Excavator guiding method and system based on satellite system

Abstract

The embodiment of the invention discloses an excavator guiding method and system based on a satellite system. The method comprises the following steps: acquiring reference data of a target area based on satellite data of a satellite system; generating first guidance information according to the reference data and equipment parameters of the excavator; in the excavation process of the nth period according to the excavation guidance of the nth period, a sensor on the excavator is used for obtaining sensing data of the nth period; according to the sensing data of the nth period, micro characteristic data of the nth period of the mining area are determined; the microcosmic characteristic data of the nth period comprises the mining guidance of the (n + 1) th period generated according to the microcosmic characteristic data of the nth period, the mining guidance of the nth period, the actual mining condition of the nth period, the mining target and hardware parameters of the excavator; the excavation guidance of the (n + 1) th period comprises advancing route data of the excavator and / or excavation control parameters of the bucket; and excavating according to the excavating guidance of the (n + 1) th period.

Description

Technical Field

[0001] This invention relates to the field of excavation technology, and in particular to an excavator guidance method and system based on a satellite system. Background Technology

[0002] In numerous fields such as civil engineering, mining, and infrastructure construction, excavators, as core earthmoving machinery, have a decisive impact on the overall cost, schedule, and quality of engineering projects due to their operational efficiency and precision. However, traditional excavator operation modes and existing intelligent guidance technologies still face several fundamental technical challenges that urgently need to be addressed. For a long time, excavator operations have relied heavily on the operator's personal experience and visual judgment. Before construction, surveyors need to use equipment such as total stations to lay out the site and set up a large number of physical stakes as positional benchmarks. During excavation, the operator needs to frequently stop the machine to estimate the relative position of the bucket and the digging depth by comparing with design drawings, observing surrounding stakes, or being directed by others. This method not only introduces serious human error, leading to frequent over-digging and under-digging, but also causes frequent interruptions to the work process, resulting in low production efficiency, and places extremely high demands on the operator's technical skills and concentration. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide an excavator guidance method and system based on a satellite system. The technical solution of the present invention is implemented as follows: A first aspect provides a satellite-based excavator guidance method, the method comprising: Based on satellite data from a satellite system, baseline data for the target area is acquired; the baseline data includes macroscopic characteristic data of the target area; the macroscopic characteristic data includes at least surface data and / or basic geological data. First guidance information is generated based on the benchmark data and the equipment parameters of the excavator; the first guidance information is used to determine the excavation target and the initial excavation calibration. During the digging process of the nth cycle according to the digging guidance of the nth cycle, the sensors on the excavator are used to acquire the sensing data of the nth cycle; the sensing data is used to describe the interaction between the bucket of the excavator and the digging area; n is a positive integer; when n equals 1, the excavator provides digging guidance according to the first guidance information; Based on the sensor data of the nth cycle, determine the microscopic characteristic data of the excavation area for the nth cycle; the microscopic characteristic data of the nth cycle includes: soil characteristic data and hydrological characteristic data of the excavation area for the nth cycle; Based on the microscopic characteristic data of the nth period, the excavation guidance of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator, the excavation guidance for the (n+1)th period is generated; the excavation guidance for the (n+1)th period includes the travel route data of the excavator and / or the excavation control parameters of the bucket. Mining is carried out according to the mining guidelines of the (n+1)th cycle.

[0004] The second party provides a satellite-based excavator guidance system, comprising a satellite communication subsystem, an inertial subsystem, and a guidance subsystem. The satellite communication subsystem communicates with a satellite and acquires satellite data. The inertial subsystem includes at least one sensor for data sensing. The guidance subsystem acquires reference data for a target area based on the satellite data. The reference data includes macroscopic characteristic data of the target area, including at least surface data and / or basic geological data. First guidance information is generated based on the reference data and the excavator's equipment parameters. This first guidance information is used to determine the excavation target and initial excavation calibration. During the excavation of the nth cycle according to the excavation guidance, the sensors on the excavator acquire the sensing data for the nth cycle. According to the sensor data, the interaction between the excavator's bucket and the excavation area is described; n is a positive integer; when n equals 1, the excavator provides excavation guidance based on the first guidance information; based on the sensor data of the nth period, the microscopic characteristic data of the excavation area for the nth period is determined; the microscopic characteristic data of the nth period includes: soil characteristic data and hydrological characteristic data of the excavated area in the nth period; based on the microscopic characteristic data of the nth period, the excavation guidance of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator, the excavation guidance for the (n+1)th period is generated; the excavation guidance for the (n+1)th period includes the excavator's travel route data and / or the bucket's excavation control parameters; excavation is carried out according to the excavation guidance for the (n+1)th period.

[0005] A third aspect provides a computer-readable storage medium storing computer-executable instructions; the computer-executable instructions, when executed by a processor, can realize the aforementioned excavator guidance method based on a satellite system.

[0006] The technical solution provided in this disclosure has the following effects: 1. Improved geological adaptability: By periodically acquiring and processing sensor data generated by the interaction between the bucket and the soil (including but not limited to vibration spectrum and resistance values), the system can retrieve the soil mechanical properties at the work site in real time, such as micro-parameters like compaction, cohesion, and moisture content. This enables the system to identify local changes in geological conditions, such as the transition from loose fill to hard undisturbed soil layers, or the encounter with localized weak interlayers, providing a data foundation for dynamic adjustments to operational strategies.

[0007] 2. Closed-loop control and dynamic optimization of the operation process were achieved: A feedback-based control closed loop was constructed. The system discretizes the operation process into continuous decision-making cycles, and the execution results of each cycle (represented by micro-characteristic data and actual mining conditions) are used as inputs to generate action guidelines for the next cycle. This iterative optimization mechanism allows mining parameters (such as trajectory, speed, and entry angle) to be adaptively adjusted based on the execution effect of the previous cycle, thereby improving the control accuracy and robustness in nonlinear and time-varying operating environments.

[0008] 3. The system achieves synergy between macro-level planning and localized execution: It utilizes satellite positioning data to establish a global spatial reference frame and excavation targets, ensuring the macro-level accuracy of the operation. At the execution layer, it integrates real-time, localized micro-geological sensing data to generate refined operational instructions tailored to the specific working conditions. This synergistic mechanism guarantees that local actions always serve the global objective, avoiding deviations from the operational path caused by localized environmental interference.

[0009] 4. Reduced reliance on subjective human factors in work results: By automating the analysis, decision-making, and guidance generation process based on multi-source sensor data, the system transforms the operator's role from leader to supervisor. This reduces over-reliance on the operator's personal experience and instantaneous judgment, helping to reduce fluctuations in work quality caused by individual differences and state variations, thereby improving the consistency and predictability of construction results. Attached Figure Description

[0010] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a satellite-based excavator guidance method provided in an embodiment of the present invention; Figure 2A This is a schematic diagram of 3D visual guidance provided in an embodiment of the present invention; Figure 2B This is a schematic diagram of 3D visual guidance provided in an embodiment of the present invention; Figure 2CThis is a schematic diagram of the terminal display in the third mode provided in an embodiment of the present invention; Figure 3 A schematic diagram of a satellite-based excavator guidance system provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a guidance system for an excavator provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device containing the guidance system of an excavator, as provided in an embodiment of the present invention. Detailed Implementation

[0011] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0012] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0013] like Figure 1 As shown, this disclosure provides a satellite-based excavator guidance method, the method comprising: S1110: Acquire baseline data for the target area based on satellite data from the satellite system; the baseline data includes macroscopic characteristic data of the target area; the macroscopic characteristic data includes at least surface data and / or basic geological data; S1120: Generate first guidance information based on the baseline data and the excavator's equipment parameters; the first guidance information is used to determine the excavation target and initial excavation calibration; S1130: During the digging process of the nth cycle according to the digging guidance of the nth cycle, the sensors on the excavator are used to acquire the sensing data of the nth cycle; the sensing data is used to describe the interaction between the bucket of the excavator and the digging area; n is a positive integer; when n equals 1, the excavator provides digging guidance according to the first guidance information; S1140: Based on the sensor data of the nth period, determine the microscopic characteristic data of the excavation area for the nth period; the microscopic characteristic data of the nth period includes: soil characteristic data and hydrological characteristic data of the excavation area in the nth period; S1150: Based on the micro-characteristic data of the nth period, the digging guidance of the nth period, the actual digging situation of the nth period, the digging target, and the hardware parameters of the excavator, generate the digging guidance for the (n+1)th period; the digging guidance for the (n+1)th period includes the excavator's travel route data and / or the bucket's digging control parameters. S1160: Mining is performed according to the mining guidelines of the (n+1)th cycle.

[0014] The guidance method provided in this embodiment can be used for guidance or control on the local excavator, or for a cloud platform connected to the excavator. Figure 2A The image shows an excavator equipped with a management platform, sensors, a GSNN antenna, etc. For example, the management platform is located in the operator's cab for viewing or use by the operator. Sensors may include tilt sensors. GNSS antennas may be integrated with other GNSS antennas. Exemplarily, a BeiDou three-antenna integrated terminal receiver can be incorporated into the excavator. This BeiDou three-antenna integrated terminal receiver can be designed specifically for engineering machinery applications, integrating data acquisition, display, radio, and 4G / 5G modules. It has two built-in BeiDou high-precision positioning and orientation boards, simultaneously outputting three sets of positioning data: 1cm horizontal accuracy, 2cm vertical accuracy, 2m orientation accuracy, 0.09° baseline, three-star eight-band, and IP65 compliant.

[0015] The GNSS antenna is an eight-band ruggedized measurement antenna, featuring high gain, miniaturization, high sensitivity, multi-system compatibility, and high reliability. Its base is made of high-strength metal and it is mounted on the excavator cab to provide position measurement for the excavator, truly meeting the needs of excavator operation in vibration and shock environments. The tilt sensor utilizes Earth's gravity, determining the tilt angle by measuring the angle between gravitational acceleration and the sensitive axis, providing high-precision data. The tilt sensor is installed on key components such as the boom, stick, and bucket to measure the tilt angle, helping the control system adjust its position and prevent overload or over-digging. The error can be controlled within 0.5 degrees. By outputting three-dimensional angle data (roll, pitch, and azimuth), the sensor provides real-time feedback on the machine's attitude, avoiding under-digging or accidents and reducing operator reliance on experience.

[0016] Before proceeding with the detailed explanation of each step, the terminology used in this embodiment is described as follows: Satellite systems: Refers to Global Navigation Satellite Systems (GNSS), including but not limited to BeiDou, GPS, GLONASS, Galileo, etc. The satellite data they provide includes satellite signals, carrier phase, ephemeris data, etc., used to calculate the receiver's position, velocity, and time information.

[0017] Baseline data: Reference data for the target area obtained by processing satellite data and other geographic information. Macroscopic characteristic data is a part of the baseline data, referring to data that describes the overall characteristics of the target area at a macroscopic scale.

[0018] Surface data: refers to the three-dimensional surface model of the target area, usually represented by a digital surface model (DSM) or a digital elevation model (DEM).

[0019] Basic geological data refers to data obtained through preliminary geological surveys, such as generalized stratigraphic maps and groundwater depth. It is relatively large in scale and relatively low in accuracy.

[0020] Equipment parameters: These refer to the inherent mechanical parameters of the excavator, including boom length, stick length, bucket specifications, maximum digging force, and maximum digging radius.

[0021] First guiding information: Initial work plan, used to determine the excavation target (such as the final design outline, depth, satellite coordinates, ground coordinates, etc. of the foundation pit) and initial excavation marking (such as the approximate area and depth of the first dig).

[0022] Sensor data: Physical quantity data collected by sensors installed on the excavator that reflects the interaction between the bucket and the digging area. This interaction is mainly reflected in mechanical and vibrational characteristics.

[0023] Microscopic characteristic data: Data derived from sensor data that describes the geological conditions of a localized area currently being excavated. Its spatial resolution is far higher than that of macroscopic characteristic data, reaching accuracy to the centimeter level. For example: Soil property data: such as soil compaction, cohesion and / or internal friction angle.

[0024] Hydrological characteristics data, such as soil moisture content and / or saturation.

[0025] Actual excavation status: refers to the actual changes in the shape of the working face after the nth excavation cycle, which can be obtained by comparing the three-dimensional point cloud or model before and after excavation.

[0026] Excavation control parameters: specific command values ​​that guide the bucket's actions, such as the bucket's desired angle of entry, desired cutting speed, and / or vibration mode.

[0027] In S1110, a global spatial reference and initial state model of the operating environment are established. Specifically, this may include: receiving satellite signals via a GNSS receiver mounted on the engineering machinery or acquired through other means; and obtaining centimeter-level precision three-dimensional coordinates of measurement points using technologies such as Real-Time Dynamic Differential (RTK). A high-precision digital surface model (DSM) or digital elevation model (DEM) of the operating area is constructed as surface data through multi-point measurement or mobile scanning. Existing geological survey reports can be digitized and imported into the system as basic geological data. These macroscopic characteristic data collectively constitute the operational baseline. The excavator itself can carry a satellite communication subsystem, which includes at least a satellite communication module, enabling satellite communication. For example, the satellite communication subsystem can communicate with the BeiDou satellite system and perform positioning and ground remote sensing scanning based on BeiDou satellites.

[0028] In S1120, global task planning and initialization are performed. Based on baseline data (such as the DEM model of the designed slope) and equipment parameters (such as the excavator's maximum digging radius and bucket size), an initial work path and starting point, i.e., the first guidance information, are generated through a path planning algorithm. This guidance is used to determine the digging target (the geometry of the final shaped surface, for example, the number of cubic meters of soil to be excavated in this shift, the digging depth, the aspect ratio, etc.) and initial calibration (the starting position and attitude of the machinery).

[0029] In S1130: Install sensors on mechanical working devices (such as buckets) or on critical load-bearing components.

[0030] Vibration sensor: Collects high-frequency vibration signals generated when the working device cuts, crushes, or impacts materials. Pressure sensor: Monitors hydraulic system pressure or resistance directly experienced by the working device.

[0031] In S1140, a mapping model between sensor data and microscopic characteristics is established. Based on this mapping, the microscopic characteristic data for the nth period (current period) is obtained. Specifically, this can be achieved in the following ways: Method 1 (Feature Library Based): A feature library of sensor data corresponding to different materials (such as loose soil, dense clay, and rock) is pre-established through experiments (e.g., specific frequency band energy of vibration signals, rate of change of pressure curves). Real-time data is matched with the feature library to qualitatively or semi-quantitatively identify soil characteristics (e.g., "hardness level: high") or water ripple characteristics (e.g., "water content: saturated").

[0032] Method 2 (based on physical model): Establish a dynamic model of the interaction between the working device and the work object, and estimate the mechanical parameters of the work object (such as internal friction angle and elastic modulus) directly from the sensor data (such as resistance value) through parameter inversion algorithm.

[0033] In S1150, adaptive operation guidance for the (n+1)th cycle (next cycle) is generated, for example, by making forward-looking decisions based on real-time sensing results. The current working face characteristics are sensed to obtain microscopic characteristic data for the nth cycle; the deviation between the actual working trajectory and the planned trajectory obtained through the positioning system yields the actual working status for the nth cycle. Based on this information, excavation guidance for the (n+1)th cycle is dynamically generated. For example, if the current area is perceived to have excessively high hardness, excavation control parameters suggesting a reduction in cutting depth and operating speed are generated to protect the equipment; if loose material is sensed ahead, instructions to adjust the travel route data are generated to improve efficiency and safety.

[0034] In S1160, digging is performed according to the digging guidance of the (n+1)th cycle. For example, the generated guidance can be automatically executed by the vehicle control system (fully automatic mode), or presented to the operator in graphical or numerical form on the human-machine interface to assist in precise operation (assistance mode). At this point, a complete "perception-decision-execution" cycle ends, and the system immediately enters the (n+1)th cycle, repeating this process until the task objective is achieved. In this embodiment, if the guidance of the (n+1)th cycle is displayed on the terminal, it can be displayed with a 3D visual effect. For example, as... Figure 2A and Figure 2B As shown, the guidance can specifically provide the depth, height, angle of the bucket tip digging, and the overall posture of the excavator. In this way, real-time 3D vision guidance, graphic and numerical information display, and task visualization eliminate the need for measurement, layout, and piling, saving time and costs.

[0035] In some embodiments, the sensing data for the nth period includes spectral data for the nth period and pressure data for the nth period; wherein, a plurality of pressure sensors are located on the inner surface of the bucket for detecting digging pressure.

[0036] In some embodiments, S1130 may specifically include at least one of the following: Obtain the spectral data of the nth period and the excavation resistance distribution data of the pressure data of the nth period; Based on the spectral data and the excavation resistance distribution data, determine the soil properties of the area to be excavated in the nth cycle; the soil properties data are used to indicate at least one of the following: the degree of soil looseness, whether it has entered a hard rock layer, and whether it is a waterlogged area.

[0037] Spectrum data: Refers to the characteristic data obtained from the original signal collected by vibration or acoustic sensors through signal processing techniques such as fast Fourier transform (FFT), which represents the distribution of signal energy over different frequency components. It reflects the "frequency fingerprint" of the vibration or sound generated when the bucket interacts with the material. When the bucket cuts materials with different densities or compositions, vibration spectra with different characteristics will be generated. For example, when cutting dense clay, a spectrum with a relatively high main frequency and concentrated energy may be generated; when passing through loose sand, a spectrum with a wider frequency distribution and dispersed energy may be generated; when hitting a rock, a spectrum feature with extremely high frequency and transient impact will be generated.

[0038] Pressure data / Digging resistance distribution data: Refers to the data measured by multiple pressure sensors or strain gauges arranged on the working surface of the bucket (such as the inner wall and cutting edge), which reflects the normal or tangential forces received by different parts of the bucket during the digging process. The digging resistance distribution data is formed after spatial integration and statistical analysis of these pressure data, and is used to describe the magnitude, direction of the resistance and its distribution on the bucket surface. Sensor arrangement: Micro pressure sensors are arranged in an array on the inner side wall, bottom plate and cutting edge of the bucket.

[0039] Data form: When the bucket is filled with materials, the bottom plate sensors show uniform high pressure; when encountering a large rock jam, pressure peaks may only appear in local sensors at the cutting edge; when operating in fluid mud, the overall resistance may be low and the fluctuations are small. This spatial distribution and temporal variation together constitute the digging resistance distribution data.

[0040] Soil property data: Parameters obtained through model inversion or feature recognition based on spectrum data and pressure data, which are used to quantitatively or qualitatively describe the physical and mechanical properties of the excavated materials. It is the specific manifestation of "microscopic characteristic data". For example, the looseness of the soil: can be inverted into a density index (such as 0 - 100%). For example, an index > A% indicates very dense, and an index < B% indicates very loose. A is greater than B.

[0041] Whether entering the hard rock layer: A formation type identifier is pre-configured, such as "1 - soil layer", "2 - strongly weathered rock", "3 - moderately weathered rock". When the system recognizes that a specific high-frequency resonance peak appears in the spectrum and the resistance increases sharply, the identifier can be switched from "1" to "2" or "3".

[0042] Whether it is a waterlogged area: An aqueous state identifier can be output, such as "dry", "wet", "saturated". For example, when the resistance is significantly reduced and the vibration spectrum shows obvious low-frequency damping characteristics, it can be judged as the "saturated" state.

[0043] In some embodiments, multiple resistance strain gauge pressure sensors are installed in a grid array on the inner wall and bottom plate of the excavator bucket to measure the pressure of material on various parts of the bucket during digging, thereby calculating pressure data and synthesizing digging resistance distribution data.

[0044] In some embodiments, large-scale earthmoving projects typically require multiple excavators to work collaboratively. In the current operating mode, each device operates independently while the system coordinates, using communication technologies such as V2X to connect multiple independent intelligent excavators into a collaborative "device cluster," sharing information, making collaborative decisions, and dynamically reallocating tasks, thereby achieving optimal system-level efficiency and functional fault tolerance.

[0045] This system, based on single-machine intelligence, adds a collaborative control layer and a communication module. Each excavator is equipped with the system described in claim 1, and an additional communication module is added, enabling it to send and receive collaborative information. The collaborative decision controller can be a distributed or centralized logical control unit. It can utilize edge computing nodes (centralized), or the decision algorithm can be deployed on each device, performing distributed decision-making through a consensus mechanism. Each excavator periodically broadcasts or multicasts its core status information, forming a shared situational awareness. The interacted information includes, but is not limited to, at least one of the following: Device identity and capability information may include, but is not limited to, at least one of the following: device ID, maximum working efficiency, working range, and other inherent parameters.

[0046] Real-time status information may include, but is not limited to, at least one of the following: satellite positioning information, equipment health status (such as hydraulic oil temperature, engine load rate), and current operating mode.

[0047] Task progress information includes, but is not limited to: the current sub-excavation target, the amount of work completed, and the geological characteristics of the current working face derived from sensor data (e.g., "Hard rock strata were discovered in the southeast corner of Area A, delaying progress by 30%").

[0048] The intention is to share information, including but not limited to: the projected action path for the next work cycle.

[0049] In some embodiments, tasks are redistributed when one or more of the following conditions exist.

[0050] A device reports a fault code or communication timeout through self-diagnosis; Significant schedule deviation (by comparing the planned schedule with the actual schedule, it was found that the schedule of the area under the responsibility of a certain equipment was seriously lagging behind). Sudden change in geological conditions (a piece of equipment reported encountering extremely unfavorable geological conditions, with the estimated processing time far exceeding the plan); Adjust the algorithm (dynamically divide the overall excavation target (such as the entire foundation pit) into several finer-grained virtual task units in the cloud or edge server).

[0051] The redistribution of tasks may include, but is not limited to, dynamically dividing the total excavation target (such as the entire foundation pit) into several more granular virtual task units in the cloud or edge server.

[0052] Auction-style task allocation can be used. For example, when tasks need to be reassigned (e.g., due to equipment A malfunction), the system packages the unfinished task units of equipment A and "auctions" them. Surrounding healthy equipment (e.g., equipment B and C) calculate a "cost" or "benefit" and submit a "bid" based on their remaining working capacity, distance from their current location to the task unit, and their perceived geological adaptability (e.g., equipment B has recently processed similar rock formations, making its process parameter library a better match). The system dynamically allocates the task unit to the optimal equipment based on the principle of "lowest bidder wins" or "highest overall benefit." This transforms task allocation from static planning to dynamic market bidding, achieving optimal system efficiency.

[0053] Thus, when device A malfunctions, its surrounding devices immediately receive a "fault offline" message via V2X or other communication methods. Based on the task allocation mechanism described above, devices B and C will automatically redraw their operational boundaries, covering the unfinished areas of device A. The new excavation target (i.e., the adjusted boundary lines) will be issued and displayed on the terminals of each device.

[0054] For large obstacles (such as huge boulders), two pieces of equipment can be temporarily instructed to work together. For example, equipment B is responsible for loosening the rocks, while equipment C is responsible for clearing debris. The two share the bucket position and action intentions via V2X, and a collision-avoidance cooperative path is generated through system or equipment negotiation to ensure absolute safety.

[0055] In some embodiments, a triaxial accelerometer is installed at the hinge point connecting the bucket and the stick to collect vibration signals generated during the excavation process.

[0056] In some embodiments, the vehicle-mounted industrial control computer has a built-in data acquisition card and a signal processing module.

[0057] In S1130, during the nth cycle of excavation, the data acquisition card synchronously acquires the raw signals from all pressure and vibration sensors at a first rate. The signal processing module performs FFT spectral analysis on the raw vibration signals, extracting spectral data within the execution frequency range, focusing on the energy values ​​of specific frequency bands. This frequency range may include one or at least two of the acoustic, infrasonic, and / or ultrasonic ranges. Simultaneously, the processing module integrates the readings from all pressure sensors to generate a pressure distribution cloud map on the bucket surface at the current moment, and calculates the magnitude, direction, and variance of the resultant force as excavation resistance distribution data.

[0058] In S1140, a method based on feature extraction and pattern classification can be used to determine soil property data. Specifically, this may include: extracting features from spectral data and inputting the extracted multi-dimensional feature vectors into a pre-trained support vector machine (SVM) classification model; the classification model outputs a soil type discrimination result (i.e., the qualitative part of the soil property data) and a confidence score of belonging to the category based on the input feature vectors; and finally quantizing the output based on the confidence score.

[0059] The extracted features may include, but are not limited to: calculating the energy proportion of the low-frequency band below the first frequency, whether there is a significant peak in the second frequency band, and the centroid frequency of the spectrum. The second frequency is higher than the first frequency; for example, the second frequency may be 2-3 kHz, but is not limited to 2-3 kHz.

[0060] Features can be extracted from the resistance distribution data, specifically including: calculating the average resistance value, the standard deviation of resistance fluctuation, and the location of the maximum pressure point.

[0061] The classification model was trained by collecting sample data (i.e., the corresponding spectrum and pressure characteristics) of four typical soil types, namely "loose sand", "dense clay", "strongly weathered rock" and "saturated silt", in the test section or simulation before construction.

[0062] The final quantitative output of soil property data can be a matrix, for example: {Soil type: strongly weathered rock, confidence level: xxx%, density index: xxx, water content: xxxx}.

[0063] This data is transmitted in real time to the downstream decision module (S1150) to generate excavation guidance for the (n+1)th cycle. For example, when "strongly weathered rock" is identified, the guidance for the next cycle will suggest "reducing the feed rate and adopting a high-frequency, low-amplitude vibration mode" to protect the equipment; when "saturated silt" is identified, it will suggest "increasing the bucket opening and passing through quickly" to avoid getting stuck.

[0064] By implementing this embodiment, the system can cross-validate soil information from different physical dimensions by fusing spectral data and pressure distribution data, thereby improving the accuracy and reliability of identification.

[0065] In this embodiment, spectrum data and pressure data are used, and the two are fused together to establish a deep mutual verification and compensation relationship in order to improve the reliability of the final soil property data output.

[0066] A real-time confidence assessment module is established for both spectral and pressure data. The confidence score (C_f) for spectral data is calculated based on the signal-to-noise ratio (SNR) and spectral sharpness (e.g., peak sharpness). A high SNR and sharp peaks result in a high confidence score. The confidence score (C_p) for pressure data is calculated based on the consistency and stability of pressure sensor readings. Small variances and smooth curves across all pressure sensor readings result in a high confidence score. The final soil classification decision function D is a weighted sum of the outputs from both classes of data: D = wf * Df + wp * Dp; where wf = Cf / (Cf + Cp), wp = Cp / (Cf + Cp), and Df and Dp are the preliminary classification results obtained based solely on spectral and pressure data, respectively.

[0067] When the vibration signal is greatly disturbed by external factors (low Cf), the system will automatically rely more on the pressure data; and vice versa. This enables dynamic mutual correction of the data and improves the robustness of the system under complex operating conditions.

[0068] In specific analysis, it's not just about analyzing data at a single point in time, but also about analyzing the changes in the spectrum and pressure data within a short time window (such as a complete cutting action), following a time-series correlation rule. For example, a smooth rise in the resistance curve should be observed, along with a steady increase in the mid-to-high frequency energy of the vibration spectrum. These two are positively correlated. When encountering internal boulders, a sharp pulse peak will first be observed in the resistance curve, followed (with a possible delay of a few milliseconds) by a high-frequency impact component in the vibration spectrum. If the system only detects the resistance pulse but not the corresponding high-frequency vibration, it can be determined that the anomaly is likely not due to soil conditions (such as mechanical jamming of the bucket itself), thus rejecting the "entering a hard rock layer" assessment and avoiding false alarms.

[0069] To address temporary failures or signal loss in a single sensor, a comprehensive degradation handling strategy must be designed. The system has multiple preset operating modes, including but not limited to: dual-modal fusion mode (normal mode), spectrum-dominated mode, stress-dominated mode, and historical data prediction mode.

[0070] When a continuous loss of pressure data is detected (e.g., sensor disconnection), the system automatically switches to "Spectrum-Dominated Mode." In this mode, soil classification relies solely on spectral data, but the system will display a message on the human-machine interface stating "Pressure data missing, recognition reliability reduced." Simultaneously, it may attempt to reconstruct an approximate resistance value as a reference using a correlation model between spectrum and resistance from historical data. Conversely, when vibration data is lost, the system switches to "Pressure-Dominated Mode." This ensures that even with partial functional impairment, the core soil identification function does not completely fail, still providing output with some reference value, achieving graceful degradation.

[0071] A lightweight time-series prediction model (such as a Kalman filter or a simple regression model) is built-in. When short-term data is missing (such as for 100-200 milliseconds), the model can predict the missing sensor data value based on the data trend just before the missing data is lost, and combine it with the excavator's own kinematic state (such as bucket speed and angle), and temporarily fill in the missing data. This provides continuous data input to the decision-making module and avoids interruption in guidance generation.

[0072] An adaptive sensing strategy based on historical processing results: This is a more creative extension, enabling the system to "intelligently" manage its sensing resources based on historical working conditions. A working condition learning module runs in the background, continuously recording the optimal combination of sensing data and its confidence level for different soil types. After a period of learning, the system may discover that when identifying "ordinary clay," the classification contribution (confidence level) of pressure data consistently exceeds 90%, while the contribution of spectral data is less than 10%. In this case, the system can generate an optimization strategy: when the soil type is identified as "ordinary clay" for several consecutive cycles and the stratum is stable, the vibration sensor and spectral analysis module can be automatically and temporarily shut down, or their sampling frequency reduced to a minimum, relying solely on pressure data for monitoring, thus reducing excavation energy consumption and costs. Once the pressure data begins to show abnormal fluctuations or trend changes, the system immediately restarts the vibration sensing system at full power to cope with possible stratum changes. This strategy can significantly reduce system power consumption and unnecessary data processing computation, making it particularly suitable for battery-powered mobile devices and extending operating time. It is also an intelligent resource allocation method.

[0073] If the system is in energy-saving mode with "spectrum acquisition off," a mandatory verification cycle is set (e.g., every 10 work cycles or every 30 seconds). During the mandatory verification cycle, the system briefly wakes up all sensors to perform a complete multimodal data acquisition and fusion judgment, comparing the results with the judgment in energy-saving mode. If the comparison results are consistent, energy-saving mode continues; if a significant deviation occurs, energy-saving mode is immediately exited, and the operator is notified that "geological conditions have changed, and the system has switched to full-sensor mode." This achieves energy saving while ensuring the long-term reliability of the system and avoiding the accumulation of misjudgment risks that may be caused by prolonged use of a single sensor.

[0074] In some embodiments, the spectrum data includes at least one of the following: The spectrum data is the spectrum data of the audio data; the audio data is collected by audio sensors distributed on the outer surface of the bucket or on the boom; The spectral data is the spectral data of the vibration data; the vibration data is collected by vibration sensors distributed on the outer surface of the bucket or on the boom.

[0075] In some embodiments, processing of the audio data may include: selecting a waterproof microphone or sound pressure sensor with high sensitivity and a wide frequency response range (e.g., 20Hz-20kHz). These are then installed inside a protective cover on the outside of the bucket or near the hydraulic lines of the boom, locations that effectively receive digging noise while minimizing direct interference from mechanical friction noise.

[0076] The raw analog electrical signal acquired by the sensor is first amplified by a preamplifier, and then filtered by an anti-aliasing filter to remove high-frequency noise and prevent signal aliasing during sampling. The conditioned analog signal is digitized by a high-precision analog-to-digital converter (ADC) at a sampling rate of no less than 40kHz to obtain raw audio data in discrete time series form. Digital filters (such as bandpass filters) are used to filter out environmental noise unrelated to the excavation operation (such as low-frequency engine noise and wind noise). A more advanced method is to use an adaptive noise reduction algorithm, using reference noise in the cab as a sample, to actively eliminate its influence from the mixed signal. The continuous audio signal stream is divided into short time segments (e.g., 50 milliseconds per frame). A Hanning window or Hamming window function is applied to each frame of signal to reduce spectral leakage caused by signal truncation. The time-domain characteristics of each frame of signal, such as short-time energy and zero-crossing rate, are calculated to preliminarily determine whether the frame contains effective excavation impact sound, thereby filtering out the effective analysis interval. For the preprocessed valid audio frames, the Fast Fourier Transform (FFT) algorithm is applied to transform them from the time domain to the frequency domain, obtaining the linear spectrum of the frame. To better conform to the characteristics of human hearing and highlight the frequency bands with concentrated energy, the linear spectrum can be converted into a Mel spectrum or a logarithmic spectrum.

[0077] Spectrum data output: Finally, the system outputs a numerical matrix that represents the energy distribution of the sound signal at different frequency components within the current time window, which is the audio spectrum data used as the basis for soil identification.

[0078] In some embodiments, processing of vibration data may include: An ICP-type triaxial accelerometer is used due to its built-in amplifier circuitry and strong anti-interference capability. It is rigidly mounted on the back of the bucket or near the boom-stalk connection pin using a magnetic base or high-strength adhesive, ensuring direct sensing of the reaction force from the soil. The sensor directly outputs a voltage signal proportional to the acceleration. This signal is synchronously sampled by an ADC at a high sampling rate (typically ≥10kHz) to obtain the raw vibration acceleration signal containing high-frequency components.

[0079] A high-pass filter is used to remove low-frequency drift caused by equipment movement, and a low-pass filter is used to remove high-frequency electrical noise. The vibration velocity signal is obtained by integrating the acceleration signal once; the vibration displacement signal is obtained by integrating it twice. The velocity signal is more meaningful for assessing the fatigue state of the equipment. The least squares method is used to fit and eliminate linear or polynomial trend terms in the signal to prevent them from affecting the accuracy of the spectrum analysis. The Welch method is used to estimate the power spectral density (PSD) of the preprocessed vibration signal (usually acceleration or velocity). This method effectively smooths random noise and obtains a stable spectrum estimate by segmenting the long signal, windowing it, calculating the FFT separately, and then averaging. In the generated PSD plot, characteristic frequency peaks related to soil-rock interaction are identified, and their center frequencies and amplitudes are recorded.

[0080] Vibration spectrum data output: The final output is vibration spectrum data that characterizes the distribution of vibration energy on the frequency axis. This data is directly related to the mechanical properties of the soil, such as stiffness and damping.

[0081] The audio spectrum reflects the propagation characteristics of airborne sound waves and is sensitive to phenomena such as cavities and fragmentation; the vibration spectrum reflects the propagation characteristics of structural waves and is directly related to the stiffness and damping of the soil. Combining these two provides more comprehensive soil "fingerprint" information. When one sensor is subjected to specific interference (e.g., audio is affected by environmental noise, vibration by engine vibration), the other data can be used for cross-validation, greatly reducing the false positive rate and enhancing the system's robustness. The fusion of these two spectral characteristics enables the system not only to distinguish between broad soil types (e.g., soil vs. rock) but also to perform more refined classifications (e.g., identifying rock strata with different degrees of weathering).

[0082] In some embodiments, S1150 may include: Based on the micro-characteristic data of the nth period, the excavation guidance of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator, the excavation guidance of the (n+1)th period is generated and arranged in order to generate the first matrix. The first matrix is ​​input into the AI ​​model, and the AI ​​model outputs the second matrix; the second matrix corresponds to the mining guidance in the (n+1)th cycle.

[0083] For example, the input information is represented as a matrix (generating the first matrix).

[0084] To enable AI model processing, the multi-source, heterogeneous input information must first be transformed into a regular numerical matrix (the first matrix).

[0085] Microscopic characteristic data vector (M_n): The soil characteristic data (such as hardness grade and moisture content) retrieved in the nth period are numerically encoded. For example, the hardness grade "high" is encoded as 0.9 and the moisture content "saturated" is encoded as 1.0, forming a k-dimensional feature vector.

[0086] Excavation guidance and status vector (G_n): contains the planned guidance (such as target depth and speed) and actual execution status (such as actual depth deviation and trajectory error) for the previous x cycles.

[0087] Mining target vector (T): Target parameters (such as target elevation and slope) of the current local working face extracted from the macro design model.

[0088] Hardware parameter vector (H): Mechanical constraint parameters of the excavator (such as maximum digging force and maximum working radius).

[0089] These vectors are concatenated and normalized to form a one-dimensional feature tensor (the first matrix). This tensor comprehensively represents the complete information of "environmental state - historical actions - machine state - task objective".

[0090] In some embodiments, a deep neural network (DNN) or a deep Q-network (DQN) is used as the AI ​​model. DNNs are suitable for direct supervised learning (learning expert policies), while DQNs are suitable for reinforcement learning (self-optimization through interaction with the environment). These AI models are trained in a simulation environment or using a large amount of historical construction data to generate a large number of "state-action" sample pairs. The state is represented by the first matrix mentioned above, and the action is the desired mining guidance (second matrix). The AI ​​model learns to output action instructions (second matrix) that maximize long-term rewards (such as highest work efficiency, lowest energy consumption, and best accuracy) given the current state (first matrix).

[0091] During construction, the first matrix, constructed in real time, is input into the pre-trained AI model. The model then performs multiple nonlinear transformations to ultimately output a second matrix.

[0092] In some embodiments, the parsing of the second matrix may include: the second matrix is ​​the digging guide for the (n+1)th cycle. It can be parsed into specific control parameters, such as: [travel speed: 0.5m / s, bucket entry angle: 45 degrees, vibration on: yes, vibration frequency: 25Hz].

[0093] In some embodiments, the excavator has a first mode, a second mode, and a third mode. S1160 may include: When the excavator is working in the first or second mode, during the process of generating the excavation guidance for the (n+1)th cycle, the AI ​​model generates control signals based on the second matrix. When the excavator is working in the first mode, it is automatically controlled to dig in the (n+1)th cycle according to the control signal. When the excavator is working in the second mode, it enters the preparatory state of automatic control after generating the control signal, and enters the execution state from the preparatory state when the operator's confirmation input on the control lever is detected. In the execution state, it performs the excavation of the n+1th cycle according to the control signal. When the excavator is operating in the third mode, the digging instructions for the (n+1)th cycle are displayed on the terminal device in the cab; the digging for the (n+1)th cycle is performed based on the operator's input based on the digging instructions for the (n+1)th cycle.

[0094] In some embodiments, the first mode is a fully automatic mode. When the excavator operates in this mode, the second matrix generated by the AI ​​model is directly calculated into the displacement or pressure setpoints of each hydraulic cylinder through the inverse dynamics model of the vehicle controller. These setpoints are sent directly to the excavator's hydraulic servo control system via the CAN bus to drive the pilot valve and the main valve, thereby automatically controlling the excavator to complete all digging actions in the n+1th cycle. The operator's role is that of a supervisor.

[0095] In some embodiments, the second mode can be a human-machine collaborative mode. This mode serves as a safe transition between fully automated mode and manual operation. After the AI ​​model generates the second matrix, the system does not execute immediately but enters a preparatory state. The system clearly displays a preview of the upcoming action (such as the expected trajectory of the bucket) on the HMI interface in the cab. The operator must actively confirm, for example, by pressing a specific button on the control lever or stepping on a foot pedal, before the system transitions from the preparatory state to the execution state and begins automatic execution. This reduces the operator's workload while giving the human final decision-making power, making it particularly suitable for complex or high-risk working conditions.

[0096] In some embodiments, the third mode is a human-led mode. In this mode, the second matrix generated by the AI ​​model is not converted into direct control signals. Instead, it is displayed to the operator on the control screen via augmented reality (AR) or graphical means. For example, a virtual "ideal bucket trajectory" may be overlaid on the screen or a text prompt suggesting "increase the bucket angle" may be given. Based on these intelligent guidance, the operator manually operates the excavator to complete the task. This mode maximizes trust in the operator's experience while providing optimal decision support.

[0097] like Figure 3 As shown in the embodiments of this disclosure, the method further includes: S1170: Detects operator status information; S1180: Based on the status information, determine whether the excavator's working mode is compatible with the operator's status; S1190: When the excavator's working mode is not compatible with the operator's status, output an adjustment prompt or automatically adjust the excavator's working mode.

[0098] In some embodiments, status information refers to objective data collected by onboard sensors that reflects the operator's current level of physiological arousal, fatigue, and / or concentration. Its core components are physiological indicators, behavioral characteristics, and / or emotional information (such as pleasure, excitement, and stress levels). For example, this information can be collected via cameras within the control room.

[0099] Specifically, the status information of the multimodal operator includes, but is not limited to: Physiological state indicators: such as heart rate (HR), heart rate variability (HRV), and skin conductance (GSR); Behavioral characteristics: such as eye features (blinking frequency, PERCLOS duration of eye closure, direction of gaze), head posture, and the fluency and precision of operational behavior; Emotional State Index: A quantitative value reflecting the operator's current emotional dimensions, such as stress level, pleasure level, and arousal level, calculated by analyzing facial expressions, tone of voice, etc.

[0100] S1170: Real-time detection of operator status information using non-invasive sensing technology.

[0101] Capacitive or optical heart rate sensors are integrated into the steering wheel or seat of the driver's cab to continuously monitor the operator's heart rate (HR) and heart rate variability (HRV). HRV is an important indicator for assessing autonomic nervous system activity and reflecting fatigue and stress. It provides objective physiological evidence for determining the operator's fatigue and stress levels.

[0102] Utilize an operator status monitoring camera based on near-infrared technology installed in the cab. By tracking the operator's eye features (such as blink frequency, duration of eye closure, PERCLOS) and head posture (such as nodding, head-up angle), the camera analyzes their attention concentration and fatigue status in real time. Directly monitor the behavioral performances most relevant to driving operation safety, such as microsleep, distraction, etc.

[0103] The system background continuously records the operator's operation signals, such as the jitter frequency of the joystick, the delayed response time of the operation, non-instructional abnormal actions, etc. Deduce the operator's status from the operation results. Rough, delayed or chaotic operations are direct manifestations of the deteriorating status.

[0104] S1180 can be used to judge the adaptability between the working mode and the operator's status. The system compares the collected status information with the preset thresholds or models to make an adaptability judgment. The system has a built-in status evaluation model. This model fuses multi-dimensional data such as the above-mentioned heart rate, eye movement, head posture, operation behavior, etc., and calculates a comprehensive operator status score.

[0105] Exemplarily, good status (score ≥ C): The operator is concentrated and responsive. At this time, the system can adapt to the third mode or the second mode with high requirements for the operator to give full play to their subjective initiative. General status (D ≤ score < C): The operator shows signs of slight fatigue or distraction. The system should recommend or automatically switch to the second mode (human-machine collaboration), and use AI for assistance and verification to add an extra safety line. Poor status (score < D): The operator is significantly fatigued, yawning frequently, having drooping eyelids or an increasing number of operation errors. The system determines it as non-adaptable and needs to intervene immediately.

[0106] S1190 can include an intervention mechanism in case of non-adaptability. When the system determines that the current mode is not adaptable to the operator's status, it starts a hierarchical intervention strategy.

[0107] When it is detected that the status has declined to the "general" level, the system gives a prompt on the display screen with eye-catching icons and text (such as "Distracted, it is recommended to enable the assistance mode"), accompanied by a gentle prompt sound. At this time, it is still up to the operator to decide whether to follow the advice. Give the operator the opportunity to correct actively and respect their dominance, which is applicable to slight fluctuations in status.

[0108] When a "poor" status is detected, or if the operator fails to respond within a specified time (e.g., 10 seconds) after receiving a prompt and the status continues to deteriorate, the system will automatically switch modes. For example, it may forcibly downgrade from mode three (human-led) to mode two (human-machine collaboration), or even directly switch to mode one (fully automatic), displaying "Fatigue driving detected, automatic protection mode activated." In critical moments, the system takes over control, becoming the last line of defense for human and machine safety, effectively preventing accidents caused by personnel condition issues.

[0109] In some embodiments, the method further includes: When the excavator is operating in the first mode, it enters the first level of safety mode; When the excavator is operating in the second mode, it enters the first or second level of safety mode. When the excavator is operating in the third mode, it will enter the second or third level of safety mode. In the first-level safety mode, the excavator's sensor system is fully activated and information is acquired. The sensor system includes a vision subsystem and a non-vision subsystem. The sensor system is used to monitor the excavator's operation, and the acquired information is used for the excavator's safe operation. In the second level of safety mode, the excavator's sensor system is partially activated and information is collected. In the third-level safety mode, the excavator's sensor system is not activated and no information is collected. For example, in the third-level safety mode, the active data collection function of the sensor system is not activated, and only the basic alarm function (e.g., electronic fence function) is retained.

[0110] In some embodiments, the vision subsystem may include, but is not limited to: all cameras (wide-angle, telephoto, infrared) running at the highest frame rate (e.g., 30fps) to execute deep learning algorithms, identify and track the dynamics of surrounding people, vehicles, and obstacles in real time, and build a high-precision 3D map of the environment.

[0111] Non-visual subsystems may include, but are not limited to: lidar performing 360° high-speed rotating scans, millimeter-wave radar continuously monitoring moving targets at medium and long distances, and ultrasonic sensors forming a seamless anti-collision barrier in the near field of the fuselage.

[0112] In the first mode, the perception system is fully activated, and all perception data is integrated in the fusion processor to generate a unified, high-refresh-rate environmental safety situation map. Any potential collision risk will trigger the system to automatically execute avoidance actions (such as stopping or detouring). This provides the highest level of uninterrupted safety protection for unmanned autonomous operations, ensuring operational safety without human intervention.

[0113] In the second security level mode, the system may disable computationally intensive visual recognition algorithms, retaining only basic object detection functions. Alternatively, it may put the LiDAR into a low-power mode, reducing the scanning frequency and focusing scanning only on critical areas (such as the bucket's direction of movement).

[0114] In some embodiments, low-power devices such as ultrasonic sensors operate continuously, but higher-level sensors (such as lidar) are only momentarily awakened from sleep mode to full power when they detect a specific intention from the operator (such as turning around to get on the vehicle) or when the system receives an external trigger signal (such as a person approaching and triggering an alarm). This event-driven wake-up or driving method reduces the number of functions and power consumption that are activated. While ensuring the effectiveness of critical safety functions (such as near-field collision avoidance), it significantly reduces the overall system power consumption and the thermal load on the computing unit, making it suitable for scenarios requiring long-term continuous operation.

[0115] In Level 3 safety mode, only the most basic sensing functions are maintained, such as the rearview camera operating at a low frame rate, or only a few ultrasonic sensors at the rear of the vehicle are activated. For example, the system provides only the most critical alarms (such as GNSS-based geofence crossing alarms) rather than proactive intervention. This maximizes energy and computing resource conservation, prioritizing hardware lifespan and power for the excavation operation itself, while still providing a basic safety baseline.

[0116] In some embodiments, the guidance for the (n+1)th period includes at least one of the following: Bottom contact protection guidance is used to issue a bottom contact warning when a sudden increase in bucket resistance is detected and the spectral data shows characteristics consistent with rock impact, and to suggest reducing the digging depth or changing the digging angle. Slope trimming guide: When trimming slopes, the slope trimming guide guides the shovel tip to the design line. It also determines whether the surface needs to be leveled by brushing the slope based on the real-time terrain inversion, and provides the lateral swing amplitude and speed of the boom. Efficiency optimization guidance is used to determine the excavation angle and speed based on the detected soil characteristics with the goal of achieving the best efficiency.

[0117] In some embodiments, bottom-out protection guidance is a type of safety protection guidance. When the system determines from sensor data that the bucket may be about to violently collide with an insurmountable hard obstacle (such as bedrock or large boulders), it automatically generates avoidance instructions to prevent equipment overload or damage.

[0118] In some embodiments, slope trimming guidance is a precision and quality-oriented approach. In slope or plane trimming operations, the system not only guides the bucket to the designed position but also generates motion commands for fine-tuning based on real-time reflection of the flatness of the shaped surface.

[0119] In some embodiments, efficiency optimization guidance is a type of economy-oriented guidance. The system aims to maximize the workload per unit time (such as the volume of earthwork excavated) or minimize the energy consumption per unit workload, and dynamically recommends the optimal excavation parameters based on the perceived soil characteristics.

[0120] In some embodiments, the method further includes: Deeply integrate sensor data (vibration spectrum, drag data, GNSS / IMU positioning data). Every shovelful of work is treated as an in-situ geological test; The precise three-dimensional location of each shovel (from GNSS / IMU) is bound to high-precision soil property data (such as bearing capacity, internal friction angle, and rock hardness index) retrieved at that location to generate a geological map based on three-dimensional geographic coordinates. By utilizing geostatistical algorithms (such as Kriging interpolation), these discrete, high-precision "geological sampling points" are fused in real time to generate a continuous, high-resolution "micro-geological atlas" covering both the excavated and partially unexcavated areas. The dynamic generation of this atlas is analogous to a printer printing a high-precision image line by line. Thus, geological exploration is simultaneously conducted during construction, breaking the traditional sequential "exploration-design-construction" engineering process and achieving real-time synchronization between exploration and construction.

[0121] Furthermore, the method may also include: not only recording the geological conditions of the excavated areas, but also using the constructed maps and machine learning algorithms (such as spatiotemporal sequence prediction models) to predict the geological conditions of the unexcavated areas in front of the bucket. For example, if the system identifies that the hard rock layer currently being excavated extends forward at a certain angle, it can predict that the area ahead will remain hard rock for at least 2-3 cycles.

[0122] Based on the prediction results, the decision-making module no longer generates instructions for "the next shovel or the next cycle," but rather an optimal strategy sequence for "the next few shovels." For example, when it predicts that hard rock strata are about to be encountered, the system will generate an instruction sequence in advance: "Cycle n+1: Reduce feed speed and switch to hydraulic breaker mode; Cycle n+Y (where Y can be any positive integer): Continue low-speed breaking and reserve maximum impact energy for potentially harder rock masses." Alternatively, as more data is collected during the excavation process, the cycle length can be appropriately increased to reduce unnecessary data processing.

[0123] Furthermore, a damage accumulation model for excavators is established. For example, a mechanical damage accumulation model for key components (such as the bucket, bucket teeth, and hydraulic system) is created. This model quantifies the fatigue damage caused to the equipment during each operation under different geological conditions (such as cutting hard rock and excavating sand). The system optimizes with the global objective of "lowest total construction cost" or "shortest total construction period." When a high-strength rock layer is predicted ahead, the system may decide: "Although directly breaking hard rock has the shortest construction period, the equipment wear cost will surge. Based on global optimization, it is recommended to take a detour and excavate from a softer soil layer on the side. Although the path is slightly longer, the total cost is lower." This allows the excavator to "actively avoid" hard layers and even "choose" a longer operating path for the sake of equipment health. This completely changes the underlying logic of construction machinery as "obeying orders and working at full capacity," giving it the ability to "think economically" and "manage its own health," thus maximizing the value of the equipment's life cycle.

[0124] like Figure 4 As shown, this disclosure provides a satellite-based excavator guidance system, including: a satellite communication subsystem 2110, an inertial subsystem 2120, and a guidance subsystem 2130; the satellite communication subsystem 2110 is used to communicate with a satellite and acquire satellite data; the inertial subsystem 2120 includes at least one sensor; the at least one sensor is used for data sensing; the guidance subsystem 2130 is used to acquire reference data of a target area based on satellite data from the satellite system; the reference data includes macroscopic characteristic data of the target area; the macroscopic characteristic data includes at least surface data and / or basic geological data; first guidance information is generated based on the reference data and the equipment parameters of the excavator; the first guidance information is used to determine the excavation target and initial excavation calibration; and the excavation is carried out according to the excavation guidance of the nth cycle. During the process, sensors on the excavator are used to acquire sensing data for the nth cycle; the sensing data is used to describe the interaction between the excavator's bucket and the excavation area; n is a positive integer; when n equals 1, the excavator provides excavation guidance based on the first guidance information; based on the sensing data of the nth cycle, the microscopic characteristic data of the excavation area for the nth cycle are determined; the microscopic characteristic data of the nth cycle includes: soil characteristic data and hydrological characteristic data of the excavated area in the nth cycle; based on the microscopic characteristic data of the nth cycle, the excavation guidance of the nth cycle, the actual excavation conditions of the nth cycle, the excavation target, and the hardware parameters of the excavator, the excavation guidance for the (n+1)th cycle is generated; the excavation guidance for the (n+1)th cycle includes the excavator's travel route data and / or the bucket's excavation control parameters; excavation is carried out according to the excavation guidance for the (n+1)th cycle.

[0125] In some embodiments, the sensing data for the nth period includes spectral data and pressure data for the nth period; wherein, a plurality of pressure sensors are located on the inner surface of the bucket for detecting digging pressure; the guidance subsystem is configured to perform at least one of the following: acquiring digging resistance distribution data of the spectral data and pressure data for the nth period; determining soil properties data of the area being excavated in the nth period based on the spectral data and digging resistance distribution data; the soil properties data are used to indicate at least one of the following: the degree of soil looseness, whether a hard rock layer has been reached, and whether it is a waterlogged area.

[0126] In some embodiments, the spectrum data includes at least one of the following: the spectrum data is the spectrum data of audio data; the audio data is collected by audio sensors distributed on the outer surface of the bucket or the boom; the spectrum data is the spectrum data of vibration data; the vibration data is collected by vibration sensors distributed on the outer surface of the bucket or the boom.

[0127] In some embodiments, the guidance subsystem is specifically used to generate a first matrix by sequentially arranging the excavation guidelines for the (n+1)th period based on the micro-characteristic data of the nth period, the excavation guidelines of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator; inputting the first matrix into the AI ​​model to obtain a second matrix output by the AI ​​model; the second matrix corresponds to the excavation guidelines of the (n+1)th period.

[0128] In some embodiments, the excavator has a first mode, a second mode, and a third mode; specifically, when the excavator is operating in the first or second mode, during the generation of the digging guidance for the (n+1)th cycle, the AI ​​model generates a control signal based on a second matrix; when the excavator is operating in the first mode, the excavator is automatically controlled to dig in the (n+1)th cycle according to the control signal; when the excavator is operating in the second mode, after generating the control signal, it enters a preparatory state for automatic control, and when a definite input from the operator on the control lever is detected, it enters an execution state from the preparatory state, and in the execution state, it executes the digging for the (n+1)th cycle according to the control signal; when the excavator is operating in the third mode, the digging guidance for the (n+1)th cycle is displayed on the terminal device in the cab; and the digging for the (n+1)th cycle is performed based on the operator's input based on the digging guidance for the (n+1)th cycle.

[0129] In some embodiments, the guidance subsystem is further configured to detect the operator's status information; determine, based on the status information, whether the excavator's working mode is compatible with the operator's status; and when the excavator's working mode is incompatible with the operator's status, output an adjustment prompt or automatically adjust the excavator's working mode.

[0130] In some embodiments, the guidance subsystem is specifically configured to: enter a first-level safety mode when the excavator is operating in a first mode; enter a first-level or second-level safety mode when the excavator is operating in a second mode; enter a second-level or third-level safety mode when the excavator is operating in a third mode; in the first-level safety mode, fully activate the excavator's sensor system and acquire collected information; the sensor system includes a vision subsystem and a non-vision subsystem, and is used to monitor the excavator's operation; the collected information is used for the safe operation of the excavator; in the second-level safety mode, partially activate the excavator's sensor system and acquire collected information; in the third-level safety mode, not activate the excavator's sensor system and acquire collected information.

[0131] In some embodiments, the guidance for the (n+1)th cycle includes at least one of the following: bottoming protection guidance, which issues a bottoming warning when a sudden increase in bucket resistance is detected and the spectral data shows characteristics consistent with rock impact, and suggests reducing the digging depth or changing the digging angle; slope trimming guidance, which guides the shovel tip to the design line when trimming slopes, and also determines whether the surface needs to be leveled by a slope brushing action based on real-time inverted terrain, and provides the boom lateral swing amplitude and speed; and efficiency optimization guidance, which determines the digging angle and speed based on the detected soil characteristics with optimal efficiency as the optimization objective.

[0132] This disclosure provides a computer-readable storage medium storing computer-executable instructions; after being executed by a processor, the computer-executable instructions can implement the excavator guidance method based on a satellite system provided by any of the aforementioned technical solutions.

[0133] Combination Figure 5 As shown in the illustration, this application provides an electronic device that can be a component of a satellite-based excavator guidance system, including a processor 10 and a memory 11. Optionally, the device may further include a communication interface 12 and a bus 9. The processor 10, communication interface 12, and memory 11 can communicate with each other via the bus 9. The communication interface 12 can be used for information transmission. The processor 10 can call logical instructions in the memory 11 to execute the satellite-based excavator guidance method described in the above embodiment.

[0134] Furthermore, the logical instructions in the aforementioned memory 11 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0135] The memory 11, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this application. The processor 10 executes functional applications and data processing by running the program instructions / modules stored in the memory 11, thereby implementing the excavator guidance method based on the satellite system in the above embodiments.

[0136] The memory 11 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 11 may include high-speed random access memory and may also include non-volatile memory.

[0137] The technical solutions of this application embodiment can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method of this application embodiment. The aforementioned storage medium can be a non-transitory storage medium, including various media capable of storing program code such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks, or it can be a transient storage medium.

[0138] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0139] The embodiments or examples disclosed in this application are not exhaustive, but merely illustrative of some embodiments or examples, and are not intended to limit the scope of protection of this disclosure. Unless contradictory, each step in a particular embodiment or example can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment or example can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment or example can be arbitrarily interchanged. Furthermore, optional methods or examples in a particular embodiment or example can be arbitrarily combined; moreover, embodiments or examples can be arbitrarily combined. For example, some or all steps of different embodiments or examples can be arbitrarily combined, and a particular embodiment or example can be arbitrarily combined with optional methods or examples of other embodiments or examples.

[0140] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0141] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0142] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0143] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0144] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for guiding an excavator, characterized in that, The method includes: Based on satellite data from a satellite system, baseline data for the target area is acquired; the baseline data includes macroscopic characteristic data of the target area; the macroscopic characteristic data includes at least surface data and / or basic geological data. First guidance information is generated based on the benchmark data and the equipment parameters of the excavator; the first guidance information is used to determine the excavation target and the initial excavation calibration. During the digging process of the nth cycle according to the digging guidance of the nth cycle, the sensors on the excavator are used to acquire the sensing data of the nth cycle; the sensing data is used to describe the interaction between the bucket of the excavator and the digging area; n is a positive integer; when n equals 1, the excavator provides digging guidance according to the first guidance information; Based on the sensor data of the nth cycle, determine the microscopic characteristic data of the excavation area for the nth cycle; the microscopic characteristic data of the nth cycle includes: soil characteristic data and hydrological characteristic data of the excavation area for the nth cycle; Based on the microscopic characteristic data of the nth period, the excavation guidance of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator, the excavation guidance for the (n+1)th period is generated; the excavation guidance for the (n+1)th period includes the travel route data of the excavator and / or the excavation control parameters of the bucket. Mining is carried out according to the mining guidelines of the (n+1)th cycle.

2. The method according to claim 1, characterized in that, The sensing data for the nth period includes the spectral data and the pressure data for the nth period; wherein, multiple pressure sensors are located on the inner surface of the bucket for detecting digging pressure; based on the sensing data for the nth period, the microscopic characteristic data of the digging area for the nth period are determined, including at least one of the following: Obtain the spectral data of the nth period and the excavation resistance distribution data of the pressure data of the nth period; Based on the spectral data and the excavation resistance distribution data, the soil properties data of the area excavated in the nth cycle are determined; the soil properties data are used to indicate at least one of the following: the degree of soil looseness, whether it has entered a hard rock layer, and whether it is a waterlogged area.

3. The method according to claim 2, characterized in that, The spectrum data includes at least one of the following: The spectrum data is the spectrum data of audio data; the audio data is collected by audio sensors distributed on the outer surface of the bucket or on the boom. The spectral data is the spectral data of vibration data; the vibration data is collected by vibration sensors distributed on the outer surface of the bucket or on the boom.

4. The method according to any one of claims 1 to 3, characterized in that, Based on the microscopic characteristic data of the nth period, the excavation guidance of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator, the excavation guidance for the (n+1)th period is generated, including: Based on the microscopic characteristic data of the nth period, the excavation guidance of the nth period, the actual excavation status of the nth period, the excavation target, and the hardware parameters of the excavator, the excavation guidance of the (n+1)th period is generated and arranged in order to generate the first matrix. The first matrix is ​​input into the AI ​​model to obtain the second matrix output by the AI ​​model; the second matrix corresponds to the mining guidance of the (n+1)th cycle.

5. The method according to claim 4, characterized in that, The excavator has a first mode, a second mode, and a third mode; the excavation according to the excavation guide of the (n+1)th cycle includes: When the excavator is operating in the first mode or the second mode, during the process of generating the excavation guidance for the (n+1)th cycle, the AI ​​model generates control signals based on the second matrix. When the excavator is working in the first mode, the excavator is automatically controlled to dig in the (n+1)th cycle according to the control signal. When the excavator is operating in the second mode, it enters the automatic control preparatory state after generating the control signal, and enters the execution state from the preparatory state when the operator's determination input on the control lever is detected. In the execution state, the excavation of the (n+1)th cycle is executed according to the control signal. When the excavator is operating in the third mode, the digging instructions for the (n+1)th cycle are displayed on the terminal device in the cab; the digging for the (n+1)th cycle is performed based on the operator's input based on the digging instructions for the (n+1)th cycle.

6. The method according to claim 5, characterized in that, The method further includes: Detect operator status information; Based on the status information, determine whether the excavator's working mode is compatible with the operator's status; When the operating mode of the excavator is not compatible with the operator's state, an adjustment prompt is output or the operating mode of the excavator is automatically adjusted.

7. The method according to claim 5, characterized in that, The method further includes: When the excavator is operating in the first mode, it enters the first level of safety mode; When the excavator is operating in the second mode, it enters the first or second level of safety mode. When the excavator is operating in the third mode, it enters the second or third level of safety mode. In the first level of safety mode, the perception system equipped on the excavator is fully activated and information is acquired; the perception system includes a vision subsystem and a non-vision subsystem, and the perception system is used to monitor the operation of the excavator; the acquired information is used for the safe operation of the excavator. In the second level of security mode, the sensing system equipped on the excavator is partially activated and information is collected. In the third level of security mode, the active data acquisition function of the sensing system is not enabled, and only the basic alarm function is retained.

8. The method according to any one of claims 1 to 3, characterized in that, The guidance for the (n+1)th period includes at least one of the following: Bottom contact protection guidance is used to issue a bottom contact warning when a sudden increase in the resistance of the bucket is detected and the spectral data shows characteristics consistent with rock impact, and to suggest reducing the digging depth or changing the digging angle. The slope trimming guide is used to guide the shovel tip to the design line when trimming the slope. It also determines whether the surface needs to be leveled by brushing the slope based on the real-time terrain inversion, and provides the lateral swing amplitude and speed of the boom. Efficiency optimization guidance, which is used to determine the excavation angle and speed based on the detected soil characteristics with the goal of achieving the best efficiency.

9. A satellite-based excavator guidance system, characterized in that, include: The system comprises a satellite communication subsystem, an inertial subsystem, and a guidance subsystem; the satellite communication subsystem is used to communicate with the satellite and acquire satellite data. The inertial subsystem includes at least one sensor; the at least one sensor is used for data sensing. The guidance subsystem is used to acquire baseline data of the target area based on satellite data from the satellite system; the baseline data includes macroscopic characteristic data of the target area; the macroscopic characteristic data includes at least surface data and / or basic geological data. First guidance information is generated based on the benchmark data and the equipment parameters of the excavator; The first guidance information is used to determine the excavation target and initial excavation calibration; During the excavation of the nth cycle according to the excavation guide of the nth cycle, the sensors on the excavator are used to acquire the sensing data of the nth cycle; the sensing data is used to describe the interaction between the bucket of the excavator and the excavation area; n is a positive integer; When n equals 1, the excavator provides excavation guidance based on the first guidance information; Based on the sensing data of the nth period, determine the microscopic characteristic data of the excavation area for the nth period; The microscopic characteristic data for the nth cycle includes: soil characteristic data and hydrological characteristic data of the area excavated in the nth cycle; based on the microscopic characteristic data for the nth cycle, the excavation guide for the nth cycle, the actual excavation status for the nth cycle, the excavation target, and the hardware parameters of the excavator, the excavation guide for the (n+1)th cycle is generated; the excavation guide for the (n+1)th cycle includes the travel route data of the excavator and / or the excavation control parameters of the bucket; excavation is carried out according to the excavation guide for the (n+1)th cycle.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions; when executed by a processor, the computer-executable instructions are able to implement the method described in any one of claims 1 to 8.

Images