Mobile work machine power transmission obstacle sensing method, system, device, and storage medium

Abstract

The application provides a mobile working machine power transmission obstacle perception method, system, device and storage medium, and relates to the field of intelligent perception. The method comprises the following steps: collecting original multi-source input data, and constructing a structured time sequence input tensor; performing standardization processing on the structured time sequence input tensor, extracting frequency domain change characteristics, and completing uniform format alignment; based on a state evolution function family and a neural ordinary differential equation model, extracting dynamic change characteristics of physical quantities and constructing a distance evolution trajectory prediction function; fusing ultrasonic ranging values and prediction function results to generate a dynamic safety margin sequence; generating an early warning response instruction based on the dynamic safety margin sequence and a multi-level risk threshold value, and realizing dynamic early warning through an acoustic and optical output module; and constructing a power supply management system, adopting a battery and solar component collaborative power supply mechanism to ensure stable operation of the system. The application improves the real-time performance and accuracy of mobile working machine obstacle perception.

Description

Technical Field

[0001] This invention relates to the field of intelligent sensing, specifically to a method, system, device, and storage medium for sensing power transmission obstacles in mobile machinery. Background Technology

[0002] Currently, controlling the safe distance between construction machinery and live power lines remains a significant challenge in power transmission line operation environments. Traditional methods mainly rely on visual inspection, fixed-position sensors, or manual early warning, lacking the ability to perceive dynamic displacement during construction in real time. This makes it difficult to promptly reflect changes in the spatial posture of components such as robotic arms and hoisting equipment, resulting in problems such as response lag and inaccurate early warning.

[0003] Existing obstacle perception technologies are mostly based on the measurement of a single physical quantity, such as electric field or ultrasonic ranging, which cannot collaboratively perceive multi-source information in complex environments. In particular, the accuracy drops significantly under environmental interference and signal obstruction. At the same time, the distance measurement results are mostly discrete values, lacking the ability to continuously model the relative motion trend of obstacles, making it difficult to support timely dynamic risk assessment and proactive control.

[0004] Furthermore, existing early warning strategies are mostly triggered by static thresholds, failing to achieve dynamic classification of risk levels and a progressive response mechanism. This results in early warning information often being either too sensitive or too delayed, affecting its effectiveness. While some systems possess certain intelligent processing capabilities, they often rely on central nodes or high-power computing units, making it difficult to adapt to the low-power operation requirements and adaptive energy management in continuous outdoor working environments.

[0005] In the current context, there is a lack of a transmission fault perception method that can integrate multi-source information, achieve dynamic evolution modeling, and possess intelligent early warning capabilities. In particular, there are still significant technical shortcomings in continuous prediction modeling, multi-level risk response, and stable power supply. There is an urgent need to propose a comprehensive solution with high precision, low power consumption, and real-time performance suitable for mobile machinery. Summary of the Invention

[0006] The purpose of this invention is to propose an artificial intelligence-based method for sensing power transmission obstacles on mobile construction machinery. This method employs multi-source sensor fusion to construct a structured temporal input tensor, models the continuous distance change process between the construction machinery and the power transmission line based on neural network constant differential equations, predicts the obstacle distance evolution trajectory, and generates dynamic early warning response commands by combining multi-level risk thresholds. Finally, real-time early warning output is achieved through an audio-visual module. Simultaneously, a low-power management mechanism is constructed that utilizes both battery and solar power to ensure stable operation of the system in complex outdoor environments.

[0007] To achieve the above objectives, a first aspect of the present invention provides a method for sensing power transmission obstacles in mobile work machinery, comprising the following steps: S1. Construct a multi-source sensing data acquisition architecture, deploy electric field sensors, ultrasonic sensors, temperature sensors, attitude position acquisition units and distance calibration devices on multiple protruding parts of the construction machinery, use signal conditioning circuit to complete low-noise amplification and filtering of electric field sensing signals, collect raw multi-source input data containing electric field, electroacoustic, temperature and attitude multi-modal signals, and send the sensing data to the alarm control unit through wireless communication module. S2. Based on the original multi-source input data, normalization, standardization and format conversion processing are performed to construct a structured time-series input tensor. The electric field intensity value sequence, ultrasonic echo ranging sequence, ambient temperature sequence and attitude position trajectory data are uniformly mapped to a unified time-series structure. The electric field intensity value is estimated based on the magnitude of the induced current generated by electrostatic induction and the effective amplitude is extracted by combining the frequency domain change characteristics. The ultrasonic propagation speed is dynamically adjusted according to the ambient temperature. S3. Based on the structured time-series input tensor, construct a neural ordinary differential equation model, train a family of differential functions to model the continuous dynamic evolution process of multi-source sensing variables in the time dimension, output a distance evolution trajectory prediction function, and simulate the distance change process between construction machinery and transmission lines during the movement process. S4. Based on the distance evolution trajectory prediction function, and by integrating the current ultrasonic ranging data, a time-look assessment of the safe distance at different times is performed to construct a dynamic safety margin sequence, which represents the degree of redundancy between the distance sensed at each time and the minimum safe distance corresponding to the voltage level. S5. Based on the comparison results between the dynamic safety margin sequence and the set multi-level risk thresholds, multi-level early warning response commands are generated, and the sound and light output module is driven by the alarm control unit to output sound and light signals in different frequency forms to realize obstacle perception and dynamic early warning. The sound and light frequency is inversely proportional to the safety margin. S6. Construct a power supply management system, adopt a collaborative power supply strategy of batteries and solar modules, and provide stable power supply to the acquisition module and control module through energy acquisition and time-sharing charge and discharge control mechanism. Support the execution of low power standby, overvoltage and over-discharge protection and early warning functions in continuous outdoor operation scenarios.

[0008] Optionally, step S1 specifically includes: S11. Electric field sensors, ultrasonic sensors, temperature sensors, attitude and position acquisition units, and distance calibration devices are deployed on multiple protruding parts of the construction machinery. Among them, the electric field sensors are used to sense the induced intensity of the power frequency electric field of the transmission line, the ultrasonic sensors are used to sense the distance information of obstacles, the temperature sensors are used to record the ambient temperature, the attitude and position acquisition units are used to collect the motion state information of the machinery itself, and the distance calibration device is used to unify the reference system of the data positions of each sensor in the initial calibration stage. S12. Construct an electric field signal conditioning circuit to perform low-noise amplification and filtering on the electric field induced signal. This includes using a high-sensitivity instrumentation amplifier to construct a low-noise amplification circuit to amplify the electric field induced signal in one stage, and using a first-order RC low-pass filter circuit to filter the sensing signal to generate an analog electric field signal with interference suppression. S13. Input the electric field simulation signal after interference suppression processing, as well as the analog signals output by the ultrasonic sensor and temperature sensor, to the microcontroller with multi-channel analog-to-digital conversion function. Use the analog-to-digital converter channel to synchronously sample each analog signal. At the same time, acquire the three-dimensional attitude data output by the attitude position acquisition unit through the digital communication interface to generate a multi-source sampling dataset containing electric field induction amplitude, electric field frequency characteristics, ultrasonic echo response time, ambient temperature and three-dimensional attitude trajectory. S14. Based on the ambient temperature value collected by the temperature sensor, the ultrasonic propagation speed is corrected, and the ultrasonic ranging value is calculated by combining the ultrasonic echo response time to form preliminary obstacle distance data. S15. Based on the multi-source sampling dataset, preliminary obstacle distance data and distance calibration reference data, construct the original multi-source input data and send it to the alarm control unit through the wireless communication module for subsequent modeling and early warning judgment.

[0009] Optionally, step S2 specifically includes: S21. Receive the raw multi-source input data forwarded by the alarm control unit, including electric field induction amplitude, electric field frequency characteristics, preliminary obstacle distance data, ambient temperature, attitude trajectory information and distance calibration reference data; S22. Normalize the electric field amplitude to generate a dimensionless electric field intensity value sequence; standardize the preliminary obstacle distance data to generate a distance measurement sequence; perform time synchronization and missing data completion operations on the ambient temperature and attitude trajectory information to generate a continuous ambient temperature sequence and attitude trajectory sequence. S23. Based on the distance calibration reference data, establish a unified spatial reference frame, map the electric field intensity value sequence, distance measurement sequence, ambient temperature sequence and attitude trajectory sequence to a unified coordinate system and time structure, and generate a standard physical measurement sequence. S26. Based on standard physical measurement sequences, a structured temporal input tensor is constructed by stacking them in chronological order. This tensor serves as the input basis for subsequent neural ordinary differential equation models and is used to describe the continuous temporal evolution state between construction machinery and power transmission barriers.

[0010] Optionally, step S3 specifically includes: S31. Receive structured temporal input tensors, construct a neural ordinary differential equation model framework for obstacle distance continuous modeling tasks, and map the dynamic relationship between construction machinery and transmission lines into the evolution process of state variables in the time dimension, where state variables include distance measurement state, electric field induction state, attitude trajectory state and ambient temperature state. S32. Define a family of state evolution functions with dynamic sensing capabilities. Input various physical measurement sequences in the structured time-series input tensor into the corresponding state sensing sub-networks to extract the dynamic change features of different physical quantities. On this basis, construct the derivative structure of the state evolution function so that the family of functions has the ability to continuously characterize the dynamic change trend of the input variables. S33. By introducing an embedded time modulation mechanism, timestamp information is explicitly encoded and embedded into the evolution function of each state variable, enhancing the model's ability to model non-equal interval observations, sudden disturbances, and high-speed displacements during construction, and realizing a fine expression of nonlinear continuous changes in time series tensors. S34. Using a numerical ordinary differential equation solver, perform time-step integration on the family of state evolution functions to generate a continuous predicted trajectory describing the distance change trend between construction machinery and transmission lines, and output the distance evolution trajectory prediction function as the basic input for subsequent safety margin assessment and early warning response. The distance evolution trajectory prediction function is expressed as: ; ; in: This is the hidden state vector, representing the encoding of the environment state at the current moment; The structured temporal input tensor contains the observation input at the current time step; For the family of differential functions in the God's ordinary differential equations, parameterized by the multilayer perceptron; and For output layer weights and bias terms; This refers to the "distance evolution trajectory prediction function" as defined in this invention; Optionally, step S4 specifically includes: S41. Construct a predicted distance sequence based on the output of the distance evolution trajectory prediction function to represent the continuous distance change process between construction machinery and power transmission obstacles over time; S42. Set a multi-level risk threshold sequence, including safety threshold, warning threshold and danger threshold, which correspond to different levels of warning response, and are used to classify and distinguish the predicted distance sequence. S43. Compare the predicted distance sequence with the multi-level risk threshold sequence frame by frame, perform interval matching operation, determine the warning level at the current moment, and output the dynamic safety margin sequence. S44. Construct a multi-level early warning response mechanism based on dynamic safety margin sequence, track and label the risk level status at different times hour by hour, and generate early warning control signals of the corresponding level to guide the alarm control unit to execute differentiated response strategies.

[0011] Optionally, step S5 specifically includes: S51. Based on the item-by-item comparison between the dynamic safety margin sequence and the multi-level risk threshold sequence, generate a set of warning level labels to mark the safety level status corresponding to each time node. S52. Based on the set of warning level labels, call the set of warning response rules and match and generate corresponding multi-level warning response instructions, including the control instruction sets corresponding to Level 1 warning, Level 2 warning and Level 3 warning; S53. Send the multi-level early warning response command to the alarm control unit and drive the sound and light output module to generate corresponding sound and light output signals. The sound and light output signals represent the early warning level in different frequency forms. S54. The frequency of the audible and visual output signal is mapped and controlled with the dynamic safety margin sequence to dynamically adjust the output frequency. The frequency of the audible and visual output is inversely proportional to the safety margin value, which is used to realize dynamic early warning feedback for power transmission fault perception in the working environment.

[0012] Optionally, step S6 specifically includes: S61. Construct a power supply management system and configure battery packs and solar modules to form a collaborative power supply structure, which are used to provide basic power supply and supplementary energy input respectively. S62. Based on the working status of the solar modules and the current load power level, execute the energy harvesting and scheduling mechanism to control the power distribution path between the solar modules and the batteries, and support time-sharing charge and discharge control strategies. S63. Power supply paths are configured for the multi-source acquisition module, the alarm control unit, and the second data processing module respectively, forming a power supply control structure with functional separation. S64. Implement overvoltage protection, over-discharge protection, and voltage status monitoring mechanisms during power supply to ensure the long-term stable operation of the energy system; S65. A low-power standby controller is set at the early warning output module. The module's operating status is dynamically switched based on the multi-level risk threshold judgment results, so as to realize the low-power early warning function in continuous outdoor operation scenarios.

[0013] A second aspect of the present invention provides a power transmission obstacle sensing system for mobile work machinery, the system comprising: The multi-source acquisition module consists of several sensors and a distance calibration device, and is deployed at multiple predetermined locations on the operating machinery to collect raw multi-source input data; wherein, the sensors include an electric field sensor, an ultrasonic sensor, a temperature sensor, and an attitude and position acquisition unit; The first data processing module performs normalization, standardization, and format conversion processing on the original multi-source input data to construct a structured temporal input tensor. The second data processing module constructs a neural ordinary differential equation model based on the structured time-series input tensor, trains a family of differential functions to model the continuous dynamic evolution process of multi-source sensing variables in the time dimension, and outputs a distance evolution trajectory prediction function. The third data processing module, based on the distance evolution trajectory prediction function, integrates current ultrasonic ranging data to perform a time-look assessment of the safe distance at different times and constructs a dynamic safety margin sequence. The early warning output module generates multi-level early warning response instructions and sends them to the alarm control unit based on the comparison results between the dynamic safety margin sequence and the set multi-level risk thresholds. The alarm control unit realizes obstacle perception and dynamic early warning based on the multi-level early warning response; The power supply management system adopts a collaborative power supply strategy of batteries and solar modules, and provides stable power to the acquisition module and control module through energy acquisition and time-sharing charge and discharge control mechanism.

[0014] Optionally, the power management system is configured with a battery pack and solar panels to form a collaborative power supply structure; Based on the working status of the solar modules and the current load power level, an energy harvesting and scheduling mechanism is executed to control the power distribution path between the solar modules and the batteries. Power supply paths are configured for the multi-source acquisition module, the alarm control unit, and the second data processing module respectively, forming a power supply control structure based on functional separation. During power supply, overvoltage protection, over-discharge protection, and voltage status monitoring mechanisms are implemented. A low-power standby controller is set at the early warning output module to dynamically switch the module's operating status based on the results of multi-level risk threshold judgments.

[0015] A third aspect of the present invention provides an electronic device comprising a processor and a memory storing computer program instructions; wherein the processor, when executing the computer program instructions, implements the mobile work machinery power transmission obstacle sensing method described in the first aspect.

[0016] In a fourth aspect, the present invention provides a computer-readable storage medium storing at least one executable instruction that, when executed on an electronic device, causes the electronic device to perform the mobile work machinery power transmission obstacle sensing method described in the first aspect.

[0017] Compared with the prior art, the present invention has at least the following technical effects: First, by constructing a structured temporal input tensor that integrates information from multiple sources such as electric field, ultrasound, temperature, and attitude, this invention achieves an accurate representation of the relative state between mobile machinery and power transmission obstacles, significantly improving the integrity and temporal consistency of data input and providing a solid data foundation for subsequent dynamic modeling.

[0018] Secondly, this invention introduces a neural ordinary differential equation model for distance prediction tasks, constructs a family of state evolution functions with continuous modeling capabilities, effectively characterizes the dynamic evolution characteristics of multi-source physical variables in the time dimension, and enhances the forward-looking analysis capability of obstacle approach trends.

[0019] Furthermore, this invention designs a method for linking dynamic safety margin and multi-level risk thresholds in the early warning mechanism. Combined with a dynamic adjustment strategy for the audio-visual output frequency, it realizes the linkage control of risk level and output signal, thereby improving the interpretability and response efficiency of the early warning response.

[0020] Finally, this invention combines a solar energy and battery-based collaborative power supply strategy to construct a low-power, zone-managed power supply structure, ensuring the stability and safety of the system during long-term continuous operation in the field. Attached Figure Description

[0021] Figure 1 This is an overall flowchart of the mobile operation machinery power transmission obstacle sensing method proposed in this invention.

[0022] Figure 2 This is a schematic diagram of the distance evolution trajectory prediction mechanism based on the neuron frequent differential equation model in the embodiment.

[0023] Figure 3 This is a diagram showing the inverse mapping relationship between the acoustic and optical output frequency and the dynamic safety margin in the embodiment. Detailed Implementation

[0024] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0025] refer to Figure 1-3 A method for sensing power transmission obstacles in mobile operating machinery based on artificial intelligence includes the following steps: S1. Construct a multi-source sensing data acquisition architecture, deploy electric field sensors, ultrasonic sensors, temperature sensors, attitude position acquisition units and distance calibration devices on multiple protruding parts of the construction machinery, use signal conditioning circuit to complete low-noise amplification and filtering of electric field sensing signals, collect raw multi-source input data containing electric field, electroacoustic, temperature and attitude multi-modal signals, and send the sensing data to the alarm control unit through wireless communication module. S2. Based on the original multi-source input data, normalization, standardization and format conversion processing are performed to construct a structured time-series input tensor. The electric field intensity value sequence, ultrasonic echo ranging sequence, ambient temperature sequence and attitude position trajectory data are uniformly mapped to a unified time-series structure. The electric field intensity value is estimated based on the magnitude of the induced current generated by electrostatic induction and the effective amplitude is extracted by combining the frequency domain change characteristics. The ultrasonic propagation speed is dynamically adjusted according to the ambient temperature. S3. Based on the structured time-series input tensor, construct a neural ordinary differential equation model, train a family of differential functions to model the continuous dynamic evolution process of multi-source sensing variables in the time dimension, output a distance evolution trajectory prediction function, and simulate the distance change process between construction machinery and transmission lines during the movement process. S4. Based on the distance evolution trajectory prediction function, and by integrating the current ultrasonic ranging data, a time-look assessment of the safe distance at different times is performed to construct a dynamic safety margin sequence, which represents the degree of redundancy between the distance sensed at each time and the minimum safe distance corresponding to the voltage level. S5. Based on the comparison results between the dynamic safety margin sequence and the set multi-level risk thresholds, multi-level early warning response commands are generated, and the sound and light output module is driven by the alarm control unit to output sound and light signals in different frequency forms to realize obstacle perception and dynamic early warning. The sound and light frequency is inversely proportional to the safety margin. S6. Construct a power supply management system, adopt a collaborative power supply strategy of batteries and solar modules, and provide stable power supply to the acquisition module and control module through energy acquisition and time-sharing charge and discharge control mechanism. Support the execution of low power standby, overvoltage and over-discharge protection and early warning functions in continuous outdoor operation scenarios.

[0026] This invention proposes an artificial intelligence-based method for detecting power transmission line faults on mobile construction machinery. It constructs an integrated intelligent sensing system encompassing multi-source sensing data acquisition, structured modeling, distance evolution prediction, dynamic early warning, and energy consumption control. Innovatively, multiple sensor fusion devices are deployed at key components of the construction machinery, and a neural network constant differential equation model is introduced to continuously and dynamically model the sensing data, enabling real-time prediction and early warning of power transmission line faults. This method combines an environmental temperature correction mechanism with an acoustic-optical linkage feedback design to establish a dynamic margin-early warning frequency mapping relationship and configures a collaborative power supply system to ensure long-term stable operation of the equipment in outdoor scenarios. It exhibits outstanding multimodal sensing capabilities, continuous modeling accuracy, and energy adaptability.

[0027] In this embodiment, step S1 specifically includes: S11. Electric field sensors, ultrasonic sensors, temperature sensors, attitude and position acquisition units, and distance calibration devices are deployed on multiple protruding parts of the construction machinery. Among them, the electric field sensors are used to sense the induced intensity of the power frequency electric field of the transmission line, the ultrasonic sensors are used to sense the distance information of obstacles, the temperature sensors are used to record the ambient temperature, the attitude and position acquisition units are used to collect the motion state information of the machinery itself, and the distance calibration device is used to unify the reference system of the data positions of each sensor in the initial calibration stage. S12. Construct an electric field signal conditioning circuit to perform low-noise amplification and filtering on the electric field induced signal. This includes using a high-sensitivity instrumentation amplifier to construct a low-noise amplification circuit to amplify the electric field induced signal in one stage, and using a first-order RC low-pass filter circuit to filter the sensing signal to generate an analog electric field signal with interference suppression. S13. Input the electric field simulation signal after interference suppression processing, as well as the analog signals output by the ultrasonic sensor and temperature sensor, to the microcontroller with multi-channel analog-to-digital conversion function. Use the analog-to-digital converter channel to synchronously sample each analog signal. At the same time, acquire the three-dimensional attitude data output by the attitude position acquisition unit through the digital communication interface to generate a multi-source sampling dataset containing electric field induction amplitude, electric field frequency characteristics, ultrasonic echo response time, ambient temperature and three-dimensional attitude trajectory. S14. Correct the ultrasonic propagation speed based on the ambient temperature value collected by the temperature sensor, and calculate the ultrasonic ranging value by combining the ultrasonic echo response time to form preliminary obstacle distance data. S15. Based on the multi-source sampling dataset, preliminary obstacle distance data and distance calibration reference data, construct the original multi-source input data and send it to the alarm control unit through the wireless communication module for subsequent modeling and early warning judgment.

[0028] This step refines the multi-source sensing data acquisition process by deploying various types of sensors, including electric field, ultrasonic, temperature, and attitude / position sensors, and introducing a distance calibration device to establish a unified reference frame, ensuring the spatial consistency and comparability of various physical signals. Furthermore, by building a dedicated electric field conditioning circuit and a multi-channel analog-to-digital conversion mechanism, the accuracy and anti-interference capability of signal acquisition are effectively improved, providing high-quality raw multi-source data input for subsequent structured processing and modeling.

[0029] In this embodiment, step S2 specifically includes: S21. Receive the raw multi-source input data forwarded by the alarm control unit, including electric field induction amplitude, electric field frequency characteristics, preliminary obstacle distance data, ambient temperature, attitude trajectory information and distance calibration reference data; S22. Normalize the electric field amplitude to generate a dimensionless electric field intensity value sequence; standardize the preliminary obstacle distance data to generate a distance measurement sequence; perform time synchronization and missing data completion operations on the ambient temperature and attitude trajectory information to generate a continuous ambient temperature sequence and attitude trajectory sequence. S23. Based on the distance calibration reference data, establish a unified spatial reference frame, map the electric field intensity value sequence, distance measurement sequence, ambient temperature sequence and attitude trajectory sequence to a unified coordinate system and time structure, and generate a standard physical measurement sequence. S26. Based on standard physical measurement sequences, a structured temporal input tensor is constructed by stacking them in chronological order. This tensor serves as the input basis for subsequent neural ordinary differential equation models and is used to describe the continuous temporal evolution state between construction machinery and power transmission barriers.

[0030] This step details the standardization and unified modeling process for the original multi-source input data, proposing to normalize, standardize, and align the multimodal physical quantity data to a unified reference frame, and further construct a structured temporal input tensor. This structure provides a unified format for continuous temporal data input to the neural network's frequent differential equation model, significantly enhancing the system's ability to fuse and process non-uniform sensing data and its temporal consistency representation capability.

[0031] In this embodiment, step S3 specifically includes: S31. Receive structured temporal input tensors, construct a neural ordinary differential equation model framework for obstacle distance continuous modeling tasks, and map the dynamic relationship between construction machinery and transmission lines into the evolution process of state variables in the time dimension, where state variables include distance measurement state, electric field induction state, attitude trajectory state and ambient temperature state. S32. Define a family of state evolution functions with dynamic sensing capabilities. Input various physical measurement sequences in the structured time-series input tensor into the corresponding state sensing sub-networks to extract the dynamic change features of different physical quantities. On this basis, construct the derivative structure of the state evolution function so that the family of functions has the ability to continuously characterize the dynamic change trend of the input variables. S33. By introducing an embedded time modulation mechanism, timestamp information is explicitly encoded and embedded into the evolution function of each state variable, enhancing the model's ability to model non-equal interval observations, sudden disturbances, and high-speed displacements during construction, and realizing a fine expression of nonlinear continuous changes in time series tensors. S34. Using a numerical ordinary differential equation solver, perform time-step integration on the family of state evolution functions to generate a continuous predicted trajectory describing the distance change trend between construction machinery and transmission lines, and output the distance evolution trajectory prediction function as the basic input for subsequent safety margin assessment and early warning response. The distance evolution trajectory prediction function is expressed as: ; ; in: ; This is the hidden state vector, representing the encoding of the environment state at the current moment; The structured temporal input tensor contains the observation input at the current time step; For the family of differential functions in the God's ordinary differential equations, parameterized by the multilayer perceptron; This is not the "distance prediction function" itself, but rather the state evolution mechanism within the prediction function; it represents: hidden states. The time-varying derivative is obtained by the neural network Modeling, with the current state as the input. With input data This is the core structure of the "divine constant differential equation model," used to model the hidden state changes over continuous time. It refers to the setting of initial conditions in the constant differential equation model of God; ; and For output layer weights and bias terms; This refers to the "distance evolution trajectory prediction function" as defined in this invention; This is the true "distance evolution trajectory prediction function"; it transforms the hidden state after evolution through ordinary differential equations. Decode into output This refers to the predicted "distance to the transmission barrier"; The constant differential equation establishes a continuous function of the hidden state's evolution over time. The "distance evolution trajectory prediction function" is based on this. Decoded result The combination of the two constitutes the input... Distance prediction The complete model; Formula source: The mathematical foundation comes from state-space dynamics models in control theory, especially the following initial value problems, which utilize the original divine frequent differential equation modeling framework: ; in It is a parameterized neural network; Formula Improvement: Remove system output from abstract state Derived as observable predicted distance values It also reflects the correlation with the input physical quantities; Input mapping encoder: converts structured temporal input tensors Mapped to initial state : ; in This represents a feedforward neural network used to extract initial features from multimodal physical inputs; God's constant differential modeler: The continuous evolution of the state over time satisfies: ; in It is a multilayer perceptron structure with temporal embedding, which fits the dynamic change patterns under structured input; Distance decoding function: Decoded as target distance value: ; in and These are linear decoding parameters, and the final output is a continuous prediction of obstacle distances.

[0032] The distance evolution trajectory prediction function continuously models the temporal trends of various state variables in a structured time-series input tensor and outputs an estimated future distance between construction machinery and power transmission lines. This function uses a neural network with a constant differential equation model as its core structure, employing a constructed family of differential functions to model the state evolution process. Each type of physical measurement sequence, such as electric field signals, ultrasonic distance, and attitude angle, is input into the corresponding state-aware sub-network to extract its continuous dynamic change characteristics. Based on this, a family of evolution functions for distance state variables is constructed, establishing a mapping relationship between the time derivatives of the model's implicit states and the channel features in the input tensor. This allows the model to learn from historical sensing data and generate a set of differentiable functions to fit the dynamic evolution of obstacle distance over continuous time periods. This prediction function has an interpretable differential structure, supports state interpolation and inference at arbitrary time points, and possesses robust handling capabilities for abrupt data points, significantly improving the accuracy and continuity of the prediction of the spatial relationship between construction machinery and power transmission lines.

[0033] In this embodiment, step S4 specifically includes: S41. Construct a predicted distance sequence based on the output of the distance evolution trajectory prediction function to represent the continuous distance change process between construction machinery and power transmission obstacles over time; S42. Set a multi-level risk threshold sequence, including safety threshold, warning threshold and danger threshold, which correspond to different levels of warning response, and are used to classify and distinguish the predicted distance sequence. S43. Compare the predicted distance sequence with the multi-level risk threshold sequence frame by frame, perform interval matching operation, determine the warning level at the current moment, and output the dynamic safety margin sequence. S44. Construct a multi-level early warning response mechanism based on dynamic safety margin sequence, track and label the risk level status at different times hour by hour, and generate early warning control signals of the corresponding level to guide the alarm control unit to execute differentiated response strategies.

[0034] The dynamic safety margin sequence is a key intermediate variable used in this invention to assess the safety status of the distance between construction machinery and transmission lines. It is generated by comparing a continuous predicted distance sequence output by a distance evolution trajectory prediction function with a set multi-level risk threshold sequence frame by frame. The risk threshold sequence consists of multiple levels, including safety thresholds, warning thresholds, and danger thresholds, each corresponding to a different level of risk perception boundary. At each moment, the system determines whether a potential risk exists by comparing the predicted machinery-obstacle distance with each level of threshold and calculates the safety margin value at that moment, i.e., the difference between the current distance and the nearest risk threshold. This safety margin value forms a sequence in the time dimension, reflecting the spatial safety buffer level at each moment within the entire prediction period, serving as the basis for generating warning response levels and controlling the frequency of audible and visual outputs. The dynamic safety margin sequence not only achieves a forward-looking assessment of future safety status but also possesses continuous, adjustable, and traceable characteristics, facilitating real-time dynamic risk labeling and multi-level warning signal generation.

[0035] In this embodiment, step S5 specifically includes: S51. Based on the item-by-item comparison between the dynamic safety margin sequence and the multi-level risk threshold sequence, generate a set of warning level labels to mark the safety level status corresponding to each time node. S52. Based on the set of warning level labels, call the set of warning response rules and match and generate corresponding multi-level warning response instructions, including the control instruction sets corresponding to Level 1 warning, Level 2 warning and Level 3 warning; S53. Send the multi-level early warning response command to the alarm control unit and drive the sound and light output module to generate corresponding sound and light output signals. The sound and light output signals represent the early warning level in different frequency forms. S54. The frequency of the audible and visual output signal is mapped and controlled with the dynamic safety margin sequence to dynamically adjust the output frequency. The frequency of the audible and visual output is inversely proportional to the safety margin value, which is used to realize dynamic early warning feedback for power transmission fault perception in the working environment.

[0036] This step focuses on the actual output mechanism of multi-level early warning signals. By establishing an inverse mapping relationship between dynamic safety margin and audio-visual output frequency, an audio-visual response adjustment strategy based on risk level is realized. This mechanism not only enhances the perceptibility and accessibility of early warning information in the field environment, but also achieves continuous mapping expression of early warning levels through dynamic frequency control, demonstrating the system's highly coupled response characteristics between perception and feedback.

[0037] In this embodiment, step S6 specifically includes: S61. Construct a power supply management system and configure battery packs and solar modules to form a collaborative power supply structure, which are used to provide basic power supply and supplementary energy input respectively. S62. Based on the working status of the solar modules and the current load power level, execute the energy harvesting and scheduling mechanism to control the power distribution path between the solar modules and the batteries, and support time-sharing charge and discharge control strategies. S63. Power supply paths are configured for the multi-source acquisition module, the alarm control unit, and the second data processing module respectively, forming a power supply control structure with functional separation. S64. Implement overvoltage protection, over-discharge protection, and voltage status monitoring mechanisms during power supply to ensure the long-term stable operation of the energy system; S65. A low-power standby controller is set at the early warning output module. The module's operating status is dynamically switched based on the multi-level risk threshold judgment results, so as to realize the low-power early warning function in continuous outdoor operation scenarios.

[0038] This step innovatively constructs a collaborative power supply system composed of batteries and solar modules, and dynamically switches the power supply path according to the load status, improving the system's energy adaptability in long-term outdoor operation scenarios. At the same time, through time-sharing charging and discharging and overvoltage / over-discharging protection mechanisms, the energy consumption of each functional module is managed separately, effectively supporting the continuous operation and stable early warning output of the entire intelligent sensing system under extreme conditions such as network outages and power outages.

[0039] To verify the feasibility of this invention in practice, it was applied to a mobile crane obstacle sensing system in a high-voltage transmission line inspection and construction scenario. In this scenario, construction machinery frequently performs lifting operations near transmission lines, posing a risk that the boom may accidentally touch the conductor or enter a dangerous electric field area under complex postures. Traditional obstacle warning systems often rely on fixed distance thresholds for judgment, which cannot accurately identify safety margin fluctuations caused by posture changes, and cannot track the continuous evolution of different types of sensing data in real time.

[0040] During the application deployment phase, electric field sensors, ultrasonic sensors, attitude and position acquisition units, ambient temperature sensors, and distance calibration devices are deployed at the end of the boom, the slewing bearing, and the front of the vehicle body of the construction machinery. After amplification and filtering of the electric field induced signals by a signal conditioning circuit, they are synchronously sampled by a multi-channel analog-to-digital converter, and then all sensor signals are transmitted to the alarm control unit in this invention via a wireless module. The acquisition module continuously records the electric field strength, ultrasonic echo time, ambient temperature, and three-dimensional attitude trajectory between the construction machinery and the conductor under different operating conditions.

[0041] After normalization, standardization, and time synchronization, the sensed data is constructed into a structured temporal input tensor, which is then input into a neural network constant differential equation model. The model extracts the continuous dynamic features of physical quantities changing over time and generates a distance evolution trajectory prediction function between construction machinery and transmission lines. At different time points, the system integrates the current ultrasonic ranging value with the predicted distance sequence to predict the distance change trend within the next 3 seconds. Based on the set multi-level risk thresholds, it generates a dynamic safety margin sequence and outputs corresponding warning level labels.

[0042] During actual operation, when the boom ascends at a speed of 2° / s to a position 3.2 meters from the conductor, the model predicts it will enter the warning zone within 2.1 seconds. The system immediately issues a yellow audible and visual warning at a frequency of 2Hz, with a current safety margin of 0.6 meters. As the boom approaches the conductor further and its posture changes abruptly, causing a sudden increase in the electric field strength, the system, based on the evolution trajectory, determines it will enter the danger zone within 1.3 seconds and issues a red audible and visual warning at a frequency of 4Hz, simultaneously triggering a forced stop signal to control the boom's movement.

[0043] In over 300 sets of construction action data, the method of this invention achieved an average early warning of high-risk actions 1.8 seconds in advance, with a root mean square error of 0.34 meters in predicted distance. Compared with traditional systems based on single-point distance measurement thresholds, the false alarm rate decreased by 38.7%, and the false alarm rate decreased by 51.2%. In energy consumption testing, this system adopted a battery and solar panel co-powered mode, ensuring normal operation for over 90% of the time under all-weather conditions, with a 22.3% reduction in average monthly energy cost per unit. In system stability evaluation, the model demonstrated strong robustness to rapid changes in the construction arm's posture, maintaining a 92.4% distance prediction accuracy even under conditions where the angular velocity of the posture change was greater than 3° / s.

[0044] The above embodiments demonstrate that the present invention can effectively sense the dynamic obstacle risks between mobile construction machinery and high-voltage transmission lines, and realize intelligent early warning response based on continuous distance prediction and multi-level early warning strategies, effectively improving the operational safety and system adaptability in power construction scenarios.

[0045] The logical approach behind the mobile work machinery power transmission obstacle sensing method disclosed in the above embodiments can be implemented entirely or partially through software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented entirely or partially in the form of a computer program product. A computer program product includes one or more computer instructions or computer programs.

[0046] When computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives (SSDs).

[0047] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0048] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0049] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0050] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for sensing power transmission obstacles in mobile operating machinery, characterized in that, Includes the following steps: S1. Deploy sensors and distance calibration devices in multiple parts of the operating machinery to collect raw multi-source input data; S2. Based on the original multi-source input data, perform normalization, standardization and format conversion processing to construct a structured temporal input tensor; S3. Based on the structured time-series input tensor, construct a neural ordinary differential equation model, train a family of differential functions to model the continuous dynamic evolution process of multi-source sensing variables in the time dimension, and output a distance evolution trajectory prediction function. S4. Based on the distance evolution trajectory prediction function, and by integrating current ultrasonic ranging data, a time-look assessment of the safe distance at different times is performed to construct a dynamic safety margin sequence. S5. Based on the comparison results between the dynamic safety margin sequence and the set multi-level risk thresholds, generate multi-level early warning response instructions, and realize obstacle perception and dynamic early warning through the alarm control unit.

2. The method for sensing power transmission obstacles in mobile operating machinery according to claim 1, characterized in that, Step S1 specifically includes: S11. Electric field sensors, ultrasonic sensors, temperature sensors, attitude and position acquisition units, and distance calibration devices are deployed on multiple protruding parts of the construction machinery. S12. Construct an electric field signal conditioning circuit to perform low-noise amplification and filtering on the electric field induced signal to generate an analog electric field signal with interference suppression. S13. Input the electric field simulation signal after interference suppression processing, as well as the analog signals output by the ultrasonic sensor and temperature sensor, to the microcontroller with multi-channel analog-to-digital conversion function. Use the analog-to-digital converter channel to synchronously sample each analog signal. At the same time, obtain the three-dimensional attitude data output by the attitude position acquisition unit through the digital communication interface to generate a multi-source sampling dataset. S14. Based on the ambient temperature value collected by the temperature sensor, the ultrasonic propagation speed is corrected, and the ultrasonic ranging value is calculated by combining the ultrasonic echo response time to form preliminary obstacle distance data. S15. Based on the multi-source sampling dataset, preliminary obstacle distance data and distance calibration reference data, construct the original multi-source input data and send it to the alarm control unit.

3. The method for sensing power transmission obstacles in mobile work machinery according to claim 1, characterized in that, Step S2 specifically includes: S21. Receive raw multi-source input data forwarded by the alarm control unit; S22. Normalize the electric field amplitude to generate an electric field strength value sequence; standardize the preliminary obstacle distance data to generate a distance measurement sequence; perform time synchronization and missing data completion operations on the ambient temperature and attitude trajectory information to generate an ambient temperature sequence and an attitude trajectory sequence. S23. Based on the distance calibration reference data, establish a unified spatial reference frame, map the electric field intensity value sequence, distance measurement sequence, ambient temperature sequence and attitude trajectory sequence to a unified coordinate system and time structure, and generate a standard physical measurement sequence. S26. Based on standard physical measurement sequences, stack them in chronological order to construct a structured temporal input tensor.

4. The method for sensing power transmission obstacles in mobile operating machinery according to claim 1, characterized in that, Step S3 specifically includes: S31. Receive structured temporal input tensors, construct a neural ordinary differential equation model framework for continuous obstacle distance modeling tasks, and map the dynamic relationship between construction machinery and transmission lines into the evolution process of state variables in the time dimension. S32. Define a family of state evolution functions with dynamic sensing capabilities. Input various physical measurement sequences in the structured time-series input tensor into the corresponding state sensing sub-networks to extract the dynamic change features of different physical quantities. On this basis, construct the derivative structure of the state evolution function. S33. By introducing an embedded time modulation mechanism, timestamp information is explicitly encoded and embedded into the evolution function of each state variable; S34. Using a numerical ordinary differential equation solver, perform time-step integration on the family of state evolution functions to generate a continuous predicted trajectory describing the distance change trend between the construction machinery and the transmission line, and output the distance evolution trajectory prediction function: In the formula, This is the hidden state vector, representing the encoding of the environment state at the current moment; The structured temporal input tensor contains the observation input at the current time step; For the family of differential functions in the God's ordinary differential equations, parameterized by the multilayer perceptron; It refers to the setting of initial conditions in the constant differential equation model of God; and These are the output layer weights and bias terms; This is the distance evolution trajectory prediction function.

5. The method for sensing power transmission obstacles in mobile operating machinery according to claim 1, characterized in that, Step S4 specifically includes: S41. Construct a predicted distance sequence based on the output of the distance evolution trajectory prediction function to represent the continuous distance change process between construction machinery and power transmission obstacles over time; S42. Set a multi-level risk threshold sequence, including safety threshold, warning threshold and danger threshold, which correspond to different levels of warning response. S43. Compare the predicted distance sequence with the multi-level risk threshold sequence frame by frame to determine the warning level at the current moment and output the dynamic safety margin sequence. S44. Construct a multi-level early warning response mechanism based on dynamic safety margin sequence, track and label the risk level status at different times hour by hour, and generate early warning control signals of the corresponding level.

6. The method for sensing power transmission obstacles in mobile work machinery according to claim 1, characterized in that, Step S5 specifically includes: S51. Generate a set of warning level labels based on the item-by-item comparison between the dynamic safety margin sequence and the multi-level risk threshold sequence; S52. Based on the set of warning level labels, call the set of warning response rules and match them to generate corresponding multi-level warning response instructions; S53. Send the multi-level early warning response command to the alarm control unit and drive the sound and light output module to generate the corresponding sound and light output signal; S54. Map the frequency of the audio-visual output signal to the dynamic safety margin sequence and dynamically adjust the output frequency.

7. A mobile working machinery power transmission obstacle sensing system, characterized in that, include: The multi-source acquisition module consists of several sensors and a distance calibration device, and is deployed at multiple predetermined locations on the operating machinery to collect raw multi-source input data; wherein, the sensors include an electric field sensor, an ultrasonic sensor, a temperature sensor, and an attitude and position acquisition unit; The first data processing module performs normalization, standardization, and format conversion processing on the original multi-source input data to construct a structured temporal input tensor. The second data processing module constructs a neural ordinary differential equation model based on the structured time-series input tensor, trains a family of differential functions to model the continuous dynamic evolution process of multi-source sensing variables in the time dimension, and outputs a distance evolution trajectory prediction function. The third data processing module, based on the distance evolution trajectory prediction function, integrates current ultrasonic ranging data to perform a time-look assessment of the safe distance at different times and constructs a dynamic safety margin sequence. The early warning output module generates multi-level early warning response instructions and sends them to the alarm control unit based on the comparison results between the dynamic safety margin sequence and the set multi-level risk thresholds. The alarm control unit realizes obstacle perception and dynamic early warning based on the multi-level early warning response; The power supply management system adopts a collaborative power supply strategy of batteries and solar modules, and provides stable power to the acquisition module and control module through energy acquisition and time-sharing charge and discharge control mechanism.

8. The mobile work machinery power transmission obstacle sensing system according to claim 7, characterized in that, The power management system is configured with battery packs and solar panels to form a collaborative power supply structure. Based on the working status of the solar modules and the current load power level, an energy harvesting and scheduling mechanism is executed to control the power distribution path between the solar modules and the batteries. Power supply paths are configured for the multi-source acquisition module, the alarm control unit, and the second data processing module respectively, forming a power supply control structure based on functional separation. During power supply, overvoltage protection, over-discharge protection, and voltage status monitoring mechanisms are implemented. A low-power standby controller is set at the early warning output module to dynamically switch the module's operating status based on the results of multi-level risk threshold judgments.

9. An electronic device, characterized in that, include: Processor and memory storing computer program instructions; When the processor executes the computer program instructions, it implements the mobile work machinery power transmission obstacle sensing method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The storage medium stores at least one executable instruction, which, when executed on an electronic device, causes the electronic device to perform the mobile work machinery power transmission obstacle sensing method as described in any one of claims 1 to 6.

Images