A method and system for evaluating and guiding driving behavior of mobile machinery

Abstract

The present invention discloses a method and system for evaluating and guiding driving behavior of mobile machinery, which belongs to the field of mobile machinery technology and is used to solve the technical problems that existing mobile machinery driving evaluation methods do not make full use of parameter characteristics, the evaluation process is rough, the specific operation cannot be associated with mechanical efficiency and energy consumption, and the influence of driving mode and working conditions on driving operation evaluation is not considered. The method comprises: collecting real-time operation data of each component of the target mobile machinery and performing data preprocessing; identifying the current operating scene of the target mobile machinery based on the real-time operation data; loading the corresponding standard operation model into a preset standard model library based on the current operating scene; extracting driver operation characteristics related to driving behavior from the real-time operation data; scoring each driving action of the current driver based on the driver operation characteristics and the standard operation model; generating targeted operation suggestions based on the calculated scores and feeding back the scores and operation suggestions to the driver.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile machinery, and in particular to a method and system for evaluating and guiding driving behavior of mobile machinery. Background Art

[0002] Mobile machinery refers to construction machinery equipped with mobility and capable of being moved between different work sites, including excavators, transport vehicles, and mobile cranes. In the mobile machinery sector, and particularly in mobile machinery operations, the driver's behavior has a crucial impact on the equipment's operating efficiency, fuel consumption, and service life. For example, the performance of an excavator driver has a significant impact on the excavator's fuel consumption, efficiency, and service life. Currently, various methods exist on the market for evaluating the operating behavior of mobile machinery drivers, but all suffer from varying degrees of deficiencies.

[0003] Early assessments of mobile machinery driver behavior primarily relied on simple counters to record the number and duration of operations, employed simple sensors to monitor key component operating parameters, or employed simple algorithms to evaluate partial operational data from a single sensor. These methods only captured surface data and failed to capture key factors such as pilot pressure changes and operational details. Engine speed is affected by a variety of factors, including the operating environment and the equipment's own operating status. Relying solely on this to assess driver behavior is highly limited and can mislead companies regarding driver management and training.

[0004] Furthermore, while existing machine operator evaluation methods can reflect a certain degree of performance, they are crude and incapable of analyzing every step and habit of the operator. For example, within a single operating cycle—excavation, boom lift and swing, unloading, and unloaded swing and reset—existing solutions can only provide a high-level assessment of the machine's material handling efficiency and overall energy consumption for one or more cycles, but fail to provide a detailed assessment of the optimal operation of each stage within each cycle. Furthermore, they are unable to effectively link dynamic operating behaviors with key indicators such as energy consumption and efficiency of mobile machinery, preventing dynamic, real-time evaluation and analysis, and consequently, providing optimal guidance for each individual operation.

[0005] In addition, in actual operations, different driving modes are suitable for different working conditions. Existing technologies have insufficient research on the relationship between driving modes and work intensity and actuator energy output requirements. They cannot conduct targeted evaluations of driver operations based on driving modes and actual working conditions, making it impossible for the evaluation results to accurately guide the driver to optimize operations.

[0006] In summary, existing methods for evaluating driving behavior of mobile machinery do not make sufficient use of pilot pressure characteristics. The evaluation process cannot be refined to each step of the operation, and it is impossible to establish a correlation between specific operations and mechanical efficiency and energy consumption. Moreover, the impact of driving mode and working conditions on driving operation evaluation is not considered. Therefore, it is impossible to obtain more representative evaluation results and better real-time guidance suggestions. Summary of the Invention

[0007] An embodiment of the present invention provides a method and system for evaluating and guiding the driving behavior of mobile machinery, which is used to solve the following technical problems: the existing mobile machinery driving behavior evaluation method does not make sufficient use of parameter characteristics, the evaluation process is rough, and it is impossible to establish a correlation between specific operations and mechanical efficiency and energy consumption. It does not consider the impact of driving mode and working conditions on driving operation evaluation.

[0008] The embodiment of the present invention adopts the following technical solutions:

[0009] In one aspect, an embodiment of the present invention provides a method for evaluating and guiding driving behavior of a mobile machine, the method comprising: collecting real-time operating data of various components of a target mobile machine and performing data preprocessing; wherein the real-time operating data includes at least pilot pressure data of a pilot handle;

[0010] identifying a current operating scenario of the target mobile machinery based on the real-time operating data;

[0011] Based on the current operation scenario, a corresponding standard operation model is loaded from a preset standard model library;

[0012] extracting driver operation features related to driving behavior from the real-time operation data;

[0013] scoring each driving action of the current driver based on the driver's operating characteristics and the standard operating model;

[0014] Based on the calculated score, targeted operational recommendations are generated, and the score and operational recommendations are fed back to the driver.

[0015] In a feasible implementation, collecting real-time operating data of various components of the target mobile machinery and performing data preprocessing specifically includes:

[0016] During the operation of the target mobile machinery, collecting pilot pressure data of each pressure channel of the left and right pilot handles;

[0017] Collecting motion data of each movable component; wherein the motion data at least includes joint angle, motion angular velocity, motion acceleration, rotational angular velocity and rotational acceleration;

[0018] Collecting energy efficiency data of the target mobile machinery; wherein the energy efficiency data includes at least instantaneous fuel consumption, instantaneous engine torque, hydraulic pump output flow, cylinder inlet and outlet pressures, and swing motor inlet and outlet flow and inlet and outlet pressures;

[0019] The pilot pressure data, the motion data and the energy efficiency data are combined into the real-time operation data, and the real-time operation data is preprocessed based on a preset preprocessing program through an edge computing unit deployed on the target mobile machinery.

[0020] In a feasible implementation, identifying the current operating scenario of the target mobile machinery based on the real-time operating data specifically includes:

[0021] Extracting pressure variation characteristics based on the pilot pressure data; wherein the pressure variation characteristics at least include: sequence characteristics and time interval characteristics of the pilot pressure changes between the pressure channels;

[0022] Extracting motion characteristics of each component based on the motion data; wherein the motion characteristics of each component include at least: the range of motion angle variation, angular velocity peak value, and motion synchronization characteristics of each component;

[0023] Obtaining geographic location information of the target mobile machinery and extracting geographic environment features;

[0024] Fusing the pressure change characteristics, the movement characteristics of each component, and the geographical environment characteristics to obtain a comprehensive scene feature;

[0025] Matching the comprehensive scene features with the preset features of each operation scene, and determining the operation scene with a matching degree exceeding a preset threshold as the current operation scene;

[0026] If the matching degree does not reach the preset threshold, the recognition cycle continues.

[0027] In a feasible implementation, before loading the corresponding standard operation model from a preset standard model library based on the current operation scenario, the method further includes:

[0028] Based on the operational factors affecting each operation scenario, and with the goal of achieving the optimal combination of energy consumption and efficiency for each operation, the theoretical operating parameters for each operation in each operation scenario are calculated.

[0029] Determining a scoring weight for each theoretical operating parameter based on a relationship between the theoretical operating parameter and the corresponding energy consumption and work efficiency, and constructing a theoretical operating model based on the theoretical operating parameters and the scoring weight;

[0030] Collecting actual operation data of senior drivers in each operation scenario, and extracting empirical operation parameters of each operation action from the actual operation data;

[0031] Determining a scoring weight for each empirical operating parameter based on a relationship between the empirical operating parameter and the corresponding energy consumption and work efficiency, and constructing an empirical operating model based on the empirical operating parameters and the scoring weight;

[0032] The theoretical operation model and the empirical operation model of each operation action in each operation scenario are associated into a group and stored in a database to construct the standard model library.

[0033] In a feasible implementation manner, extracting driver operation features related to driving behavior from the real-time operation data specifically includes:

[0034] Extracting the pilot handle control characteristics of the driver based on the pilot pressure data; wherein the pilot handle control characteristics include at least: time proportion characteristics of the pilot pressure data in different amplitude regions, loading characteristics and unloading characteristics of the pilot pressure monitoring data;

[0035] Extracting the driver's component operation characteristics based on the motion data of each movable component of the target mobile machinery; wherein the component operation characteristics include at least: motion trajectory characteristics of each component, motion speed change characteristics, and operation coordination characteristics;

[0036] Extracting the driver's operating energy consumption characteristics based on the target mobile machinery's energy efficiency data; wherein the operating energy consumption characteristics include at least: energy consumption per unit workload and energy consumption distribution characteristics at different operating stages;

[0037] The extracted pilot handle control features, component operation features, and operation energy consumption features are summarized and organized to form a complete feature data set to obtain the driver operation features.

[0038] In a feasible implementation, extracting the pilot handle control feature of the driver based on the pilot pressure data specifically includes:

[0039] Based on a preset amplitude region division rule, the pilot pressure data is divided into different amplitude regions; wherein the amplitude regions include at least a dead zone, a control zone, and a saturation zone;

[0040] monitoring the duration of the pilot pressure data in different amplitude regions, and calculating the time proportion of each amplitude region to obtain the time proportion feature;

[0041] Extracting the loading time and loading change rate of the pilot pressure monitoring data according to the change time between the initial value and the maximum value of the pilot pressure, and determining the maximum value of the loading change rate and its corresponding time position to obtain the loading feature;

[0042] At the same time, the unloading time and the unloading change rate of the pilot pressure monitoring data are extracted, and the maximum value of the unloading change rate and its corresponding time position are determined to obtain the unloading feature.

[0043] In a feasible implementation, based on the driver's operating characteristics and the standard operating model, each driving action of the current driver is scored, specifically including:

[0044] inputting the extracted driver operation characteristics into the theoretical operation model and the empirical operation model in the standard operation model respectively;

[0045] Comparing the driver's operating characteristics with the theoretical operating parameters in the theoretical operating model, performing a weighted scoring calculation based on the comparison result and the scoring weight of each theoretical operating parameter, and obtaining a first scoring result;

[0046] Comparing the driver's operating characteristics with the empirical operating parameters in the empirical operating model, and performing a weighted scoring calculation based on the comparison result and the scoring weight of each empirical operating parameter to obtain a second scoring result;

[0047] The first scoring result and the second scoring result are weighted and summed according to the preset model weight to obtain the final score.

[0048] In a feasible implementation manner, after scoring each driving action of the current driver based on the driver's operating characteristics and the standard operating model, the method further includes:

[0049] The collected real-time operation data, processed feature data, final scores, and operation suggestions are classified and stored in a distributed data storage system;

[0050] Regularly back up stored data and perform regular maintenance on distributed data storage systems;

[0051] Extracting the operating data and corresponding scoring results within a period of time from the distributed data storage system according to a preset time period;

[0052] Input the extracted running data into the current scoring model, compare the scoring results of the current scoring model with the stored scoring results, and analyze the model prediction error;

[0053] According to the model prediction error, the parameters of the current scoring model are adjusted to optimize the adaptability of the current scoring model to different working conditions and driver operating styles.

[0054] In a feasible embodiment, the method further includes:

[0055] Real-time monitoring of sensor data of target mobile machinery, extracting data anomaly features from the sensor data using a data anomaly detection algorithm, and determining whether a sensor failure has occurred based on the data features; wherein the data anomaly features include at least: data anomaly range features, data mutation features, and data correlation features;

[0056] If a sensor fails, the redundant sensor of the sensor is activated to continue data collection;

[0057] Perform fault diagnosis on faulty sensors and send alarm information to inform the location of the faulty sensor and possible fault causes;

[0058] Data interpolation or prediction algorithms are used to fill in missing or abnormal data during the fault period.

[0059] On the other hand, an embodiment of the present invention further provides a mobile machinery driving behavior evaluation and guidance system, characterized in that the system includes:

[0060] A data acquisition module, configured to acquire real-time operating data of various components of the target mobile machinery and perform data preprocessing; wherein the real-time operating data at least includes pilot pressure data of the pilot handle;

[0061] a scoring calculation module configured to identify a current operating scenario of the target mobile machinery based on the real-time operating data; load a corresponding standard operating model from a preset standard model library based on the current operating scenario; extract driver operating characteristics related to driving behavior from the real-time operating data; and score each driving action of the current driver based on the driver operating characteristics and the standard operating model;

[0062] The decision-making and feedback module is used to generate targeted operational suggestions based on the calculated scores and to feed back the scores and operational suggestions to the driver.

[0063] Compared with the prior art, the method and system for evaluating and guiding driving behavior of mobile machinery provided by the embodiments of the present invention have the following beneficial effects:

[0064] 1. Comprehensive Driver Performance Assessment: By collecting multi-dimensional data, including operational characteristics, actuator motion data, and energy efficiency data, this method comprehensively and meticulously correlates driver operations with overall vehicle performance. Compared to existing technologies that rely solely on single or small amounts of data, this method analyzes driver behavior from multiple perspectives, more accurately assessing driver performance and providing a robust basis for developing targeted improvement recommendations.

[0065] 2. Micro- and Macro-Integrated Driving Characteristic Analysis: This analysis analyzes pilot handle driving characteristics from both a microscopic (four-dimensional analysis of each pilot pressure) and macroscopic (complex action matrix) perspective. This not only provides a deep understanding of the impact of each operational detail on driving behavior, but also captures the overall characteristics of complex operating situations. This comprehensive analysis provides a more accurate assessment of the driver's operating style and skill, helping to identify potential operational issues and providing more precise guidance for improvement.

[0066] 3. Precise scenario-based scoring and operational guidance: Optimal path templates and skilled operator driving templates are established for different operational scenarios as the basis for scoring. Scoring is centered around energy consumption and efficiency, with detailed operational guidance for each action. Compared to existing technologies, this invention provides drivers with operational recommendations that better meet their actual needs based on the characteristics of actual operational scenarios. This helps drivers optimize their operations in each specific scenario, effectively improving operational efficiency and reducing energy consumption – a significant advantage unmatched by existing technologies.

[0067] 4. Quantitative Comprehensive Scoring System: By introducing a comprehensive scoring formula, a quantitative mathematical relationship is established between energy consumption, efficiency, and driving characteristics in the time domain, making the scoring more scientific and accurate. This quantitative scoring system can more intuitively reflect the differences between driver operation and standard templates, providing drivers with more targeted improvement directions, further enhancing the value of this invention in practical applications.

[0068] In summary, the present invention provides a scientific, objective, and comprehensive method for scoring the driving behavior of mobile machinery. Through multi-sensor information fusion, it conducts an in-depth analysis of various parameter characteristics of mobile machinery, especially the pilot pressure characteristics, which makes up for the defect of insufficient utilization of pilot pressure characteristics in the existing technology.

[0069] Furthermore, the present invention combines model-based calculation of optimal driving operational characteristics with the extraction of operational characteristics of skilled machine operators, enabling accurate assessment of driver operational behavior. Furthermore, the scoring model effectively correlates operational parameters with the energy consumption and efficiency of mobile machinery. This allows for effective and timely evaluation of each driver's operation during operation, providing real-time guidance and recommendations for optimal operation. This demonstrates that the present invention can help businesses improve operational efficiency, reduce fuel consumption, extend equipment life, and enhance equipment management. It also provides drivers with targeted improvement recommendations, promoting the improvement of their operational skills. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0071] Figure 1 A flow chart of a method for evaluating and guiding driving behavior of mobile machinery provided by an embodiment of the present invention;

[0072] Figure 2 A schematic structural diagram of a mobile machinery driving behavior evaluation and guidance system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0073] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0074] The embodiment of the present invention provides a method for evaluating and guiding driving behavior of mobile machinery, such as Figure 1 As shown, the method for evaluating and guiding driving behavior of mobile machinery specifically includes steps S101-S106:

[0075] S101. Collect real-time operation data of various components of the target mobile machinery and perform data preprocessing; wherein the real-time operation data at least includes pilot pressure data of the pilot handle.

[0076] Specifically, during the operation of the target mobile machinery, pilot pressure data of each pressure channel of the left and right pilot handles are collected.

[0077] And collect motion data of each movable component; wherein the motion data at least includes joint angle, motion angular velocity, motion acceleration, rotational angular velocity and rotational acceleration.

[0078] and collecting energy efficiency data of the target mobile machinery; wherein the energy efficiency data at least includes instantaneous fuel consumption, instantaneous engine torque, hydraulic pump output flow, cylinder inlet and outlet pressures, and the inlet and outlet flow and inlet and outlet pressures of the rotary motor.

[0079] The pilot pressure data, motion data and energy efficiency data are combined into real-time operation data, and the real-time operation data is preprocessed based on a preset preprocessing program through the edge computing unit deployed on the target mobile machinery.

[0080] As a feasible implementation method, multi-dimensional data collection mainly includes: 1) Driver operation characteristic signal collection: using high-precision pressure sensors to synchronously collect the time domain pressure signals P of the 12 pressure channels of the left and right pilot handles of the target mobile machinery i Here, i represents the i-th pressure channel. The 12 pressure channels include boom raising, boom lowering, arm retraction, arm extension, bucket retraction, bucket extension, left slewing, right slewing, left forward, left backward, right forward, and right backward.

[0081] 2) Mobile machinery motion characteristic signal acquisition: Taking the IMU sensor as an example (displacement sensor can also be used): the angles of the boom joint, arm joint, and bucket joint are collected through the IMU unit θ arm-i , angular velocity ω arm-i and acceleration a arm-i , rotational angular velocity ω rotation , angular velocity ω rotation and acceleration a rotation (Here, i represents the i-th joint) and is used to correspond to the driver's operating characteristics. For example, after violent driving with a wide opening, the large moment of inertia of the upper mechanism causes no movement, and the hydraulic oil output by the hydraulic pump is lost through overflow, which is detrimental to energy efficiency. This reflects the actual movement of the actuator due to inertia and other factors after the driver's operation.

[0082] 3) Energy efficiency characteristic data collection of the whole machine: collect the instantaneous fuel consumption Q(t) of the mobile machinery, the instantaneous torque T(t) output by the engine, and the output flow rate Q(t) of the hydraulic pump. pump (t), pressure signal P pump (t), the inlet and outlet pressure signals P of the boom cylinder, bucket cylinder, and stick cylindercylinder-in-i (t), P cylinder-out-i (t) (where i represents the i-th cylinder), the inlet flow rate Q of the rotary motor motor-in (t), inlet pressure P motor-in (t), outlet flow Q motor-out (t), outlet pressure P motor-out (t). Through this data, the driver's operating characteristics are carefully correlated with the energy consumption and efficiency characteristics of the entire machine, thereby comprehensively evaluating the driver's level and making well-founded improvement suggestions.

[0083] S102: Identify the current operation scene of the target mobile machinery based on the real-time operation data.

[0084] Specifically, based on the pilot pressure data, pressure variation characteristics are extracted; wherein the pressure variation characteristics at least include: sequence characteristics and time interval characteristics of the pilot pressure changes between the pressure channels.

[0085] Based on the motion data, the motion characteristics of each component are extracted; wherein the motion characteristics of each component at least include: the range of motion angle variation, the peak angular velocity, and the synchronization characteristics of the motion of each component.

[0086] Obtain geographic location information of the target mobile machinery and extract geographic environment features;

[0087] Finally, the pressure change characteristics, component motion characteristics, and geographic environment characteristics are fused to generate a comprehensive scene feature. This comprehensive scene feature is then matched with the preset characteristics of each operation scene. The operation scene with a matching degree exceeding a preset threshold is determined as the current operation scene. If the matching degree does not reach the preset threshold, the recognition cycle continues.

[0088] S103. Based on the current operation scenario, load the corresponding standard operation model in the preset standard model library.

[0089] Specifically, based on the operational factors affecting each operating scenario, and aiming for optimal energy consumption and efficiency for each action, the theoretical operating parameters for each operating action are calculated. Based on the relationship between the theoretical operating parameters and their corresponding energy consumption and work efficiency, scoring weights are determined for each theoretical operating parameter. Based on these theoretical operating parameters and scoring weights, a theoretical operating model is constructed.

[0090] Furthermore, the system collects actual operational data from senior drivers in each operating scenario and extracts empirical operational parameters for each action from this data. Based on the relationship between these empirical operational parameters and their corresponding energy consumption and work efficiency, it determines scoring weights for each empirical operational parameter. Based on these empirical operational parameters and their scoring weights, it constructs an empirical operational model.

[0091] Furthermore, the theoretical operation model and empirical operation model for each operation action in each work scenario are associated as a group and stored in the database to build a standard model library. In actual use, based on the current work scenario identified, the corresponding standard operation model is matched and loaded from the pre-set standard model library.

[0092] As a feasible implementation, taking excavators as an example, the present invention provides two standard operating models. One is for different scenarios, such as leveling, earthwork, trenching, slope leveling, and hammering. By establishing a simulation model, the optimal operating parameters for driving maneuvers in each scenario are calculated, with the goal of achieving the best combination of energy consumption E and efficiency F. The second is to extract driving characteristics from a large number of experienced excavator operators, including pilot pressure variation patterns and component motion parameters. These characteristics are then integrated to form empirical operating templates for different scenarios. Finally, the models are integrated for comprehensive scoring and operational guidance.

[0093] S104: Extracting driver operation features related to driving behavior from the real-time operation data.

[0094] Specifically, based on the pilot pressure data, the pilot handle control characteristics of the driver are extracted; wherein the pilot handle control characteristics include at least: the time proportion characteristics of the pilot pressure data in different amplitude regions, the loading characteristics and unloading characteristics of the pilot pressure monitoring data.

[0095] Furthermore, based on the motion data of each movable component of the target mobile machinery, the driver's component operation characteristics are extracted; wherein the component operation characteristics at least include: each component's motion trajectory characteristics, motion speed change characteristics, and operation coordination characteristics.

[0096] Furthermore, based on the energy efficiency data of the target mobile machinery, the driver's operating energy consumption characteristics are extracted; wherein the operating energy consumption characteristics at least include: energy consumption per unit workload and energy consumption distribution characteristics in different operating stages.

[0097] Finally, the extracted pilot handle control features, component operation features, and operation energy consumption features are summarized and organized to form a complete feature data set, and the driver operation features are obtained.

[0098] As a feasible implementation method, based on the pilot pressure data, the pilot handle control characteristics of the driver are extracted, specifically including:

[0099] Based on preset amplitude region division rules, the pilot pressure data is divided into different amplitude regions, where the amplitude regions include at least a dead zone, a control zone, and a saturation zone. The duration of the pilot pressure data in different amplitude regions is monitored, and the time proportion of each amplitude region is calculated to obtain the time proportion feature.

[0100] Based on the change time between the initial and maximum pilot pressure values, the loading time and loading change rate of the pilot pressure monitoring data are extracted. The maximum loading change rate and its corresponding time position are determined to obtain the loading signature. Simultaneously, the unloading time and unloading change rate of the pilot pressure monitoring data are extracted. The maximum unloading change rate and its corresponding time position are determined to obtain the unloading signature.

[0101] As a feasible implementation method, the mathematical model for extracting the pilot handle driving features is as follows:

[0102] 1) Duration of a single operation:

[0103] A complete operation process is when the pilot handle is opened, returns to the middle position, and the pilot pressure drives the multi-way valve to control the valve core to complete opening and closing once.

[0104] The duration of a single operation of the pilot pressure of the i-th channel is expressed as: T duration-i =t end-i -t start-i Among them, t start-i is the moment when the pilot pressure of the i-th channel begins to change, t end-i is the time when the pilot pressure of the i-th circuit returns to its initial state (neutral control). This formula is used to describe the duration of the actuator action required during the actual operation process and reflects the driver's operation time for the action.

[0105] 2) Pilot system pressure amplitude distribution during a single operation:

[0106] For the pilot pressure of the i-th channel, let the time when the pressure amplitude is in the dead zone (Pi<6bar) be t dead-i The time in the control area (6bar≤Pi≤23bar) is t control-i The time in the saturation zone (Pi>23bar) is t saturation-i , a single complete operation cycle time is T cycle-i .

[0107] The distribution ratios of pressure amplitude in each area are:

[0108] Dead time ratio: R dead-i =T dead-i ×100% / T cycle-i .

[0109] Control zone time percentage: R control-i =T control-i ×100% / T cycle-i .

[0110] Saturation time ratio: R saturation-i =T saturation-i ×100% / Tcycle-i .

[0111] The above formula is used to reflect the maintenance status of the pilot handle operation process in different pressure control zones during actual operation, and reflects the driver's operating habits under different operation intensity requirements.

[0112] 3) Pilot pressure loading / unloading time characteristics during a single operation:

[0113] For the pilot pressure of the i-th line, the pilot pressure loading time T load-i The time for the pressure to rise from the initial value to the maximum value, the pilot pressure unloading time T unload-i It is the time for the pressure to drop from the maximum value to the initial value. load-i =t max-i -t init-i ;T unload-i =t end-max-i -t max-i ; where t init-i is the moment when the pressure starts to rise, t max-i is the moment when the pressure reaches its maximum value, t end-max-i The time when the pressure starts to drop from the maximum value to the initial value. These two formulas are used to reflect the intensity of the pilot handle opening and resetting operation in the actual operation process. The shorter the loading / unloading time, the more intense the operation.

[0114] 4) Maximum value of the pilot pressure loading / unloading rate of change during a single operation and its time position parameters during the operation:

[0115] For the pilot pressure of the i-th line, the pilot pressure loading change rate r load-i (t) is the pressure change per unit time during the pressure rise process, and the pilot pressure unloading change rate r unload-i (t) is the pressure change per unit time during the pressure drop process. load-i (t)=dP i (t) / dt; here, t is the time of pressure rise. unload-i (t)=dP i (t) / dt; here, t is the time of pressure drop stage.

[0116] Maximum load change rate r load-max-i =max{r load-i (t)}, the corresponding time position t load-max-pos-i is the moment when the maximum value appears; the maximum value of the unloading change rate r unload-max-i =max{r unload-i (t)}, the corresponding time position t unload-max-pos-iThe maximum value is the moment when the value is reached. These parameters are used to reflect the intensity of the driving style during the pilot handle operation in the actual operation process. The larger the maximum value of the rate of change, the more intense the operation style.

[0117] 5) Compound action characteristic parameters (12-way pilot pressure linkage):

[0118] Considering the compound action formed by some simultaneous linkage operations in the 12-way pilot pressure, it is assumed that these linkage pilot pressures form a matrix M, where M ij Represents the relevant parameters of the pilot pressure of the i-th channel in the j-th operation stage. The characteristics of the compound action can be analyzed by establishing a matrix operation model. For example, the comprehensive strength index of the compound action is calculated as follows: Where m and n are the number of rows and columns of matrix M, respectively, and w ij is a weighting factor, set according to the importance of different pilot pressures in complex maneuvers. This way, the driver's operation is scored from the perspective of complex maneuvers, more comprehensively reflecting the driving characteristics in complex maneuvers.

[0119] S105 : Score each driving action of the current driver based on the driver's operating characteristics and the standard operating model.

[0120] Specifically, the extracted driver operation characteristics are input into the theoretical operation model and the experience operation model in the standard operation model respectively.

[0121] Furthermore, the driver's operating characteristics are compared with the theoretical operating parameters in the theoretical operating model, and a weighted scoring calculation is performed based on the comparison result and the scoring weight of each theoretical operating parameter to obtain a first scoring result.

[0122] The driver's operating characteristics are compared with the experiential operating parameters in the experiential operating model, and a weighted scoring calculation is performed based on the comparison result and the scoring weight of each experiential operating parameter to obtain a second scoring result.

[0123] Finally, the first scoring result and the second scoring result are weighted and summed according to the pre-set model weight to obtain the final score.

[0124] As a feasible implementation method, suppose that in a certain operation scenario, a complete operation cycle contains k actions, and the time of each action is t action-j (j=1,2,⋯,k), where j represents the action stage.

[0125] Then the total time of the operation cycle is T cycle-total for: .

[0126] The efficiency F is: .

[0127] The energy consumption E of stage j in time t is: .

[0128] To establish a mathematical correlation between driving characteristics, energy consumption, and efficiency in the time domain, this paper defines a comprehensive driving characteristic index C, which comprehensively considers factors such as pilot pressure characteristics, the motion characteristics of each component, and operating time. Taking pilot pressure as an example, the driving characteristic index of pilot pressure can be expressed as:

[0129] C pressure-i =α dead-i R dead-i +α control-i R control-i +α saturation-i R saturation-i .in, α dead-i 、 α control-i 、 α saturation-i They are the dead zone weight coefficient, control zone weight coefficient, and saturation zone weight coefficient, respectively. R dead-i 、 R control-i 、 R saturation-i They are the dead zone pilot pressure, the control zone pilot pressure, and the saturation zone pilot pressure, respectively. Similar processing is performed on other driving characteristic factors, and the comprehensive driving characteristic index C is finally obtained.

[0130] The comprehensive score S can be calculated using the following formula:

[0131] .

[0132] Among them, E model (01,02),F model (01,02), C model (01,02) are the standard values ​​of energy consumption, work efficiency and comprehensive driving characteristic indicators obtained based on the theoretical operation model or the empirical operation model respectively; 、 、 The weighting factor is set based on the importance of energy consumption, work efficiency, and driving characteristics in the scoring. This comprehensive scoring formula quantitatively compares the driver's actual operations in each specific scenario with the standard template, scoring each action and providing optimized operational guidance. For example, in an earthwork excavation scenario, if the driver's pilot pressure change in a particular excavation action differs from the optimal template and results in excessive energy consumption or low efficiency, the system can provide targeted adjustment suggestions, down to the optimization of each action.

[0133] S106: Generate targeted operational suggestions based on the calculated scores; and feed back the scores and operational suggestions to the driver.

[0134] Specifically, according to the calculated score, corresponding operation suggestions are matched, and the score and operation suggestions are fed back to the driver through a visual interface and / or voice.

[0135] Next, we will use a specific earthwork excavation scenario as an example to describe the implementation process of the above method in detail:

[0136] 1. Startup and Initialization: When the target mobile machine's engine is ignited, the mobile machine driving behavior evaluation and guidance system starts operating simultaneously. The system's hardware components complete power-on self-tests in sequence to ensure proper connectivity and functionality. Simultaneously, the system's software program loads, initializing the operating system, driver, and core algorithm modules of the scoring system, preparing for subsequent data collection and processing.

[0137] Sensor Initialization: High-precision pressure sensors installed on the 12 pressure channels of the left and right pilot handles of the target mobile machinery are set to operate at a sampling frequency of 100Hz. This high-frequency sampling accurately captures subtle, instantaneous changes in pilot pressure. For example, the rapid rise and fall of pilot pressure during operations such as digging and swinging can be recorded in detail, providing high-precision data support for subsequent analysis of the driver's operating force and rhythm. IMU sensors are located in the boom, arm, bucket joints, and swing mechanism, and are set to a sampling frequency of 100Hz. This frequency is sufficient to monitor the angle, angular velocity, and acceleration of each component in three dimensions in real time, accurately reproducing the motion trajectory and dynamic characteristics of each component. The engine oil system's instantaneous fuel consumption sensor, the engine output shaft torque sensor, and the flow and pressure sensors in the hydraulic pump, cylinder, and motor are each set to an appropriate sampling frequency based on the changing characteristics of their respective monitored parameters. For example, a fuel consumption sensor may sample at a lower frequency, such as 50 Hz, because changes in engine fuel consumption are relatively slow; while a hydraulic pump flow sensor may sample at a higher frequency, such as 100 Hz, to capture instantaneous changes in flow during rapid hydraulic system operation.

[0138] Data storage: Ensure that data is stored in an efficient binary format to reduce storage space and improve data read and write speeds. The storage path is set to a specific directory in the distributed file system to ensure data security and scalability. Furthermore, a separate storage subdirectory is allocated for each sensor's data, and an index file is created to facilitate subsequent data retrieval and management.

[0139] In the edge computing unit, the parameters of various data processing algorithms are initialized. For example, for the pressure data filtering algorithm, the filter window size is set to 5 to effectively remove high-frequency noise interference; for the IMU data fusion algorithm, the fusion coefficient of the acceleration and gyroscope data is adjusted to improve the accuracy of the attitude solution.

[0140] 2. Operation scene recognition

[0141] After the system starts, sensors begin collecting data in real time. This data undergoes preliminary signal conditioning and format conversion in the edge computing unit before being transmitted to the central processing unit. The central processing unit uses complex pattern recognition algorithms to analyze the data from multiple dimensions to identify the current operating scenario.

[0142] Pilot pressure signal analysis: The pilot pressure signals for earthmoving operations exhibit unique patterns. For example, at the start of excavation, the arm pilot pressure rises rapidly, and the boom pilot pressure adjusts accordingly to control excavation depth. The system analyzes the rising and falling edges, amplitude range, and the sequence and time intervals of pressure changes across the 12 pilot pressure channels to determine whether the operation pattern matches the earthmoving operation pattern.

[0143] Analysis of the motion characteristics of each component: The system combines motion data from the boom, arm, bucket joint, and slewing mechanism collected by the IMU sensors to analyze the motion trajectory, velocity changes, and coordinated motion relationships of each component. During excavation, the movements of the boom, arm, and bucket coordinate with each other in a specific sequence and rhythm. The system further confirms whether the scenario is an excavation scenario by detecting characteristics such as the angular range of movement, peak angular velocity, and synchronization of movement of each component. For example, when the arm is extended and inserted into the soil, the boom will lift appropriately to maintain balance. The system can assist in this judgment by analyzing the time synchronization and angular coordination of these two movements.

[0144] Utilizing Geographic Location and Environmental Information: If the target mobile machine is equipped with GPS and other geolocation sensors, as well as operating environment sensors, the system will use this information to assist in determining the operating scenario. For example, if GPS data indicates the target mobile machine is located in a designated excavation area on a construction site, and the values ​​detected by the operating environment sensors are within the typical range for earthwork excavation, this will increase confidence that the current scenario is earthwork excavation.

[0145] Comprehensive Judgment and Decision-Making: The system uses comprehensive judgment logic to integrate and analyze the aforementioned features. It uses machine learning algorithms (such as support vector machines) to classify and judge the collected data features. When the degree of match between each feature and the earthwork excavation scenario exceeds a set threshold, the current operation scenario is determined to be earthwork excavation, and the earthwork excavation scenario data collection process is initiated. If the match does not reach the threshold, the identification cycle continues.

[0146] 3. Earthwork excavation scene data collection and preprocessing:

[0147] Data Collection: After confirming that the excavation scenario is real-time, each sensor begins synchronous data collection at a set frequency. High-precision pressure sensors monitor the 12 pilot pressure channels in real time. Each sampling point accurately records the pilot pressure value at a specific moment. These continuous sampling points form a curve showing the change in pilot pressure over time. For example, in a typical excavation operation, the pilot pressure curve will show a process of gradually rising from the initial value, passing through the control zone to the saturation zone, and then gradually falling back to the initial value as the excavation operation ends. The pressure sensor can accurately capture every detail of the pressure change during this complete process.

[0148] IMU sensors capture the angles, angular velocities, and accelerations of various components. This data not only depicts the spatial trajectory of each mechanism in a mobile machine but also reveals changes in velocity and acceleration. For example, IMU sensor data can accurately analyze the angular rate of change of the boom during lifting and the acceleration of the boom during extension and retraction, thereby understanding the smoothness and coordination of the operator's operation.

[0149] Various sensors installed in key locations, such as the engine oil system, output shaft, and hydraulic pumps, cylinders, and motors, collect relevant data at their assigned frequencies. Fuel consumption sensors measure the engine's instantaneous fuel consumption in real time, accurately recording the amount of fuel consumed, providing essential data for analyzing energy consumption during excavation operations. Torque sensors monitor the engine's instantaneous torque output, reflecting the engine's load under different operating conditions. Flow sensors and pressure sensors at the hydraulic pumps, cylinders, and motors, respectively, collect flow and pressure data from the hydraulic system. This data is crucial for understanding the hydraulic system's operating status and the transmission and distribution of power during excavation operations. For example, by analyzing changes in hydraulic pump flow, it is possible to understand the system's demand for hydraulic power during different operating phases.

[0150] Furthermore, the collected data is transmitted in real time to the edge computing unit via a data transmission line. During the data transmission process, data verification and error correction mechanisms are implemented to ensure data integrity and accuracy. For example, a cyclic redundancy check algorithm is used to verify each data packet. Upon receiving a data packet, the receiver recalculates the CRC value and compares it with the CRC value sent by the sender. If there is a mismatch, the packet is requested to be resent, thus preventing data transmission errors from affecting subsequent processing.

[0151] The edge computing unit preprocesses incoming data to improve data quality and reduce the burden on the central processing unit. For the pilot pressure data collected by the pressure sensor, the edge computing unit uses a median filter algorithm to remove sudden noise interference. This effectively removes abnormal pressure values ​​caused by electromagnetic interference or sensor noise, making the pressure data smoother and more accurately reflect the driver's operating intention. For the angle, angular velocity, and acceleration data collected by the IMU sensor, the edge computing unit uses a complementary filtering algorithm to fuse acceleration and gyroscope data to improve attitude resolution accuracy. Accelerometers measure an object's acceleration in a gravitational field, thereby obtaining attitude information, but are susceptible to vibration and shock during dynamic motion. Gyroscopes measure an object's angular velocity and can determine attitude changes through integration, but this can lead to cumulative errors over time. By rationally fusing the data from these two sensors, the complementary filtering algorithm leverages the accuracy of the accelerometer in the low-frequency band and the fast response of the gyroscope in the high-frequency band to more accurately calculate the attitude information of each component. For example, when calculating the boom angle, the complementary filtering algorithm is used to adjust the angle calculation result in real time based on the gravity component measured by the acceleration sensor and the angular velocity change measured by the gyroscope, making the obtained boom angle more accurate and reliable.

[0152] For data collected by energy consumption and power sensors, the edge computing unit detects and removes outliers based on the device's rated parameter range. For example, an engine's instantaneous fuel consumption should fluctuate within a certain range under normal operating conditions. If the collected fuel consumption value far exceeds or falls below this range, it may be an outlier caused by a sensor failure or data transmission error. The data is identified as an outlier and removed. Missing data points are filled in using linear interpolation or a prediction algorithm based on historical data to ensure data continuity. The preprocessed data is transmitted to the central processing unit via a high-speed data bus for further analysis.

[0153] 4. Feature extraction:

[0154] Data pre-processed by the edge computing unit is transmitted stably and quickly to the central processing unit via a high-speed data bus. During data transmission, data caching and flow control mechanisms are used to ensure orderly data transmission and avoid data loss or processing overload.

[0155] Based on the pilot pressure data, the central processing unit divides the amplitude area into the preset dead zone (0-6bar), control zone (6-23bar), and saturation zone (23-40bar), and calculates a series of key characteristic parameters to deeply characterize the driver's operating style.

[0156] During a complete operation, the central processing unit monitors the duration of the pilot pressure in different amplitude ranges and calculates the time percentage in each range. For example, for the boom pilot pressure, the entire process from the pressure starting to rise and leaving the dead zone to the pressure returning to the dead zone is considered a complete operation. During this process, the time the pressure is in the dead zone, the control zone, and the saturation zone is calculated. The dead zone time percentage, control zone time percentage, and saturation zone time percentage are then calculated separately. These time percentage parameters can reflect the driver's preference for using different pressure zones when operating the boom. For example, if the saturation zone time percentage is too high, it may indicate that the driver is applying excessive force during excavation, resulting in increased energy consumption and increased equipment wear.

[0157] The time it takes for the pilot pressure to rise from its initial value to its maximum value is recorded as the loading time, and the time it takes for the pilot pressure to fall from its maximum value to its initial value is recorded as the unloading time. For example, during a boom raising operation, the boom pilot pressure begins to rise from its initial value. The time difference between reaching the maximum pressure required to raise the boom is recorded as the loading time. When the boom is raised and the pilot pressure gradually falls back to its initial value, the time difference between this process is recorded as the unloading time. The loading and unloading times reflect the speed and rhythm of the operator's operation. A short loading time may indicate a hasty operation, which could cause damage to the equipment; a long unloading time may affect operational efficiency. The pilot pressure signal is differentiated during its rising and falling phases to obtain the loading and unloading pressure change rates. The maximum loading change rate and its corresponding time location are determined, as well as the maximum unloading change rate and its corresponding time location. For example, when a bucket is digging, the loading change rate of the pilot pressure reflects the speed and force of the bucket's insertion into the material. A larger maximum value indicates a more abrupt insertion, potentially causing unnecessary damage to the material and equipment. These rate-of-change parameters and their temporal location information can describe the driver's operating style and ability to control the equipment in more detail.

[0158] Furthermore, the central processing unit combines IMU data with energy consumption and power sensor data to extract characteristic parameters of various aspects such as movement and energy consumption, and comprehensively evaluates the driver's operating level.

[0159] Using the angle and angular velocity data collected by the IMU sensors, the central processing unit can reconstruct the spatial motion trajectory of components such as the boom, arm, and bucket. For example, a curve showing the change in boom angle over time during an excavation operation can be plotted to visually demonstrate the boom's motion. Simultaneously, the angular velocity data can be used to calculate the speed of each component at different times and analyze speed changes. For example, during arm extension, the change in arm angular velocity can be observed to determine whether the extension speed is uniform. Uneven speed changes can affect excavation smoothness and efficiency.

[0160] During excavation, boom raising and arm retraction should start and end at appropriate times to achieve efficient excavation. The system quantifies operational coordination by calculating metrics such as the time difference between boom and arm movements and the synchronization rate of angle changes. Prolonged delays in arm retraction after boom raising, or unsynchronized angle changes between the two, can reduce operational coordination and impact excavation efficiency and quality. Analysis of these coordination parameters accurately assesses the driver's level of coordination between various components during earthmoving operations.

[0161] By combining instantaneous fuel consumption data with workload-related parameters to calculate energy consumption per unit of workload, it is possible to monitor the energy consumption variations during each operational phase and link. By analyzing these energy consumption variations, the driver's operational energy consumption at each phase and link can be assessed, and then compared with the standard energy consumption to determine whether the operation is energy-efficient. If energy consumption is higher than the standard, it may indicate that the driver's operating method is wasting energy. The specific cause will be further analyzed and improvement suggestions will be provided. This method is more detailed than existing evaluation methods, and the resulting improvement suggestions are more targeted, directly indicating which operational actions and characteristics are inappropriate and need improvement.

[0162] Furthermore, the energy consumption distribution during different operation phases is analyzed. For example, for an excavator, the excavation operation is divided into the following phases: excavation, boom lift and swing, unloading, and unloaded swing and reset. The energy consumption percentage of each phase relative to the total energy consumption is calculated. This energy consumption distribution analysis can identify operational phases with high energy consumption and provide targeted optimization recommendations. For example, if energy consumption during the excavation phase is too high, it may be due to excessive excavation depth or speed, requiring adjustment of operating parameters to reduce energy consumption. Further analysis can also be conducted on the relationship between energy consumption and operational characteristics during each phase, such as the correlation between pilot pressure and energy consumption during the excavation phase. This provides a more specific basis for optimizing operations. For example, during the excavation phase, when the bucket is fully immersed in the soil, if the "opening characteristics" of the handles for operating the arm retracting and the handles for operating the bucket inward are very drastic, such as maximum opening or an excessive rate of change in the opening amplitude. At this point, violent operation can no longer accelerate the movement of the bucket and dipper arm, but should be assisted by boom lifting. This situation will be fed back to the system and reflected as high energy consumption and poor excavation efficiency of the excavator, thereby obtaining a guide for system evaluation and providing a specific basis for operation.

[0163] Finally, the extracted features, such as pilot pressure, motion, and energy consumption, are aggregated and organized into a complete feature dataset. After standardization and normalization, these features are output for subsequent scoring calculations. Standardization and normalization eliminate differences in dimensionality and numerical ranges between features, ensuring that each feature has the same weighting in the scoring calculation, improving the accuracy and fairness of the scoring.

[0164] 5. Scoring calculation matches earthwork excavation scoring model:

[0165] The scoring calculation module automatically matches the appropriate scoring model from a scoring model library based on the current earthwork excavation scenario. This library contains a variety of models. For earthwork excavation scenarios, scoring is performed using a combination of optimal operating models based on theoretical dynamics and kinematics, and empirical models derived from statistical analysis of data from a large number of experienced operators.

[0166] Models based on theoretical dynamics and kinematics are constructed based on the machine's physical structure, operating principles, and the mechanical properties of earthmoving. Through mathematical modeling, they determine the optimal operating parameters for each step of an earthmoving operation under ideal conditions, such as the optimal pilot pressure curve, the optimal speed and acceleration of each component's movement, and corresponding energy efficiency indicators. For example, using an excavator as an example, the optimal rate of increase in pilot pressure when the bucket arm penetrates the soil at a specific excavation depth is calculated based on parameters such as the excavator's bucket capacity and the mechanical properties of the soil, thereby achieving efficient excavation with minimal energy consumption.

[0167] The empirical model based on skilled operator data statistically analyzes a large number of actual operational data from experienced operators in earthmoving operations to extract representative operational characteristics and corresponding scoring criteria. For example, the range of pilot pressure variation, the appropriate range of component movement speeds, and operational coordination indicators under different excavation conditions are collated and summarized to form standard parameters for the empirical model.

[0168] At the same time, the relationship between these operational characteristics and the final operational results (such as excavation efficiency, energy consumption, and excavation quality) is analyzed to determine the weights of different operational characteristics in the scoring. Extracted parameters such as pilot pressure characteristics, motion characteristics, and energy consumption characteristics are input into a matching scoring model. The scoring model compares actual operating parameters with preset optimal or standard parameters, using a specific scoring algorithm to derive scores for each item. The final score is then calculated through weighted summation, comprehensively assessing the driver's operational proficiency in earthwork excavation scenarios.

[0169] For optimal operating models based on theoretical dynamics and kinematics, the actual pilot pressure and motion characteristics are compared with the optimal parameters calculated by the model. For example, if the model calculates that at a certain excavation depth, the arm pilot pressure should reasonably spend 60% of its time in the control zone, but in actual operation this percentage is only 50%, the pilot pressure characteristic's score for this item is calculated using a preset scoring algorithm. Similarly, various parameters such as motion characteristics and energy consumption characteristics are compared and scored. Each score is weighted according to its importance in the theoretical model, for example, a weight of 0.4 for the pilot pressure characteristic, 0.3 for the motion characteristic, and 0.3 for the energy consumption characteristic.

[0170] For empirical models based on skilled operator data, actual operating characteristics are matched against standard operating characteristics extracted from the empirical model. For example, the empirical model stipulates that during the excavation and insertion phase, the coordinated movement speed of the boom and arm should be within a specific range. The degree to which the coordinated movement speed of the boom and arm in actual operation matches this range determines the score for this item. By quantifying the degree of match between each operating characteristic and the empirical model standard, each score based on the empirical model is obtained. Each score is weighted according to its importance in the empirical model. For example, the weight of the operational coordination characteristic is 0.5, and the sum of the weights of other operating characteristics is 0.5.

[0171] The scores calculated based on the theoretical model and the empirical model are weighted and summed according to the pre-set model weights to obtain the final score.

[0172] Assume that the total score based on the theoretical model accounts for 0.6, and the total score based on the empirical model accounts for 0.4. Assume that the leading pressure characteristic score based on the theoretical model is S 11 , the motion feature score is S12 , the energy consumption characteristic score is S 13 ; The operational coordination feature score based on the empirical model is S 21 , other operating characteristics scores are S 22 , then the final score is (S=0.6×(0.4×S 11 +0.3×S 12 +0.3×S 13 )+0.4×(0.5×S 21 +0.5×S 22 This final score comprehensively reflects how close the driver's operating level in earthwork excavation scenarios is to optimal or skilled operation.

[0173] 6. Result feedback and storage

[0174] The decision-making and feedback module will provide the calculated scoring results and targeted operation suggestions to the driver through a visual interface or voice prompts.

[0175] Visual interface display: The cab is equipped with a high-resolution LCD screen, which displays the scoring results to the driver in the form of intuitive numbers and charts. For example, a circular progress bar is used to represent the overall score. The fill ratio of the progress bar corresponds to the score, and the specific score number is displayed next to it. At the same time, the scores of each feature are displayed in the form of a bar graph, so that the driver can clearly understand his performance in pilot pressure control, operation coordination, energy consumption control, etc. In addition, the average score comparison with the drivers of the same type of machinery and detailed scoring information at different operation stages will be displayed to help the driver have a more comprehensive understanding of his own operating level. The visual interface can also provide standard operation trajectory guidance for each step of the operation, such as the angle changes of each joint, the trajectory changes of the bucket teeth, etc., to achieve dynamic guidance of the actual trajectory and the optimal trajectory.

[0176] Generation and display of operational suggestions: Based on the scoring results, the system generates targeted operational suggestions. If the score is low and the energy consumption is too high, analysis shows that this is due to the pilot pressure staying in the saturated zone for too long. The system will prompt the driver on the display screen, "During the excavation process, try to minimize the time the pilot pressure stays in the saturated zone, and appropriately reduce the excavation force to reduce energy consumption." In the case of poor operational coordination, such as asynchronous movement of the boom and arm, the system will prompt, "Pay attention to the synchronization of the boom and arm movements, and maintain coordination when starting and stopping to improve operational efficiency." These operational suggestions not only point out the problem, but also provide specific improvement directions to help drivers improve their operating skills. The operational suggestions will be displayed in detail on the display screen in text form, accompanied by relevant animation demonstrations to intuitively show the correct operating methods.

[0177] Voice prompt assistance: In addition to the visual interface display, the system also uses voice prompts to convey scoring results and important operational suggestions to the driver in the form of voice. Taking an excavator as an example, after the excavator completes an excavation cycle, the system automatically announces, "This operation is scored 70 points. Please pay attention to energy consumption control. During the start-up phase of the rotation process, you stayed in the pilot pressure saturation zone for a long time. It is recommended to adjust the operating force appropriately." At this time, the wide-opening rotation operation is ineffective, and the system acceleration overflow loss is large, which is a point-deducting feature. Voice prompts allow the driver to obtain key information in a timely manner while focusing on operating the excavator. The volume, speed, and tone of the voice prompts can be set according to the driver's preferences to improve the user experience.

[0178] Furthermore, the collected raw data, processed feature data, scoring results, operation suggestions and other information are stored in a distributed data storage system for subsequent query, analysis and model optimization.

[0179] Distributed storage architecture: Data storage systems utilize a distributed storage architecture, consisting of multiple storage nodes distributed across different physical locations. This architecture improves data security and availability. Even if a storage node fails, other nodes can continue to store and provide data services. For example, distributed storage systems such as Ceph store data across multiple servers, ensuring data reliability through data redundancy and replication. Each storage node is equipped with high-speed hard drives and redundant power supplies to ensure fast data storage and retrieval.

[0180] Data classification storage: Different types of data are classified and stored. The original sensor data is organized and stored according to the acquisition time and sensor type to facilitate subsequent tracing and re-analysis. The processed feature data, such as the pilot pressure feature parameters, motion feature parameters, energy consumption feature parameters, etc., are stored in a special feature database and indexed and associated with the corresponding original data. The scoring results and operation suggestions are stored in the scoring database, and relevant information such as the scoring time and operation scene are also recorded. For example, the original pressure sensor data collected in the earth excavation scene at 10 am on June 1, 2025 is stored in a folder named with the date and time, and the corresponding pilot pressure feature parameters are stored in the corresponding table of the feature database and associated with the original data and scoring results through a unique job ID.

[0181] Stored data is regularly backed up to prevent data loss. Backup data is stored in an offsite data center to mitigate extreme situations such as natural disasters and hardware failures. Furthermore, the storage system undergoes regular maintenance, including data cleansing and index optimization, to ensure efficient data storage and fast retrieval. For example, a full backup is performed monthly, and incremental backups are performed weekly. The database is also reindexed and defragmented on monthly system maintenance days to improve data query efficiency. Furthermore, stored data undergoes quality checks, deleting duplicate or invalid data to ensure data accuracy and integrity.

[0182] 7. Scene switching detection and processing:

[0183] During operation, the system continuously monitors changes in sensor data characteristics and operating behavior patterns to determine whether a scene switch occurs.

[0184] The system uses the following methods to detect scene switching:

[0185] Operational behavior pattern analysis: By analyzing the changing patterns of the pilot pressure signal, the motion combinations of each actuator, and the operation frequency, the system determines whether the operational behavior pattern has changed significantly. For example, in an earthmoving scenario, the pilot pressure is primarily concentrated in channels related to actions such as excavation, lifting, and rotation, and the operation frequency is relatively high. In contrast, in a site leveling scenario, the pilot pressure is more closely related to horizontal adjustment of the boom and minor movements of the bucket, and the operation frequency is relatively low. If the system detects a significant change in these characteristics, it may indicate a scenario switch.

[0186] Changes in sensor data characteristics: This section focuses on changes in the statistical characteristics of sensor data, such as the mean, variance, and amplitude range of each sensor data point. These statistical characteristics of sensor data vary in different operating scenarios. For example, in a crushing hammer scenario, the pressure and flow fluctuations in the hydraulic system will be more dramatic than in an earthmoving scenario, and the pressure amplitude range may be wider. The system uses set thresholds to determine whether changes in these statistical characteristics exceed the normal range. If so, further analysis is conducted to determine whether the change is caused by a scenario change.

[0187] Changes in geographic location and environmental information: If the target mobile machine is equipped with both geographic and environmental sensors, the system will use this information to assist in determining scene switching. For example, if the location moves from an excavation area to a leveled area, or if environmental sensors detect significant changes in soil type or terrain, the likelihood of a scene switching increases.

[0188] Scenario Switch Processing: If a scenario switch is detected, the system immediately stops the scoring process for the current scenario and returns to the initialization phase, re-performing scenario recognition and parameter configuration to adapt to the new operation scenario. The specific processing steps are as follows: Stop Current Scenario Scoring: The system suspends tasks related to the current scenario, including data collection, feature extraction, and score calculation, to ensure that erroneous scenario data is not mixed into subsequent processing. For example, the system stops the calculation task of the earthwork excavation scenario scoring model to free up related computing resources. Scenario Recognition and Parameter Configuration: The system re-enters the scenario recognition process and analyzes the new sensor data according to the methods used in the operation scenario recognition phase to determine the current operation scenario. Once a new scenario is identified, such as a site leveling scenario, the system reloads the corresponding scoring model and parameter settings based on the specific characteristics of the scenario. For example, a scoring model for the site leveling scenario might be loaded, which might prioritize characteristic parameters such as boom and bucket leveling control and leveling index. Simultaneously, the system adjusts the sensor acquisition strategy, such as appropriately increasing the sampling frequency of the boom tilt sensor and bucket angle sensor to more accurately monitor key parameters in the site leveling operation.

[0189] 8. Sensor fault handling and detection:

[0190] The system monitors sensor data in real time and uses data anomaly detection algorithms to determine whether any sensor failures have occurred.

[0191] The system uses the following fault detection methods:

[0192] Data range detection: Based on the sensor's measurement range and the normal operating parameters of the equipment, the appropriate range of each sensor's data is set. For example, the measurement range of a pilot pressure sensor is typically (0-60 bar). If the collected pressure value exceeds this range, such as a pressure value greater than 60 bar, the system will determine that the data is abnormal, which may indicate a sensor failure.

[0193] Data mutation detection: Analyzes the changing trends of sensor data to detect any data mutations. Under normal circumstances, sensor data should show continuous and smooth changes. For example, if the angle data collected by the IMU sensor shows a sudden and large jump between adjacent sampling points, and this jump does not conform to the movement pattern under normal operation, the system may determine that there is a sensor failure.

[0194] Correlation analysis: This system uses the correlation between data from different sensors to detect faults. For example, there's a correlation between an engine's instantaneous fuel consumption and parameters like engine output torque and hydraulic pump flow. If, at a certain moment, the instantaneous fuel consumption data suddenly increases while related parameters like engine output torque and hydraulic pump flow do not change accordingly, the system, through correlation analysis, determines that the fuel consumption sensor may be faulty.

[0195] Start redundant sensors to obtain alternative data: If redundant sensors exist, the system quickly starts them to obtain alternative data to ensure the continuity of data collection. During system design, some key sensors, such as pilot pressure sensors and IMU sensors, are equipped with redundant sensors. When the main sensor fails, the system automatically switches to the redundant sensor for data collection. For example, when the pilot pressure sensor fails, the backup pilot pressure sensor immediately starts working and collects pressure data at the same sampling frequency, ensuring that the central processing unit can continue to obtain pilot pressure information and maintain the normal operation of the data processing process.

[0196] The system uses a fault diagnosis algorithm to accurately locate the faulty sensor and sends a fault alarm message to the driver and maintenance personnel, informing them of the location of the faulty sensor and the possible cause of the fault. The fault diagnosis algorithm combines the working principle of the sensor, historical fault data, and currently collected abnormal data for analysis. For example, if the data collected by the pilot pressure sensor is always zero, the fault diagnosis algorithm first checks whether the power connection of the sensor is normal, and then analyzes whether there is a break or short circuit in the sensor's signal transmission line. By gradually eliminating possible causes of the fault, the fault point is accurately located. Once the faulty sensor and its possible cause are determined, the system prompts the driver through the display screen, "Pilot pressure sensor (channel 3) fault, possible cause: signal transmission line fault, please check", and sends a fault alarm message to the maintenance personnel's mobile phone or maintenance management system, notifying them to carry out repairs in a timely manner.

[0197] In order to reduce the impact of faults on the scoring results, the system uses data interpolation or prediction algorithms to fill in the missing or abnormal data during the fault period.

[0198] Data interpolation algorithm: For data loss caused by short-term sensor failure, the system uses a linear interpolation algorithm to fill the data.

[0199] Data Prediction Algorithm: For prolonged sensor failures or data anomalies, the system employs a machine learning-based prediction algorithm to fill in the gaps. For example, historical sensor data and mobile machinery operating status data are used to train a time series prediction model, such as a long-short-term memory network model. When a pilot pressure sensor fails, the pre-failure pilot pressure data and other relevant sensor data are input into the trained model, which then predicts the pilot pressure data during the failure period. This machine learning-based prediction algorithm can more accurately fill in the gaps in long-term missing data, minimizing the impact on scoring results. The recovered data is then transmitted to the feature extraction stage of the central processing unit for subsequent data processing and scoring.

[0200] 9. Regular model optimization and regular data extraction:

[0201] The system extracts actual operational data from a distributed data storage system over a set period (e.g., the past month) and the corresponding scoring results, based on a set time period (e.g., monthly). The data extraction process filters data based on pre-defined time ranges and data types. During the extraction process, the system verifies the data's integrity and accuracy to ensure that the extracted data is not corrupted or lost. For example, this is achieved by checking the checksum of data files and verifying the continuity of data records.

[0202] The extracted actual operational data is fed into the current scoring model, and the model's predictions are compared with the actual scores to analyze the model's prediction error. The system comprehensively evaluates the model's prediction performance by calculating multiple error metrics, including mean squared error, mean absolute error, and other metrics such as relative error.

[0203] Based on the error obtained from the analysis, the parameters of the scoring model are adjusted using a machine learning algorithm to optimize the model's adaptability to different working conditions and driver operating styles, thereby improving scoring accuracy. Taking the neural network-based scoring model as an example, the stochastic gradient descent method randomly selects a small number of data samples in each iteration to calculate the gradient of the loss function with respect to the model parameters, and then updates the model parameters in the opposite direction of the gradient, so that the loss function gradually decreases, thereby improving the model's prediction accuracy. In each iteration, the update formula for the model parameters is: ,in is the learning rate, which controls the step size of parameter updates, is the loss function. The adaptive moment estimation method, based on stochastic gradient descent, combines the advantages of momentum and adaptive learning rate adjustment to more efficiently adjust model parameters. Through multiple iterations, the model parameters are continuously adjusted to better fit the relationship between actual operational data and scoring results. During this adjustment process, cross-validation and other techniques are also employed to prevent overfitting and ensure good generalization of the model to new data.

[0204] At the same time, the system continuously collects operational data from newly experienced operators in various operating scenarios to update the empirical model and ensure that the scoring criteria reflect the latest and most optimized operating methods. The system collects new data from experienced operators through various means, such as collaborating with professional mobile machinery training institutions to organize specialized operational tests for experienced operators and collecting operational data from outstanding operators at actual work sites. The collected data covers operational information for mobile machinery of different brands and models under various operating conditions. After preprocessing and feature extraction, the newly collected data is merged with the existing empirical model data and reanalyzed. For example, clustering algorithms are used to analyze both new and existing data to identify typical operating patterns under different operating characteristics, which in turn updates the standard operating characteristic parameters and scoring criteria in the empirical model. Furthermore, the system analyzes the latest operating techniques and technological development trends in the industry and incorporates these factors into the updating of the empirical model. After model optimization, the optimized model is applied to subsequent scoring processes, continuously improving the system's assessment capabilities and providing drivers with more accurate scores and more targeted operational recommendations.

[0205] In summary, through the detailed implementation cases above, the mechanical driving behavior scoring system can comprehensively and accurately evaluate the driver's operating behavior in earthwork excavation and other operation scenarios. It also has the ability to respond to various special situations and continuously improve system performance through regular model optimization, providing strong support for improving mechanical operation efficiency, reducing energy consumption, and improving equipment management level.

[0206] In addition, the embodiment of the present invention also provides a mobile machinery driving behavior evaluation and guidance system, such as Figure 2 As shown, the mobile machinery driving behavior evaluation and guidance system 200 specifically includes:

[0207] The data acquisition module 210 is used to collect real-time operation data of each component of the target mobile machinery and perform data preprocessing; wherein the real-time operation data at least includes pilot pressure data of the pilot handle;

[0208] The scoring calculation module 220 is configured to identify the current operating scenario of the target mobile machine based on the real-time operating data; load a corresponding standard operating model from a preset standard model library based on the current operating scenario; extract driver operating characteristics related to driving behavior from the real-time operating data; and score each driving action of the current driver based on the driver operating characteristics and the standard operating model;

[0209] The decision and feedback module 230 is used to generate targeted operational suggestions based on the calculated scores, and to feed back the scores and operational suggestions to the driver.

[0210] The various embodiments of the present invention are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the device, apparatus, and non-volatile computer storage medium embodiments are generally similar to the method embodiments, so their descriptions are simplified. For relevant details, refer to the descriptions of the method embodiments.

[0211] The above description of specific embodiments of the present invention is provided. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0212] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations may be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for evaluating and guiding driving behavior of mobile machinery, characterized in that: The method comprises: Collecting real-time operating data of various components of the target mobile machinery and performing data preprocessing; wherein the real-time operating data at least includes pilot pressure data of the pilot handle; identifying a current operating scenario of the target mobile machinery based on the real-time operating data; Based on the operational factors affecting each operation scenario, and with the goal of achieving the optimal combination of energy consumption and efficiency for each operation, the theoretical operating parameters for each operation in each operation scenario are calculated. Determining a scoring weight for each theoretical operating parameter based on a relationship between the theoretical operating parameter and the corresponding energy consumption and work efficiency, and constructing a theoretical operating model based on the theoretical operating parameters and the scoring weight; Collecting actual operation data of senior drivers in each operation scenario, and extracting empirical operation parameters of each operation action from the actual operation data; Determining a scoring weight for each empirical operating parameter based on a relationship between the empirical operating parameter and the corresponding energy consumption and work efficiency, and constructing an empirical operating model based on the empirical operating parameters and the scoring weight; The theoretical operation model and the empirical operation model of each operation action in each operation scenario are associated into a group, stored in the database, and a standard model library is constructed; Based on the current operation scenario, a corresponding standard operation model is loaded from a preset standard model library; Extracting driver operation features related to driving behavior from the real-time operation data specifically includes: Extracting the pilot handle control features of the driver based on the pilot pressure data; wherein the pilot handle control features include at least: time proportion features of the pilot pressure data in different amplitude regions, loading features and unloading features of the pilot pressure monitoring data; specifically including: Based on a preset amplitude region division rule, the pilot pressure data is divided into different amplitude regions; wherein the amplitude regions include at least a dead zone, a control zone, and a saturation zone; the duration of the pilot pressure data in different amplitude regions is monitored, and the time proportion of each amplitude region is calculated to obtain the time proportion feature; based on the change time between the initial value and the maximum value of the pilot pressure, the loading time and the loading change rate of the pilot pressure monitoring data are extracted, and the maximum value of the loading change rate and its corresponding time position are determined to obtain the loading feature; at the same time, the unloading time and the unloading change rate of the pilot pressure monitoring data are extracted, and the maximum value of the unloading change rate and its corresponding time position are determined to obtain the unloading feature; Extracting the driver's component operation characteristics based on the motion data of each movable component of the target mobile machinery; wherein the component operation characteristics include at least: motion trajectory characteristics of each component, motion speed change characteristics, and operation coordination characteristics; Extracting the driver's operating energy consumption characteristics based on the target mobile machinery's energy efficiency data; wherein the operating energy consumption characteristics include at least: energy consumption per unit workload and energy consumption distribution characteristics at different operating stages; Summarize and organize the extracted pilot handle control features, component operation features, and operation energy consumption features to form a complete feature data set, thereby obtaining the driver operation features; scoring each driving action of the current driver based on the driver's operating characteristics and the standard operating model; Based on the calculated score, targeted operational recommendations are generated, and the score and operational recommendations are fed back to the driver.

2. A method for evaluating and guiding driving behavior of mobile machinery according to claim 1, characterized in that: Collect real-time operating data of each component of the target mobile machinery and perform data preprocessing, including: During the operation of the target mobile machinery, collecting pilot pressure data of each pressure channel of the left and right pilot handles; Collecting motion data of each movable component; wherein the motion data at least includes joint angle, motion angular velocity, motion acceleration, rotational angular velocity and rotational acceleration; Collecting energy efficiency data of the target mobile machinery; wherein the energy efficiency data includes at least instantaneous fuel consumption, instantaneous engine torque, hydraulic pump output flow, cylinder inlet and outlet pressures, and swing motor inlet and outlet flow and inlet and outlet pressures; The pilot pressure data, the motion data and the energy efficiency data are combined into the real-time operation data, and the real-time operation data is preprocessed based on a preset preprocessing program through an edge computing unit deployed on the target mobile machinery.

3. A method for evaluating and guiding driving behavior of mobile machinery according to claim 2, characterized in that: Identifying a current operating scenario of the target mobile machinery based on the real-time operating data specifically includes: Extracting pressure variation characteristics based on the pilot pressure data; wherein the pressure variation characteristics at least include: sequence characteristics and time interval characteristics of the pilot pressure changes between the pressure channels; Extracting motion characteristics of each component based on the motion data; wherein the motion characteristics of each component include at least: the range of motion angle variation, angular velocity peak value, and motion synchronization characteristics of each component; Obtaining geographic location information of the target mobile machinery and extracting geographic environment features; Fusing the pressure change characteristics, the movement characteristics of each component, and the geographical environment characteristics to obtain a comprehensive scene feature; Matching the comprehensive scene features with the preset features of each operation scene, and determining the operation scene with a matching degree exceeding a preset threshold as the current operation scene; If the matching degree does not reach the preset threshold, the recognition cycle continues.

4. A method for evaluating and guiding driving behavior of mobile machinery according to claim 1, characterized in that: Based on the driver's operating characteristics and the standard operating model, each driving action of the current driver is scored, specifically including: inputting the extracted driver operation characteristics into the theoretical operation model and the empirical operation model in the standard operation model respectively; Comparing the driver's operating characteristics with the theoretical operating parameters in the theoretical operating model, performing a weighted scoring calculation based on the comparison result and the scoring weight of each theoretical operating parameter, and obtaining a first scoring result; Comparing the driver's operating characteristics with the empirical operating parameters in the empirical operating model, and performing a weighted scoring calculation based on the comparison result and the scoring weight of each empirical operating parameter to obtain a second scoring result; The first scoring result and the second scoring result are weighted and summed according to the preset model weight to obtain the final score.

5. The method for evaluating and guiding driving behavior of mobile machinery according to claim 1, characterized in that: After scoring each driving action of the current driver based on the driver's operating characteristics and the standard operating model, the method further includes: The collected real-time operation data, processed feature data, final scores, and operation suggestions are classified and stored in a distributed data storage system; Regularly back up stored data and perform regular maintenance on distributed data storage systems; Extracting the operating data and corresponding scoring results within a period of time from the distributed data storage system according to a preset time period; Input the extracted running data into the current scoring model, compare the scoring results of the current scoring model with the stored scoring results, and analyze the model prediction error; According to the model prediction error, the parameters of the current scoring model are adjusted to optimize the adaptability of the current scoring model to different working conditions and driver operating styles.

6. A method for evaluating and guiding driving behavior of mobile machinery according to claim 1, characterized in that: The method further comprises: Real-time monitoring of sensor data of target mobile machinery, extracting data anomaly features from the sensor data using a data anomaly detection algorithm, and determining whether a sensor failure has occurred based on the data features; wherein the data anomaly features include at least: data anomaly range features, data mutation features, and data correlation features; If a sensor fails, the redundant sensor of the sensor is activated to continue data collection; Perform fault diagnosis on faulty sensors and send alarm information to inform the location of the faulty sensor and possible fault causes; Data interpolation or prediction algorithms are used to fill in missing or abnormal data during the fault period.

7. A mobile machinery driving behavior evaluation and guidance system, using a mobile machinery driving behavior evaluation and guidance method according to any one of claims 1 to 6, characterized in that: The system comprises: A data acquisition module, configured to collect real-time operating data of various components of the target mobile machinery and perform data preprocessing; wherein the real-time operating data at least includes pilot pressure data of the pilot handle; a scoring calculation module configured to identify a current operating scenario of the target mobile machinery based on the real-time operating data; load a corresponding standard operating model from a preset standard model library based on the current operating scenario; extract driver operating characteristics related to driving behavior from the real-time operating data; and score each driving action of the current driver based on the driver operating characteristics and the standard operating model; The decision-making and feedback module is used to generate targeted operational suggestions based on the calculated scores and to feed back the scores and operational suggestions to the driver.

Images