A compaction method, a road roller, and a storage medium for a narrow or corner area

Abstract

The present application relates to the technical field of road roller, and particularly relates to a compaction method for narrow or corner areas, a road roller and a storage medium. The compaction method comprises the following steps: receiving a compaction quality heat map operation instruction; pointing a vibration head to a target area; driving a robot end to sweep along a preset path, and recording touch point coordinates when a contact force exceeds a set threshold to obtain a discrete boundary point set; fitting a physical boundary curve according to the discrete boundary point set and expanding a safety distance outward to establish a feasible area boundary constraint; generating an initial guide track located in the feasible area; calculating a position correction amount based on an impedance control law; comparing the corrected expected position with the boundary constraint; if the comparison result satisfies the condition, taking the position as a target position; otherwise, correcting to the boundary and adjusting the impedance parameter; and driving the vibration head to perform compaction operation in the narrow or corner area. Through the compaction method, the robot can realize the compaction operation in the complex corner area without dead angle and low collision risk.

Description

Technical Field

[0001] This invention relates to the field of road roller technology, and more particularly to a compaction method, road roller, and storage medium for narrow or corner areas. Background Technology

[0002] In road, bridge, and various infrastructure construction projects, vibratory rollers are the primary equipment for large-area material compaction. Their working principle utilizes a vibrating steel wheel to generate high-frequency impact force, causing the compacted material particles to rearrange and compact, thereby achieving the specified load-bearing strength. However, in actual construction sites, besides large open areas, there are also narrow or obstructed corner areas such as behind bridge abutments, near curbs, in backfilled trenches, and at building corners. Due to space limitations, the steel wheels and body of traditional rollers cannot easily enter or maneuver in these areas, thus becoming blind spots for compaction operations.

[0003] To address this issue, an attempt was made to install a small robot on the body of a conventional road roller. This robot carries a vibratory compaction device to supplement the compaction of corner areas. The process is roughly as follows: the operator or the host system pre-plans a robotic arm trajectory that avoids known obstacles based on the ideal geometry of the work area. Then, under open-loop or semi-closed-loop control, the robot drives the vibratory head along this pre-planned trajectory to compact the designated area. The pre-planned trajectory is typically a sequence of path points generated offline based on an idealized model of the work area (such as a rectangle, arc, or other regular shapes).

[0004] Because the boundaries of corner areas at construction sites are often extremely irregular, such as the winding boundary between backfill areas and existing structures, and the preset trajectory is difficult to reflect these real and irregular physical boundaries in real time, the robot may cause leakage areas due to the trajectory deviating from the real boundaries during actual operation, or it may cause collisions due to the trajectory being too close to obstacles. In severe cases, this may damage equipment or completed structures. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of the embodiments of the present invention is to provide a compaction method for narrow or corner areas.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions: In a first aspect, embodiments of the present invention provide a compaction method for narrow or corner areas. A vibratory compaction robot is mounted on a road roller, with a vibratory head at the robot's end. The compaction method includes: receiving a compaction quality heatmap operation command; performing global positioning and attitude preparation to point the vibratory head towards the target area; driving the robot's end to sweep along a preset path when the vibratory head is not activated, recording the coordinates of the contact point when the contact force exceeds a set threshold to obtain a discrete boundary point set; fitting a physical boundary curve based on the discrete boundary point set and expanding outwards by a safe distance to establish a feasible area boundary constraint; generating an initial guide trajectory within the feasible area based on the compaction quality heatmap; activating the vibratory head; reading the end contact force; calculating the position correction based on an impedance control law; comparing the corrected desired position with the boundary constraint; if the constraint is satisfied, using it as the target position; otherwise, correcting to the boundary and adjusting the impedance parameters; and driving the vibratory head to perform compaction operations in narrow or corner areas.

[0007] Secondly, embodiments of the present invention also provide a road roller for the aforementioned compaction method for narrow or corner areas, comprising a road roller body and a boom-mounted vibratory compaction robot mounted thereon; the boom-mounted vibratory compaction robot includes a robot-road roller integrated connection structure, a multi-degree-of-freedom robot, a communication and data link device, and an adaptive impedance controller; the robot-road roller integrated connection structure includes a mounting base, a passive vibration damping base, an active inertial compensation platform, and a through mechanism, wherein the mounting base is fixed to the rear frame of the road roller, and the passive vibration damping base is disposed on the mounting base. Between the mounting base and the active inertial compensation platform, the active inertial compensation platform is set on the passive vibration damping base, and the through mechanism is integrated inside the robot base to achieve continuous rotation; the multi-degree-of-freedom robot includes a waist rotation joint, shoulder joint, elbow joint, wrist rotation joint, wrist pitch joint, wrist yaw joint, a vibrating head, and a six-dimensional force / torque sensor installed on the end flange; the communication and data link device includes a 4G / 5G industrial router and a vehicle-mounted Wi-Fi Bluetooth adapter; the adaptive impedance controller is installed in the main control cabinet of the robot base.

[0008] Thirdly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the compaction method for narrow or corner areas.

[0009] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages: The compaction method of this invention solves the problems of preset trajectories being unable to adapt to irregular corners, easy under-compaction, and collisions by first probing the boundary to establish a constraint model and then combining force feedback to correct the trajectory in real time. At the same time, since feasible area constraints based on real physical boundaries are established before compaction, and the position correction amount is continuously compared with the boundary constraints and forcibly corrected during the compaction process, the robot can achieve compaction operations with no blind spots and low collision risk in complex corner areas.

[0010] Advantages of additional aspects of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In addition, the dimensions or spacing between the components are exaggerated to show the position of each component, and the schematic diagrams are for illustrative purposes only.

[0012] Figure 1 This is a flowchart of the compaction method provided in an embodiment of the present invention; Figure 2 This is an overall schematic diagram of the road roller provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the arm-mounted vibration compaction robot provided in an embodiment of the present invention; In the diagram: 1. Mounting base; 2. Passive vibration damping base; 3. Active inertial compensation platform; 4. Robot base; 5. Vibration head; 6. Six-dimensional force / torque sensor; 7. Protective enclosure; 8. Main control cabinet; 9. Sealed electrical cabinet; 10. Waist rotation joint; 11. Shoulder joint; 12. Elbow joint; 13. Wrist rotation joint; 14. Wrist pitch joint; 15. Wrist yaw joint; Detailed Implementation To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the linguistic context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0013] Generally speaking, the terms "comprising" and "including" only indicate that the steps and elements are explicitly identified, and these steps and elements do not constitute an exclusive list. The method or apparatus may also include other steps or elements.

[0014] Example 1 like Figure 1 As shown, in a typical embodiment of this disclosure, a compaction method for narrow or corner areas is provided. The method is executed by a boom-mounted vibratory compaction robot mounted on a road roller, with a vibratory head at the robot's end. The vibratory head applies impact force to the ground to achieve material compaction.

[0015] Compaction methods include: S100 receives instructions for compaction quality heat mapping operations; The S200 performs global positioning and attitude preparation, pointing the vibrating head towards the target area; When the vibrating head is not activated, the S300 drives the robot end to sweep along a preset path. When the contact force exceeds the set threshold, the coordinates of the contact point are recorded to obtain a discrete boundary point set. S400 fits the physical boundary curve based on the discrete boundary point set and expands the safety distance outward to establish feasible region boundary constraints; The S500 generates an initial guide trajectory within the feasible area based on the compaction quality heat map, starts the vibrating head, reads the end contact force, calculates the position correction based on the impedance control law, compares the corrected desired position with the boundary constraints, and uses the position as the target position if it meets the requirements; otherwise, it corrects to the boundary and adjusts the impedance parameters to drive the vibrating head to perform compaction operations in narrow or corner areas.

[0016] A compaction quality heat map is a data file that graphically marks the expected compaction requirements or detected weak compaction areas in the work area. For example, it can be generated based on previous inspections or design drawings. The robot receives the instructions through a 4G / 5G network and parses them to know all the target points that need to be compacted and their priorities.

[0017] Global localization refers to the robot calculating its position relative to the vehicle's coordinate system using an internal encoder and a pre-set model of a road roller. Attitude preparation refers to the control system driving the robot's lumbar joints and arm joints to move the vibrating head roughly towards the target work area.

[0018] The sweeping operation, performed without the vibratory head activated, aims to detect the actual physical boundaries before compaction, thus preventing collisions caused by directly activating vibration. During sweeping, the robot's end effector moves at a set collision detection threshold. A six-dimensional force / torque sensor monitors the contact force in real time. Once the threshold is exceeded, the three-dimensional coordinates of the current end effector's center point are recorded as a boundary contact point. After recording, the robot retracts and offsets to continue sweeping, thereby obtaining a discrete set of boundary points.

[0019] A continuous spatial curve is generated as the physical boundary profile, and then the curve is expanded outward by a safe distance to obtain a feasible region.

[0020] The system plans an initial guide trajectory, such as a zigzag or spiral path, within a feasible area based on the compaction quality heatmap. After the vibrating head is activated, the adaptive impedance controller reads the contact force fed back by the six-dimensional force / torque sensor in real time at a set control cycle. The actual contact force is compared with the expected target contact force to obtain the force error. The position correction is calculated using the impedance control law and superimposed on the expected position of the initial guide trajectory to obtain the corrected expected position Xc. Then, it checks whether Xc satisfies the boundary constraints. If it does, Xc is sent as the target position to the underlying position controller; if it does not, the target position is forcibly corrected to the intersection point on the boundary, and the damping term coefficient in the impedance control law is increased to make the robot behave more rigidly when approaching the boundary. Finally, the inverse kinematics calculation is used to convert the results into angular displacement commands for each joint, driving the motors to move.

[0021] This method solves the problems of preset trajectories being unable to adapt to irregular corners, prone to under-compaction, and collisions by first probing the boundary to establish a constraint model and then combining force feedback to correct the trajectory in real time. At the same time, because feasible area constraints based on real physical boundaries are established before compaction, and the position correction amount is continuously compared with the boundary constraints and forcibly corrected during the compaction process, the robot can achieve compaction operations with no blind spots and low collision risk in complex corner areas.

[0022] In some further specific examples disclosed herein, the safety distance is equal to the sum of the maximum envelope radius of the vibrating head and the preset safety margin, ensuring that the robot will not collide even with minor control errors when operating close to the boundary. The cubic B-spline curve fitting algorithm has good local support; modifying a control point only affects the local shape of the curve, making it suitable for handling locally irregular boundaries, such as unevenness at corners or meandering boundaries in trench backfill areas. Compared to global fitting methods, B-splines can more realistically preserve the local details of the boundary, avoiding the loss of key contour information due to smoothing, thereby improving the accuracy of the boundary constraint model.

[0023] In other specific examples disclosed herein, a gentle force is applied to the working surface to collect contact force signal characteristics. The system uses these characteristics to identify the physical properties of the working surface, including hardness, roughness, and flatness. For example, the force response curves produced by a hard, flat concrete surface are completely different from those of a soft, backfilled soil area. After identifying the properties, the adaptive impedance controller automatically matches and loads a set of corresponding optimal impedance parameters from its internal parameter library, including desired inertia parameters, damping parameters, and stiffness parameters. For example, for soft ground, low stiffness and high damping parameters are selected to make the boom behave more compliantly to conform to the ground undulations; for hard ground, high stiffness and low damping parameters are selected to maintain pressure stability.

[0024] Without accurate identification of contact surface properties, impedance parameters cannot be adjusted accordingly, leading to over-compression on soft surfaces or under-compression on hard surfaces. By pre-tuning, the controller can operate with appropriate flexibility and rigidity from the start of compaction, avoiding oscillation adjustments after startup.

[0025] In a further specific example of this disclosure, the impedance control law is used to calculate the position correction using the formula Md×ΔXdd+Bd×ΔXd+Kd×ΔX=ef, where ef is the force error, i.e., the difference between the expected contact force and the actual feedback force; Md, Bd, and Kd are the target inertia, damping, and stiffness matrices, respectively, all of which are diagonal matrices; ΔXdd is the corrected acceleration, ΔXd is the corrected velocity, and ΔX is the position correction.

[0026] This equation describes the dynamic relationship between the end-effector position correction and the force error. In practical control, the inertial term has a relatively small impact; adjustments are mainly made using damping and stiffness terms to achieve stable tracking of the contact force. The required position correction can be obtained by calculating the equation. The introduction of this formula allows the robot to simulate the behavior of a mass-spring-damped system when in contact with the ground, ensuring both accurate tracking of the desired force and adaptive adjustment of the arm's compliance based on changes in ground stiffness. Compared to simple PID force control, impedance control better handles dynamic changes in the contact environment, avoiding force overshoot or oscillations.

[0027] In other examples disclosed herein, during the sweeping process, after recording a boundary contact point, the control system controls the robot's end effector to retract a safe distance along its original path and slightly deviate from the path before continuing sweeping. The retraction prevents the end effector from continuing to compress obstacles after contact, thus avoiding damage; the path deviation ensures that subsequent sweeps do not repeatedly touch the same point, but rather gradually cover the entire boundary area. Through multiple sweeps, the system can obtain a series of discrete boundary contact points, forming a point set. This operation is closely integrated with the boundary detection step: without retraction and deviation, the sweeping process may fail to obtain a complete boundary contour due to repeated contact with the same point. The retraction distance and deviation can be dynamically adjusted according to the end effector size and detection accuracy.

[0028] In a further specific example of this disclosure, different regions in the compaction quality heatmap have different compaction requirement levels; for example, the area behind a bridge abutment typically requires higher compaction levels than areas farther from the structure. Prioritizing the planning and execution of guide trajectories for high-priority areas before processing low-priority areas ensures that critical weak areas are compacted first, preventing high-priority areas from being overlooked due to insufficient time or power. The heatmap not only provides the location information of target points but also provides priority weights, making trajectory planning no longer a simple geometric coverage but a task-oriented intelligent scheduling.

[0029] The specific steps of the compaction method are as follows: S100: System initialization and task instruction reception.

[0030] S110: After the system is powered on, a power-on self-test is performed. All joint modules of the robot, the adaptive impedance controller, and the passive vibration damping base and active inertial compensation platform in the integrated connection structure are initialized.

[0031] S111: The 4G / 5G industrial router and vehicle-mounted Wi-Fi / Bluetooth adapter are powered on to establish internal and external communication links with the remote control center and the main road roller. The network connection establishment time is less than 3 seconds.

[0032] S112: The robot receives work instructions from the main roller or control center via a 4G / 5G network. These instructions include a compaction quality heatmap. The compaction quality heatmap is a data file that graphically identifies the expected compaction requirements or detected weak compaction areas within the work area. The robot analyzes this data to determine all target points requiring compaction and their priorities.

[0033] S200: Robot global localization and attitude preparation.

[0034] S210: After the main roller travels to the vicinity of the first target work area, the boom-mounted vibratory compaction robot (hereinafter referred to as the robot) initiates global positioning. The robot uses its internal encoder and a pre-set roller body model to calculate its position relative to the vehicle body coordinate system.

[0035] S220: The robot begins posture preparation. Thanks to the through mechanism inside the base, the robot's swivel joint can drive the entire arm system to rotate continuously. The control system calculates an optimal starting posture based on the orientation of the target work point and drives the swivel joint and each arm joint to move to this posture. At this time, the vibrating head is roughly pointing towards the target work area, but has not yet made contact with any surface.

[0036] S300: Intelligent contact surface recognition and controller parameter self-tuning.

[0037] S310: When the vibrating head is guided above the work surface, the robot moves downward at an extremely slow speed, making it contact the work surface with a pre-set gentle force. The contact force threshold is 10N.

[0038] S311: A six-dimensional force / torque sensor installed at the end acquires force and torque signals in real time at the moment of contact. The control system analyzes the waveform, peak value, and settling time of these signals. For example, the force response curves of a hard, flat concrete surface are completely different from those of a soft soil backfill area. The system uses these characteristics to intelligently identify the physical properties of the current working surface, including hardness, roughness, and flatness.

[0039] S320: After identifying the contact surface properties, the adaptive impedance controller automatically matches and loads a set of optimal impedance parameters corresponding to that property from its internal parameter library. This parameter set includes the desired inertia parameter, damping parameter, and stiffness parameter. For example, for soft ground, the controller will select a set of low stiffness, high damping parameters to make the boom behave more compliantly and conform to the ground undulations; for hard ground, it will select high stiffness, low damping parameters to maintain pressure stability.

[0040] S400. Establish a boundary constraint model for a narrow space.

[0041] This step is used in the robot's control system to establish a calculable virtual boundary for the current working area, which includes a safety distance, as a rigid constraint for subsequent trajectory planning.

[0042] S410. Perform compliant probing and boundary point acquisition: Before the compaction operation begins, the robot does not immediately start the vibrating head, but first uses its end end as a compliant probe to perform boundary detection.

[0043] S411. Set pathfinding safety parameters. The adaptive impedance controller first adjusts the robot's end effector stiffness to a low level, for example, a translational stiffness of 500 N / m, and simultaneously sets a collision detection threshold Ft, for example, 15 N. This low stiffness setting ensures that the end effector does not generate a violent impact force when it comes into contact with an obstacle, thus providing a buffering protection.

[0044] S412. Perform a sweeping motion. The control system drives the robot end effector to carry the non-activated vibrating head and slowly sweep along a pre-set basic path slightly larger than the theoretical working area. The speed of this sweeping motion is controlled within 50 mm / s.

[0045] S413. Record the coordinates of the contact point. During the sweeping process, a six-dimensional force / torque sensor is installed between the robot's end effector flange and the vibrating head, continuously monitoring the contact force at a sampling rate of 1kHz. When the detected contact force Fc first exceeds the preset collision threshold Ft, i.e., Fc is greater than 10N, the system determines that the end effector has touched the physical boundary. At this time, the control system immediately records the encoder readings of each joint of the robot and calculates the three-dimensional spatial coordinates of the robot's end effector center point, i.e., the geometric center of the vibrating head, in the robot's base coordinate system using a forward kinematics model, denoted as a boundary contact point Pi=(xi, yi, zi). Subsequently, the system controls the robot's end effector to retreat a safe distance along the original path, for example, 20mm, and then slightly deviates from the path before continuing to sweep. Through multiple sweeps, the system can obtain a series of discrete boundary contact points, forming a point set {P1, P2, ..., P...} n}

[0046] S420. Generate a continuous boundary and apply a safe expansion.

[0047] S421. Fitting the physical boundary curve. After obtaining the discrete point set {Pi}, the system uses a cubic B-spline curve fitting algorithm to generate a smooth and continuous spatial curve B(u), which is the real physical boundary contour perceived by the robot. Utilizing the good local support of the B-spline curve, modifying a control point only affects the local shape of the curve, adapting to the handling of locally irregular boundaries.

[0048] S422. Establishing Constraint Inequalities. To ensure the robot does not collide with the physical boundary during operation, while maximizing its proximity to the boundary, the fitted physical boundary curve B(u) needs to be expanded outward by a safe distance ds. The value of ds consists of two parts: ds = re + dm. Where re is the maximum envelope radius of the vibrating head or end effector, for example, 75mm, and dm is a preset small safety margin, for example, 5mm. The expanded boundary Bs(u) can be obtained by shifting each point on the original curve outward by a distance ds along its normal direction. Finally, the boundary constraint of the confined space is expressed as the following mathematical inequality: The robot's tool center point coordinates Pt=(x, y, z) must always lie within the feasible region Ω enclosed by the bloated boundary Bs(u), i.e., Pt∈Ω. Equivalently, a signed distance function SDF(Pt) can be defined, representing the signed distance function from point Pt to the bloated boundary Bs(u), obtained analytically from the boundary curve or through discrete sampling. A positive value indicates that the point is within the feasible region Ω. When SDF(Pt) is greater than 0, it indicates that the point is within the feasible region; when SDF(Pt) is less than 0, it indicates that intrusion has occurred. The control objective is to ensure that SDF(Pt) is greater than or equal to 0 throughout the entire movement process.

[0049] S500. Real-time dynamic trajectory planning and tracking based on boundary constraints. After the boundary constraint model is established, the robot can begin the load compaction operation.

[0050] S510. Generate initial guide trajectory. Based on the priority indicated by the compaction quality heatmap and the established feasible region Ω, the system plans an initial guide trajectory located within Ω. This trajectory consists of a series of discrete path points. Guide trajectories for high-priority regions are planned and executed first, followed by lower-priority regions. Depending on the task requirements, the system typically uses a zigzag or spiral path.

[0051] S520. Initiate real-time force-controlled compaction cycle. The vibratory head starts, with an electric motor frequency of 100Hz. The adaptive impedance controller begins running the core control cycle with a period of 1ms. Within each cycle, the controller executes the following sub-steps: S521. Read force feedback. Read the current contact force Ff=(Fx, Fy, Fz) and torque Mf=(Mx, My, Mz) between the end effector and the ground from the six-dimensional force / torque sensor.

[0052] S522. Calculate the position correction. The controller compares the actual contact force Ff with the desired target contact force Fd to obtain the force error ef = Fd - Ff. Fd is automatically matched and applied from the controller's internal desired force parameter library based on the contact surface properties identified by S300; for example, a typical value of 200N is used for medium-hardness backfill soil. Then, the end position correction ΔX required to bring the force error to zero is calculated using the impedance control law. The impedance control law formula is: Md×(ΔXdd)+Bd×(ΔXd)+Kd×(ΔX)=ef; Where ΔXdd is the corrected acceleration, ΔXd is the corrected velocity, and ΔX is the position correction. Md, Bd, and Kd are the target inertia, damping, and stiffness matrices preset by the controller, respectively, all of which are diagonal matrices. In this step, the controller mainly focuses on steady-state force tracking; therefore, the inertia term Md×(ΔXdd) has a relatively small impact, and adjustment is mainly achieved through the damping term Bd×(ΔXd) and the stiffness term Kd×(ΔX). By solving this second-order differential equation, the required position correction ΔX can be obtained.

[0053] S523. Merge boundary constraints. The position correction ΔX calculated in step S522 is superimposed on the desired position Xd of the initial guiding trajectory to obtain the corrected desired position Xc = Xd + ΔX. Subsequently, the system checks whether Xc satisfies the boundary constraint inequality SDF(Xc) greater than or equal to 0 in step S422.

[0054] If the conditions are met, Xc is directly sent as the new target position to the robot's underlying position controller.

[0055] If the conditions are not met, meaning Xc is outside the feasible region Ω, a collision is likely to occur, triggering the constraint handling logic. The system calculates the vector from the current position to Xc and solves for its intersection with the boundary Bs(u). Subsequently, the target position is forcibly corrected to the intersection on the boundary, and the damping coefficient Bd in the impedance control law is increased, for example, by 20%, to simulate the collision rebound effect. This makes the robot behave more rigidly when approaching the boundary, preventing the continuous application of excessive force in an attempt to intrude into the boundary.

[0056] S524. Execute motion commands. The final determined target position Xt is converted into angular displacement commands for each joint of the robot through inverse kinematics calculation, including the waist rotation joint, shoulder joint, elbow joint, wrist rotation joint, wrist pitch joint, and wrist yaw joint, and sent to the servo drivers of each joint module to drive the motors to move.

[0057] S530. Continuous Operation and Iteration. The cycles S520 to S524 described above run continuously at a frequency of 1kHz, driving the vibratory head to compact the material along a trajectory that conforms to the physical boundary and meets the constant contact force requirement. Throughout the process, the passive vibration damping base and the active inertial compensation platform continuously operate, isolating vibration interference from the roller body and ensuring the stability of the force control system. The measured vibration isolation efficiency is no less than 90%. This allows the trajectory tracking error throughout the entire motion process—the deviation between the actual tool center point position and Xt—to be effectively controlled within 1mm.

[0058] To address the inherent drawbacks of traditional preset trajectories, such as their inability to adapt to irregular edges and corners, leading to issues like under-compression, collisions, and trajectory deviations, this embodiment utilizes force sensors to perceive boundary positions in real time after posture preparation and establishes a boundary constraint model for confined spaces. This enables real-time dynamic adjustment of the motion trajectory. This mechanism allows the robot to automatically avoid obstacles and closely conform to complex boundaries during operations, reducing trajectory tracking errors, under-compression rates, and collision risks. It achieves safe compaction with no blind spots and near-zero collisions, significantly improving the robot's operational flexibility and safety in extremely confined spaces.

[0059] Example 2 This embodiment provides a road roller for the above-described compaction method. For example... Figure 2 , Figure 3 As shown, the road roller includes a vehicle body and a boom-mounted vibratory compaction robot mounted on it. The robot specifically comprises three core components: a robot-road roller integrated connection structure, a multi-degree-of-freedom robot, a communication and data link device, and an adaptive impedance controller.

[0060] The integrated connection structure includes a mounting base 1, a passive vibration damping base 2, an active inertia compensation platform 3, and a through mechanism.

[0061] Mounting base 1 is a cast steel or welded steel structure platform with dense reinforcing ribs. It is fixed to the main load-bearing beam of the rear frame of the road roller by high-strength preload bolts. The position avoids the direct impact zone of the rollers, while obtaining the best working field of vision and ensuring that the robot's load and working reaction force can be safely transferred to the host.

[0062] The passive vibration damping base 2 is a high-performance rubber-metal composite vibration damping pad, positioned between the active vibration isolation platform and the mounting platform. Its materials and stiffness have been specially selected and tested to effectively filter out most of the high-frequency, low-amplitude operational vibrations transmitted from the roller's steel wheel, protecting the precision reducers and encoders within the robot's joints. The passive vibration damping base 2 is positioned between the mounting base 1 and the active inertial compensation platform 3, filtering out high-frequency, low-amplitude vibrations. The active inertial compensation platform 3 is positioned on top of the passive vibration damping base 2, using a built-in accelerometer to sense residual vibrations in real time and generate reverse micro-motions to actively counteract the vibrations transmitted to the robot base 4. A through-type mechanism is integrated inside the robot base 4, enabling the robot to rotate continuously without cable entanglement.

[0063] The multi-degree-of-freedom robot includes six degrees of freedom: waist rotation joint 10, shoulder joint 11, elbow joint 12, wrist rotation joint 13, wrist pitch joint 14, and wrist yaw joint 15, as well as a six-dimensional force / torque sensor 6 and a vibrating head 5 mounted on the end flange.

[0064] The communication and data link equipment includes a 4G / 5G industrial router and an onboard Wi-Fi Bluetooth adapter for exchanging data with the remote control center and the main roller.

[0065] An adaptive impedance controller is installed in the main control cabinet 8 of the robot base 4 and is used to execute the control algorithm in the aforementioned method.

[0066] This road roller provides a stable benchmark for the precise force control of the robot in strong vibration environments by using two levels of vibration isolation: passive vibration reduction and active compensation. It achieves continuous rotation through a through-type mechanism, expanding the robot's workspace. Through the cooperation of a six-dimensional force / torque sensor 6 and an impedance controller, it achieves compliant operation guided by force.

[0067] Without effective vibration isolation, force sensor readings would be overwhelmed by the vibration of the main unit, making impedance control unstable; without a through-mechanism, the robot cannot flexibly adjust its posture in confined areas; without a communication link, the robot cannot receive compaction quality heat maps and cannot coordinate with the main roller. Therefore, this overall structure solves the comprehensive mechanical and control challenges of mounting a robot on a high-vibration roller.

[0068] In a further specific example of this disclosure, the passive vibration damping base 2 employs a rubber-metal composite damping pad, a material with excellent damping characteristics that effectively absorbs high-frequency vibration energy. The active inertial compensation platform 3 is driven by a piezoelectric ceramic or an electromagnetic actuator. The piezoelectric ceramic is suitable for high-frequency, small-displacement compensation, while the electromagnetic actuator is suitable for low-frequency, large-displacement compensation; both can be selected based on the actual vibration spectrum. An accelerometer is built-in for real-time sensing of residual vibration and feedback control. The through-mechanism includes an electric slip ring and a hydraulic rotary joint; the electric slip ring transmits power and signals, while the hydraulic rotary joint transmits hydraulic oil. This enables the robot to perform continuous 360-degree rotation without tangling cables or oil pipes.

[0069] In other specific examples disclosed herein, the waist rotation joint 10, shoulder joint 11, elbow joint 12, wrist rotation joint 13, wrist pitch joint 14, and wrist deflection joint 15 respectively employ an integrated module of a hollow shaft servo motor, a harmonic reducer, and dual encoders. The hollow shaft design allows cables to pass through the center of the motor, the harmonic reducer provides a large reduction ratio and high precision, and the dual encoders are used to detect joint positions and achieve full closed-loop control. The boom adopts an aerospace-grade aluminum alloy hollow structure with internal wiring; all motor power lines, encoder feedback lines, and sensor cables pass through the inside of the boom, effectively protecting the cables and reducing motion interference.

[0070] The vibratory head 5 is a small, high-power electric or hydraulic vibratory tamper. The electric version has a power of 3kW, a vibration frequency of 100Hz, and an excitation force of 10kN; the hydraulic vibratory tamper has a working pressure of 20MPa and a flow rate of 15L / min. It is installed at the robot end effector via a quick-change device, with a changeover time of ≤30 seconds. The adaptive impedance controller contains a parameter library to store impedance parameter sets and desired contact force parameters corresponding to different working surface properties. For example, the typical desired contact force for medium-hardness backfill soil is 200N.

[0071] The 4G / 5G industrial router is integrated into a sealed electrical cabinet 9 with an IP67 or higher protection rating on the body of the road roller, and has an operating temperature range of -40°C to +85°C.

[0072] The vehicle-mounted Wi-Fi / Bluetooth adapter is integrated into the protective housing 7 of the robot base 4, with an IP65 protection rating and an operating temperature range of -20°C to +70°C.

[0073] This embodiment is not simply a robot installation; rather, it achieves deep integration between the robot and the main road roller at the physical, power, and information levels through a robot-road roller integrated connection structure, a through-mechanism, and communication and data link devices. This allows the arm-mounted robot to function as an intelligent terminal of the main road roller, seamlessly cooperating with it. While the main machine is compacting large areas, it automatically and efficiently completes the compaction tasks of all corner areas, ultimately freeing manual labor from dangerous and arduous corner work, achieving full mechanization and automation of this process. By constructing a highly integrated and collaborative unmanned compaction operation system, the overall efficiency and safety of corner area compaction are greatly improved.

[0074] By deeply integrating a six-dimensional force / torque sensor with an adaptive impedance controller, the robot acquires human-like tactile senses. Before operation, it can intelligently identify contact surface attributes and self-tune parameters; during operation, it can sense force signals in real time and dynamically adjust the arm's flexibility and rigidity, automatically achieving fit or disturbance rejection, effectively avoiding over- or under-pressure. Combined with compaction quality heatmap guidance and in-situ verification, an intelligent closed loop of perception-decision-execution-verification is formed, ensuring that the compaction quality at each point meets the set standards, achieving a leap from human experience to data-driven operation.

[0075] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the compaction method described in Embodiment 1. The storage medium can be a static random access memory, an electrically erasable programmable read-only memory, an erasable programmable read-only memory, a programmable read-only memory, a read-only memory, a disk storage device, a flash memory, a magnetic disk, or an optical disk, etc. This medium can be installed in the controller within the robot's main control cabinet or in a server at a remote control center.

[0076] While the specific embodiments of the present invention have been described above, they are not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A compaction method for narrow or corner areas, characterized in that, The road roller is equipped with a boom-mounted vibratory compaction robot, with a vibratory head at the end of the robot. The compaction methods include: Receive instructions for compaction quality heat mapping operations; Perform global positioning and attitude preparation to point the vibrating head toward the target area; When the vibrating head is not activated, the robot end sweeps along a preset path. When the contact force exceeds the set threshold, the coordinates of the contact point are recorded to obtain a discrete boundary point set. Feasible region boundary constraints are established by fitting physical boundary curves to discrete boundary point sets and expanding outward safety distances. The initial guide trajectory within the feasible area is generated based on the compaction quality heat map. The vibratory head is started, the end contact force is read, the position correction is calculated based on the impedance control law, the corrected expected position is compared with the boundary constraints, and if they are satisfied, it is taken as the target position; otherwise, it is corrected to the boundary and the impedance parameters are adjusted to drive the vibratory head to perform compaction operations in narrow or corner areas.

2. The compaction method for narrow or corner areas as described in claim 1, characterized in that, The safety distance is equal to the sum of the maximum envelope radius of the vibrating head and the preset safety margin; the physical boundary curve is fitted using a cubic B-spline curve fitting algorithm.

3. The compaction method for narrow or corner areas as described in claim 1, characterized in that, Before sweeping, the process also includes intelligent identification of the contact surface and self-tuning of controller parameters: the working surface is contacted with a gentle force, the contact force signal characteristics are collected to identify the physical properties of the working surface, and then the adaptive impedance controller is matched and loaded with the corresponding impedance parameter set, which includes the desired inertial parameter, damping parameter and stiffness parameter.

4. The compaction method for narrow or corner areas as described in claim 3, characterized in that, The impedance control law uses the formula Md×ΔXdd+Bd×ΔXd+Kd×ΔX=ef to calculate the position correction, where ef is the force error, Md, Bd, and Kd are the target inertia, damping, and stiffness matrices, respectively, ΔXdd is the corrected acceleration, ΔXd is the corrected velocity, and ΔX is the position correction.

5. The compaction method for narrow or corner areas as described in claim 1, characterized in that, During the sweeping process, after recording a boundary touch point, the robot end effector is controlled to retreat a safe distance along the original path, and then continue sweeping after deviating from the path.

6. The compaction method for narrow or corner areas as described in claim 1, characterized in that, Based on the priority indicated by the compaction quality heat map, guide trajectories for high-priority areas are planned and executed first, and then low-priority areas are processed.

7. A road roller for use in the compaction method for confined or corner areas as described in any one of claims 1-6, characterized in that, This includes the road roller chassis and the boom-mounted vibratory compaction robot mounted on it. The arm-mounted vibratory compaction robot includes a robot-roller integrated connection structure, a multi-degree-of-freedom robot, a communication and data link device, and an adaptive impedance controller. The robot-roller integrated connection structure includes a mounting base, a passive vibration damping base, an active inertia compensation platform, and a through mechanism. The mounting base is fixed to the rear frame of the roller. The passive vibration damping base is located between the mounting base and the active inertia compensation platform. The active inertia compensation platform is located on the passive vibration damping base. The through mechanism is integrated inside the robot base to achieve continuous rotation. The multi-degree-of-freedom robot includes a waist rotation joint, a shoulder joint, an elbow joint, a wrist rotation joint, a wrist pitch joint, a wrist yaw joint, a vibrating head, and a six-dimensional force / torque sensor mounted on the end flange. The communication and data link device includes a 4G / 5G industrial router and a vehicle-mounted Wi-Fi Bluetooth adapter. The adaptive impedance controller is installed in the main control cabinet of the robot base.

8. The road roller as described in claim 7, characterized in that, The passive vibration damping base is a rubber-metal composite vibration damping pad, and the active inertial compensation platform is driven by a piezoelectric ceramic or electromagnetic actuator and has a built-in accelerometer; the through mechanism includes an electric slip ring and a hydraulic rotary joint, enabling the robot to perform 360-degree continuous rotation without getting tangled in cables.

9. The road roller as described in claim 7, characterized in that, The waist rotation joint, shoulder joint, elbow joint, wrist rotation joint, wrist pitch joint, and wrist deflection joint each adopt an integrated module of hollow shaft servo motor, harmonic reducer, and dual encoder. The boom adopts a hollow structure with internal wiring. The vibrating head is an electric vibrating tamper or a hydraulic vibrating tamper. The adaptive impedance controller contains a parameter library to store impedance parameter groups and desired contact force parameters corresponding to different working surface attributes.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the compaction method for narrow or corner areas as described in any one of claims 1-6.

Images