1=0, then do not adjust θ' and Δd'; otherwise, proceed to S720; S720, adjust θ' and Δd' based on NUM1 and the total number of reflective cones NUM deployed.

3. The AGV-based intelligent reflective cone dynamic deployment and retrieval system according to claim 2, characterized in that, Step S720 includes the following steps: S721, based on NUM1 and NUM, determine the impact rate of the reflective cone τ = NUM1 / NUM; S722, Based on τ and the preset dynamic adjustment coefficient mapping table QR, determine the dynamic adjustment coefficient γ1 of θ' and the dynamic adjustment coefficient γ2 corresponding to Δd'; S723, adjust θ' to θ'×γ1 and Δd' to Δd'×γ2.

4. The intelligent reflective cone dynamic deployment and retrieval system based on AGV according to claim 1, characterized in that, The AGV transportation platform is equipped with a 16-line lidar and a laser SLAM module, which are used to build point cloud maps of the construction area in real time and enable autonomous navigation.

5. The intelligent reflective cone dynamic deployment and retrieval system based on AGV according to claim 1, characterized in that, The AGV transport platform includes a six-degree-of-freedom collaborative robotic arm. The end effector of the six-degree-of-freedom collaborative robotic arm integrates a three-finger flexible gripper, a vacuum suction cup, and an air jet cleaning device. The wrist is equipped with an RGB-D camera to realize the pose recognition of the reflective cone.

6. The intelligent reflective cone dynamic deployment and retrieval system based on AGV according to claim 1, characterized in that, The reflective cone is a deformable intelligent reflective cone with a three-stage telescopic structure; it has a built-in counterweight self-balancing chassis, a multispectral warning light at the top of the cone, a cone pressure sensor, and a 5G-V2X communication module; the bottom of the cone is equipped with a ring-shaped electromagnet array to achieve magnetic docking with the robotic arm; and the surface of the cone is coated with a thermochromic self-healing coating.

7. The intelligent reflective cone dynamic deployment and retrieval system based on AGV according to claim 1, characterized in that, Static parameters include: effective road width and radius of curvature; Dynamic parameters include: real-time traffic speed, traffic density, and historical accident data in construction areas; Environmental parameters include: weather conditions, road conditions, and current time.

Images