Highway 3D digital paving construction process in karst landform region

Abstract

The present application relates to the technical field of road engineering construction, and more particularly to a highway 3D digital paving construction process in karst topography, which comprises six major procedures of modeling partitioning, positioning calibration, double-source verification, self-adaptive paving, rolling coordination, and acceptance closed loop. First, multiple types of data are integrated to build a four-dimensional digital twin model and divide the settlement risk area. Then, the roadbed coordinate calibration is completed relying on multi-source positioning. During construction, radar and point cloud are used for double-source geological verification, and the paving and rolling parameters are adjusted. After completion, the pavement is scanned, and the whole-process data is archived to generate a digital completion file. In the present application, the double-source collaborative verification can real-time review the underground karst conditions and dynamically update the construction parameters. Combined with the mountain anti-shielding positioning and whole-process linkage control, a complete digital management and control closed loop is formed, which makes up for the defects of traditional static surveying and construction that cannot adapt to the changes in the field geology, and improves the geological adaptation capability and forming quality of the karst road section paving construction.

Description

Technical Field

[0001] This invention relates to the field of road engineering construction technology, and in particular to a 3D digital paving construction process for highways in karst landform areas. Background Technology

[0002] In highway pavement construction, 3D digital paving technology can automate the control of paving elevation and alignment by relying on digital models and positioning systems, effectively reducing the workload of manual layout and improving the stability of construction accuracy. Given the complex geological conditions and uneven distribution of karst in karst areas, current construction typically begins with a full-line geological survey to identify the distribution and scale of unfavorable karst geological bodies. Based on this, different levels of settlement risk sections are delineated, and corresponding differentiated paving process schemes are matched to reduce the pavement quality risks caused by uneven foundations.

[0003] However, the results of the preliminary geological survey are mostly static conclusions. Due to the limitations of detection accuracy and on-site conditions, some shallow small karst anomalies are easily missed. Furthermore, there is a lack of effective geological verification and dynamic parameter adjustment mechanisms during construction. Risk zoning and process parameters are difficult to modify according to the actual geological conditions on site, resulting in insufficient adaptability of geological risk management and difficulty in fully adapting to the complex geological conditions of karst areas. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a 3D digital paving construction process for highways in karst landform areas. It aims to improve the existing paving construction in karst landform areas, which relies on static geological surveys to delineate risk sections, lacks geological verification methods and dynamic adjustment mechanisms for process parameters, and has insufficient adaptability to geological risk management.

[0005] This invention provides the following technical solution: a 3D digital paving construction process for highways in karst landform areas, comprising the following steps: S1. Modeling and Zoning: Integrate engineering geological survey data, karst detection data and road design data to construct a four-dimensional digital twin model of the road. Divide the route into multiple levels of settlement risk zones and match the corresponding paving process parameters and mixture formula parameters. The model has a built-in karst risk feature library and dynamic iteration interface. S2. Positioning calibration: A relay-type GNSS reference station is set up along the construction line. The paver is equipped with a multi-source fusion positioning unit. Before paving, the global verification of the roadbed morphology and unified spatial coordinate mapping are completed. S3. Dual-source verification: During the paving operation, the geological detection unit at the front of the paver collects the radar echo of the roadbed ahead, matches it with the karst risk feature database to verify the boundary of the risk zone, and simultaneously uses laser point cloud to invert the actual stiffness of the roadbed. After the two types of perception results are verified together, the settlement risk zone and corresponding parameters are dynamically updated and sent back to the four-dimensional road digital twin model. S4, Adaptive Paving: Calls real-time updated risk zone parameters, automatically adjusts the screed reference elevation and elevation angle, synchronously links the mixing plant to adjust the dosage of modified components in the mixture, and enables dual-dimensional attitude compensation of cross slope and longitudinal slope in complex terrain sections. S5. Compaction Coordination: The compaction equipment is connected to a unified data control system, which matches the compaction operation parameters with the current road section risk level, subgrade stiffness and mixture temperature. In high settlement risk sections, the excitation parameters and compaction sequence are automatically adjusted. S6. Acceptance closed loop: After paving is completed, a vehicle-mounted 3D laser scanning is used to generate an actual point cloud model of the road surface. This model is then compared and analyzed with a 4D road digital twin model to repair abnormal areas. All data throughout the process is linked to the 4D road digital twin model to construct a digital as-built archive for the entire life cycle.

[0006] By adopting the above technical solutions, this invention addresses the challenges of complex geology and terrain constraints in highway paving in karst landform areas by constructing a full-process digital management and control system. Through dual-source collaborative verification of geological exploration and stiffness inversion, settlement risk zones and supporting process parameters can be dynamically updated during construction, improving the adaptability and accuracy of geological risk management. Combined with multi-source fusion positioning and point cloud blind spot compensation mechanisms, positioning accuracy is ensured in signal obstruction scenarios, supporting precise, pileless operations. By integrating risk zone parameters, adaptive adjustment of paving posture and collaborative management of compaction parameters are achieved, balancing forming accuracy and operational safety in karst subgrades. A digital twin model linking the entire process data forms an as-built archive, enabling data integration between construction and maintenance stages and providing data support for subsequent road maintenance.

[0007] Preferably, in step S1, the four-dimensional road digital twin model includes three-dimensional spatial geometry and settlement prediction time dimension; the dynamic iteration interface supports real-time transmission of construction data and synchronous update of parameters; the multi-level settlement risk zone includes at least a normal zone, a potential risk zone and a significant risk zone, and each zone is set with a differentiated paving thickness redundancy coefficient, a mixture modification scheme and a compaction control standard, and the paving thickness redundancy coefficient is determined based on the settlement prediction amplitude of the corresponding zone.

[0008] Preferably, in step S2, the multi-source fusion positioning unit integrates a GNSS receiver, an inertial measurement unit, and a multi-line laser scanner; the global verification of the roadbed topography is based on a four-dimensional road digital twin model; when the GNSS signal is lost for more than a set time, the system automatically switches to a positioning mode combining laser point cloud registration and inertial measurement unit, and corrects the cumulative positioning error of the inertial measurement unit through the inherent topographic features of the roadbed.

[0009] Preferably, in step S3, the geological detection unit is a small three-dimensional ground-penetrating radar that moves synchronously with the paver and performs real-time scanning and acquisition of the shallow underground layers of the road section to be paved within a set range ahead; the karst risk feature database stores radar echo characteristic parameters of typical karst geological bodies as a benchmark for echo matching.

[0010] Preferably, in step S3, the actual stiffness of the roadbed is obtained by inversion based on the elastic foundation beam model combined with the micro-deformation parameters of the roadbed surface; the two types of perception results are verified in a coordinated manner, with the abnormal geological bodies detected and identified by radar as the trigger for adjusting the risk level, and the roadbed stiffness inversion results as the basis for determining the risk level. When there is a conflict between the two types of perception results, the final determination is made according to the principle of the higher risk level.

[0011] Preferably, in step S4, the reference elevation of the screed plate is determined based on the thickness redundancy coefficient and the predicted settlement amplitude of the corresponding risk zone, and a thickness margin is actively reserved in the high settlement risk section.

[0012] Preferably, in step S4, the dual-dimensional attitude compensation of cross slope and longitudinal slope is achieved by collecting paver attitude data in real time through an inertial measurement unit and combining it with the design alignment parameters in the four-dimensional road digital twin model to correct the height of both sides of the screed and the front and rear tilt angles, respectively.

[0013] Preferably, in step S5, the vibration parameter adjustment of the high settlement risk section is to reduce the vibration amplitude and optimize the rolling sequence, and match the paver travel speed to ensure the temperature window for compaction operation.

[0014] Preferably, in step S4, during the adaptive paving process, the morphology of the formed road surface is detected in real time by the rear laser scanning unit at the tail of the paver, and the thickness and smoothness deviations are calculated. When the deviation exceeds the warning threshold, the screed parameters are adjusted immediately.

[0015] Preferably, in step S6, the digital as-built archive has a built-in settlement early warning association interface, which can be directly connected to the road maintenance management system to provide data support for settlement monitoring and preventive maintenance in karst sections during the operation period.

[0016] The present invention has the following beneficial effects: 1. In this invention, a dual-source collaborative verification mechanism is constructed by forward geological exploration and roadbed stiffness inversion. This mechanism can verify the boundary of karst risk zone and supplement the identification of unlabeled shallow abnormal geological bodies during paving operations. Combined with established judgment rules, the settlement risk zoning and supporting process parameters are dynamically updated to improve the adaptability and accuracy of geological risk management and improve the control limitations of traditional static exploration and construction.

[0017] 2. In this invention, a relay-type GNSS reference station is used in combination with a multi-source fusion positioning unit. Combined with the blind spot filling mechanism of roadbed point cloud registration, the positioning accuracy can be maintained stably in the case of signal obstruction in karst mountainous areas. Global topography verification and coordinate mapping are completed before paving, realizing pileless precision operation and improving the construction adaptability in complex terrain.

[0018] 3. In this invention, the benchmark elevation of the screed is adjusted based on dynamically updated risk zoning parameters, and the terrain attitude deviation is eliminated by combining cross slope and longitudinal slope dual-dimensional attitude compensation. With the real-time detection and feedback of the morphology during the paving process, the paving thickness and the forming accuracy of the line can be improved, and the construction quality requirements of different settlement risk sections can be adapted.

[0019] 4. In this invention, the compaction equipment is incorporated into a unified management and control system. The compaction operation parameters are matched with the road section risk level, subgrade stiffness and mixture temperature. In high-risk sections, the vibration amplitude is reduced and the compaction sequence is optimized. This can reduce the additional disturbance of strong vibration operation to karst subgrade while ensuring compaction quality.

[0020] 5. In this invention, the entire process of construction and testing data is linked to a four-dimensional digital twin model of the road to build a digital as-built archive for the entire life cycle. An interface for settlement early warning is reserved, which can realize the data connection between the construction and operation and maintenance stages and provide data support for the later road maintenance and management. Attached Figure Description

[0021] Figure 1 This is a flowchart of a method for 3D digital paving construction of highways in karst landform areas proposed in this invention. Figure 2 This is a framework diagram of a 3D digital paving construction process for highways in karst landform areas proposed in this invention. Figure 3 This is a flowchart of the dual-source verification and dynamic iteration proposed in this invention; Figure 4 This is the adaptive paving control logic diagram proposed in this invention. Detailed Implementation

[0022] The technical solutions in 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.

[0023] In an embodiment of the present invention, the present invention provides a 3D digital paving construction process for highways in karst landform areas, such as... Figure 1 As shown, it includes the following steps: S1. Modeling and Zoning: Integrate engineering geological survey data, karst detection data and road design data to construct a four-dimensional digital twin model of the road. Divide the route into multiple levels of settlement risk zones and match the corresponding paving process parameters and mixture formula parameters. The model has a built-in karst risk feature library and dynamic iteration interface. Furthermore, the four-dimensional road digital twin model includes three-dimensional spatial geometry and settlement prediction time dimension; the dynamic iteration interface supports real-time transmission of construction data and synchronous update of parameters; the multi-level settlement risk zone includes at least normal zone, potential risk zone and significant risk zone, and each zone is set with differentiated paving thickness redundancy coefficient, mixture modification scheme and compaction control standard, and the paving thickness redundancy coefficient is determined based on the settlement prediction amplitude of the corresponding zone.

[0024] Specifically, engineering geological survey data were obtained from the test results of geotechnical mechanical parameters of boreholes along the route, karst detection data were obtained from the full-line scanning and interpretation results of vehicle-mounted 3D ground-penetrating radar, and road design data included horizontal, vertical, and cross-sectional alignments and pavement structure layer design parameters. A 3D engineering modeling platform was used to complete the spatial coordinate registration of multi-source data and construct a four-dimensional digital twin model of the road. The three-dimensional spatial dimension corresponds to the topography, structural layers, and underground karst geometry of the entire road, while the temporal dimension corresponds to the evolution process of medium- and long-term subgrade settlement prediction. The model has a built-in karst risk feature library, storing three types of radar echo characteristic parameters: reflected wave amplitude, two-way travel time, and phase reversal of typical karst adverse geological bodies. It also provides an open dynamic iteration interface to receive real-time feedback data during the construction phase and update model attributes synchronously. Based on the burial depth and size of karst cavities and the initial resilient modulus of the subgrade, combined with settlement prediction results, the entire route is divided into three levels of settlement risk zones. Corresponding paving process parameter sets and mixture formula parameter sets are matched for each zone, and the paving thickness redundancy coefficient is set positively correlated with the settlement prediction amplitude of each zone. Taking the K12+300~K12+800 section of a karst mountain expressway as an example, this section has many shallow dissolution fissures and small caves. It was initially divided into two categories: potential risk area and significant risk area, with different combinations of pre-set process parameters.

[0025] S2. Positioning calibration: A relay-type GNSS reference station is set up along the construction line. The paver is equipped with a multi-source fusion positioning unit. Before paving, the global verification of the roadbed morphology and unified spatial coordinate mapping are completed. Furthermore, the multi-source fusion positioning unit integrates a GNSS receiver, an inertial measurement unit, and a multi-line laser scanner; the global verification of the roadbed topography is based on a four-dimensional digital twin model of the road; when the GNSS signal is lost for more than a set time, the system automatically switches to a positioning mode that combines laser point cloud registration with the inertial measurement unit, and corrects the cumulative positioning error of the inertial measurement unit through the inherent topographic features of the roadbed.

[0026] Specifically, mobile GNSS reference stations are deployed at fixed intervals along the construction route. These stations are uniformly connected to the project's independent construction coordinate system and broadcast RTK differential correction data via data transmission radios, forming a relay-style differential signal coverage link. The multi-source fusion positioning unit on the paver uses an onboard industrial control computer as its computing core, simultaneously collecting coordinate data from the GNSS receiver, angular velocity and acceleration data from the inertial measurement unit, and roadbed surface point cloud data from a multi-line laser scanner. An extended Kalman filter algorithm is used to complete the multi-source data fusion, outputting the paver's real-time spatial coordinates and three-dimensional attitude parameters. The state update process is calculated based on the following formula: X k =F k,k-1 X k-1 +W k-1 ; In the formula, X k Let F be the system state vector at time k, containing the paver's three-dimensional coordinates, velocity, attitude angles, and inertial measurement unit zero bias, recursively derived from the state at the previous time step; k,k-1 W is the state transition matrix from time k-1 to time k, derived from the vehicle kinematics model; k-1 The system process noise vector follows a zero-mean Gaussian distribution. The output positioning results are used to match the risk zoning parameters of the current road segment and to provide the initial pose for laser point cloud registration. Before the paving operation starts, the paver travels at low speed along the road segment to be paved for one leg. The airborne laser scanner collects the full-width roadbed surface point cloud, and global registration is performed using the iterative nearest point algorithm and the roadbed design point cloud in the digital twin model to complete spatial coordinate mapping and system self-calibration without the need for manual placement of stakeouts. When the GNSS signal is lost due to mountain obstruction for more than a set time, the system automatically switches to a point cloud registration and inertial measurement combined mode. The accumulated error of the inertial measurement unit is corrected by the inherent morphological characteristics of the roadbed to maintain stable positioning accuracy.

[0027] S3. Dual-source verification: During the paving operation, the geological detection unit at the front of the paver collects the radar echo of the roadbed ahead, matches it with the karst risk feature database to verify the boundary of the risk zone, and simultaneously uses laser point cloud to invert the actual stiffness of the roadbed. After the two types of perception results are verified together, the settlement risk zone and corresponding parameters are dynamically updated and sent back to the four-dimensional road digital twin model. Furthermore, the geological exploration unit is a small three-dimensional ground-penetrating radar that moves synchronously with the paver to scan and collect data on the shallow underground layers of the road section to be paved within a set range in front; the karst risk feature database stores the radar echo characteristic parameters of typical karst geological bodies as a benchmark for echo matching. Furthermore, the actual stiffness of the roadbed is obtained by inversion based on the elastic foundation beam model combined with the micro-deformation parameters of the roadbed surface; the two types of perception results are verified together, with the abnormal geological bodies detected and identified by radar as the trigger for adjusting the risk level, and the roadbed stiffness inversion results as the basis for judging the risk level. When the two types of perception results conflict, the final judgment is made according to the principle of the higher risk level.

[0028] Specifically, during the paving operation, a small 3D ground-penetrating radar at the front of the paver continuously collects shallow ground-level radar echo data of the road section to be paved. The control system matches and compares the amplitude, travel time, and phase characteristics of the real-time echoes with preset parameters in the karst risk characteristic database. When the matching degree reaches a preset threshold, the actual boundary and development degree of the preset risk zone are verified and confirmed. At the same time, shallow anomalous geological bodies that were not marked in the previous survey are identified and their spatial locations are marked. Simultaneously, the roadbed surface point cloud data collected by the airborne laser scanner is reused to extract the roadbed surface micro-variables. Combined with the roadbed structural layer thickness parameters obtained from the previous detection, the actual resilient modulus of the roadbed in the current working section is calculated based on the elastic foundation beam model. The calculation formula is as follows: ; In the formula, E0 is the roadbed resilient modulus, and the calculation result is used to correct the settlement risk level of the corresponding road section; D is the equivalent calculated diameter of the roadbed, determined by the load action range of the laser scan; μ is the Poisson's ratio of the roadbed soil, taken from the previous geological survey test data; l is the radius of action of the flexible load, calculated from the grounding size of the paver; p is the equivalent uniformly distributed load, calculated from the paver's self-weight and working load; w is the roadbed surface deformation, extracted from the laser point cloud data. The radar detection results and stiffness inversion results are jointly verified. The abnormal geological bodies identified by the radar are used as the trigger for adjusting the risk level, and the stiffness inversion results are used as the judgment basis for adjusting the risk level. When the two types of results conflict, the higher risk level is used to complete the final judgment. After the verification is completed, the settlement risk level and supporting process parameters of the corresponding road section are dynamically updated, and the updated data is synchronized to the four-dimensional road digital twin model through the dynamic iteration interface. Taking K12+520 in the aforementioned section as an example, shallow small karst caves that were not marked in the early survey were identified by forward radar. The synchronous stiffness inversion showed that the resilient modulus of the roadbed in this section was lower than the initial predicted value. After collaborative verification, the risk level of this local road section was raised, and the corresponding process parameters were adjusted accordingly.

[0029] S4, Adaptive Paving: Calls real-time updated risk zone parameters, automatically adjusts the screed reference elevation and elevation angle, synchronously links the mixing plant to adjust the dosage of modified components in the mixture, and enables dual-dimensional attitude compensation of cross slope and longitudinal slope in complex terrain sections. Furthermore, the reference elevation of the screed is determined based on the thickness redundancy coefficient and the predicted settlement amplitude of the corresponding risk zone, and a thickness margin is proactively reserved in high settlement risk sections. Furthermore, the cross slope and longitudinal slope dual-dimensional attitude compensation is achieved by collecting paver attitude data in real time through inertial measurement units and combining it with the design alignment parameters in the four-dimensional road digital twin model to correct the height of the screed on both sides and the front and rear tilt angles respectively. Furthermore, during the adaptive paving process, the rear laser scanning unit at the rear of the paver is used to detect the shape of the formed road surface in real time, calculate the thickness and smoothness deviation, and provide immediate feedback to adjust the screed parameters when the deviation exceeds the warning threshold.

[0030] Specifically, the paver control system retrieves the risk zoning parameters of the current working section in real time from the four-dimensional road digital twin model. Based on the thickness redundancy coefficient and settlement prediction amplitude of the corresponding zone, it calculates the real-time benchmark elevation of the screed. The stroke of the screed lifting cylinder is adjusted via an electro-hydraulic proportional valve to proactively reserve thickness margin in high settlement risk areas. The benchmark elevation calculation logic is as follows: H real =H des +α·S; In the formula, H real The calculation results for the actual control elevation of the ironing board are directly output to the hydraulic servo system; H des The road surface design elevation for the corresponding location is retrieved from the digital twin model; α is the thickness redundancy coefficient for the corresponding risk zone; and S is the predicted settlement amplitude for the corresponding location. Simultaneously, combining vehicle attitude data collected by the inertial measurement unit with the design alignment parameters within the model, dual-dimensional attitude compensation for both cross and longitudinal slopes is completed: in the cross slope direction, precise shaping of the designed cross slope is achieved by adjusting the difference in hydraulic cylinders on both sides of the screed; in the longitudinal slope direction, thickness deviation caused by vehicle pitch is eliminated by adjusting the front and rear tilt angles of the screed. The mixing plant control system is linked to the digital twin model via a dedicated data link, dynamically adjusting the blending ratio of crack-resistant modified components and fibers in the mixture based on the real-time risk level of the current road section. During paving, the rear-mounted laser scanning unit at the rear of the paver continuously scans the newly formed road surface, calculating the paving thickness and smoothness indicators at fixed step lengths. When the deviation exceeds the preset warning threshold, it immediately feeds back to the screed control system, fine-tuning the screed parameters for the corresponding location in the next work cycle.

[0031] S5. Compaction Coordination: The compaction equipment is connected to a unified data control system, which matches the compaction operation parameters with the current road section risk level, subgrade stiffness and mixture temperature. In high settlement risk sections, the excitation parameters and compaction sequence are automatically adjusted. Furthermore, the vibration parameters in high settlement risk areas are adjusted to reduce vibration amplitude and optimize rolling sequence, while matching the paver travel speed to ensure the temperature window for compaction operations.

[0032] Specifically, both the double-drum roller and the pneumatic tire roller are equipped with GNSS positioning modules and infrared temperature sensors, and are uniformly connected to the construction management and control data system. The system obtains the settlement risk level and subgrade stiffness parameters of the current working section from the four-dimensional digital twin model of the road, combines it with the surface temperature of the mixture collected by infrared sensors on site, and synchronously matches the paver's travel speed to generate corresponding rolling follow-up distance, number of rolling passes, vibration amplitude, and frequency parameters, which are then sent to the rolling equipment. For high settlement risk sections, the system automatically reduces the vibration amplitude of the rolling equipment, increases the proportion of static compaction, and adjusts the sequence of initial compaction, intermediate compaction, and final compaction to avoid strong vibration causing additional disturbance to the karst subgrade; at the same time, it matches the paving operation travel speed and controls the rolling follow-up distance to ensure that the compaction operation is completed within the suitable temperature range of the mixture.

[0033] S6. Acceptance closed loop: After paving is completed, a vehicle-mounted 3D laser scanning is used to generate an actual point cloud model of the road surface. This model is then compared and analyzed with a 4D road digital twin model to repair abnormal areas. All data throughout the process is linked to the 4D road digital twin model to construct a digital as-built archive for the entire life cycle.

[0034] Furthermore, the digital as-built archives have a built-in settlement early warning interface, which can be directly connected to the road maintenance management system to provide data support for settlement monitoring and preventive maintenance in karst sections during the operation period.

[0035] Specifically, after the paving work is completed and the road surface temperature drops to ambient temperature, a vehicle-mounted 3D laser scanning system travels at a constant speed along the route to collect high-density point cloud data of the entire road surface, generating a point cloud model of the actual road surface morphology. This actual point cloud model is then precisely registered and compared with the design road surface model in the 4D road digital twin model. Deviations in paving thickness, cross slope, and smoothness across the entire road section are calculated. Areas with deviations exceeding limits are located and targeted milling repairs are performed. After repair, the model is scanned again for verification until design requirements are met. Geological survey data, subgrade stiffness inversion data, paving process parameters, compaction operation data, final quality acceptance data, and repair records from the entire construction process are integrated and linked to the corresponding spatial locations in the 4D road digital twin model, constructing a digital as-built archive covering the entire lifecycle. The archive includes a built-in settlement early warning interface, which can directly connect to the road maintenance management system, providing basic data support for settlement monitoring and preventative maintenance in karst sections during the operational phase.

[0036] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A 3D digital paving construction technology for highways in karst landform areas, characterized in that: Includes the following steps: S1. Modeling and Zoning: Integrate engineering geological survey data, karst detection data and road design data to construct a four-dimensional digital twin model of the road. Divide the route into multiple levels of settlement risk zones and match the corresponding paving process parameters and mixture formula parameters. The model has a built-in karst risk feature library and dynamic iteration interface. S2. Positioning calibration: A relay-type GNSS reference station is set up along the construction line. The paver is equipped with a multi-source fusion positioning unit. Before paving, the global verification of the roadbed morphology and unified spatial coordinate mapping are completed. S3. Dual-source verification: During the paving operation, the geological detection unit at the front of the paver collects the radar echo of the roadbed ahead, matches it with the karst risk feature database to verify the boundary of the risk zone, and simultaneously uses laser point cloud to invert the actual stiffness of the roadbed. After the two types of perception results are verified together, the settlement risk zone and corresponding parameters are dynamically updated and sent back to the four-dimensional road digital twin model. S4, Adaptive Paving: Calls real-time updated risk zone parameters, automatically adjusts the screed reference elevation and elevation angle, synchronously links the mixing plant to adjust the dosage of modified components in the mixture, and enables dual-dimensional attitude compensation of cross slope and longitudinal slope in complex terrain sections. S5. Compaction Coordination: The compaction equipment is connected to a unified data control system, which matches the compaction operation parameters with the current road section risk level, subgrade stiffness and mixture temperature. In high settlement risk sections, the excitation parameters and compaction sequence are automatically adjusted. S6. Acceptance closed loop: After paving is completed, a vehicle-mounted 3D laser scanning is used to generate an actual point cloud model of the road surface. This model is then compared and analyzed with a 4D road digital twin model to repair abnormal areas. All data throughout the process is linked to the 4D road digital twin model to construct a digital as-built archive for the entire life cycle.

2. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S1, the four-dimensional road digital twin model includes three-dimensional spatial geometry and settlement prediction time dimension; the dynamic iteration interface supports real-time transmission of construction data and synchronous update of parameters; the multi-level settlement risk zone includes at least a normal zone, a potential risk zone and a significant risk zone, and each zone is set with a differentiated paving thickness redundancy coefficient, a mixture modification scheme and a compaction control standard, and the paving thickness redundancy coefficient is determined based on the settlement prediction amplitude of the corresponding zone.

3. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S2, the multi-source fusion positioning unit integrates a GNSS receiver, an inertial measurement unit, and a multi-line laser scanner; the global verification of the roadbed topography is based on a four-dimensional road digital twin model; when the GNSS signal is lost for more than a set time, the system automatically switches to a positioning mode combining laser point cloud registration and inertial measurement unit, and corrects the cumulative positioning error of the inertial measurement unit through the inherent topographic features of the roadbed.

4. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S3, the geological detection unit is a small three-dimensional ground-penetrating radar that moves synchronously with the paver and performs real-time scanning and acquisition of the shallow underground layers of the road section to be paved within a set range ahead; the karst risk feature database stores radar echo characteristic parameters of typical karst geological bodies as a benchmark for echo matching.

5. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S3, the actual stiffness of the roadbed is obtained by inversion based on the elastic foundation beam model combined with the micro-deformation parameters of the roadbed surface; the two types of perception results are jointly verified, with the abnormal geological bodies detected and identified by radar as the trigger for adjusting the risk level, and the roadbed stiffness inversion results as the basis for judging the risk level. When there is a conflict between the two types of perception results, the final judgment is made according to the principle of the higher risk level.

6. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S4, the reference elevation of the screed plate is determined based on the thickness redundancy coefficient and the predicted settlement amplitude of the corresponding risk zone, and a thickness margin is actively reserved in the high settlement risk section.

7. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S4, the dual-dimensional attitude compensation of cross slope and longitudinal slope is achieved by collecting paver attitude data in real time through an inertial measurement unit and combining it with the design alignment parameters in the four-dimensional road digital twin model to correct the height of both sides of the screed and the front and rear tilt angles respectively.

8. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S5, the vibration parameters of the high settlement risk section are adjusted to reduce the vibration amplitude and optimize the rolling sequence, and to match the paver's walking speed to ensure the temperature window for compaction.

9. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S4, during the adaptive paving process, the morphology of the formed road surface is detected in real time by the rear laser scanning unit at the tail of the paver, and the thickness and smoothness deviations are calculated. When the deviation exceeds the warning threshold, the parameters of the screed are adjusted immediately.

10. The 3D digital paving construction technology for highways in karst landform areas according to claim 1, characterized in that: In step S6, the digital as-built archive has a built-in settlement early warning association interface, which can be directly connected to the road maintenance management system to provide data support for settlement monitoring and preventive maintenance in karst sections during the operation period.

Images