Camping car control system and method based on PWM and multi-sensor fusion

Abstract

The invention provides a camping car control system and method based on PWM and multi-sensor fusion. For a pure electric camping car, environment, vehicle state and user context data are collected through a multi-source sensor, and after space-time alignment and confidence evaluation, a multi-level information fusion algorithm is adopted to generate an environment situation and vehicle health state cognitive model. And quantifying the comprehensive utility of each electricity load in combination with the dynamic utility optimization model, and generating a personalized control strategy. The optimal PWM duty ratio of each load is calculated through an adaptive solver under multiple constraints of a lithium battery power limit value, SOC, temperature, health state and the like, and system-level tracking deviation is introduced as a feedback triggering mechanism to realize closed-loop re-optimization. According to the method, collaborative optimization of safety, comfort and energy efficiency of the pure electric camping car under the condition of limited electric energy is achieved, the service life of the battery is effectively prolonged, and the method is suitable for energy management of intelligent limousines and mobile energy storage vehicles.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent vehicle control, in particular to a camping vehicle control system and method based on PWM and multi-sensor fusion. BACKGROUND

[0002] With the maturity of new energy technology, pure electric camping vehicles are gradually becoming an important development direction in the field of outdoor travel. Unlike traditional fuel or hybrid motor homes, pure electric camping vehicles completely rely on on-board power battery systems for power supply, not only for driving the vehicle, but also for providing continuous power for high-power life loads such as air conditioning, refrigerators, induction cookers, lighting, water pumps, audio-visual equipment, inverters, etc. inside the vehicle. Energy supply is completely limited by battery capacity and charging conditions, especially in remote areas or long off-grid use scenarios, the challenge of energy management is particularly prominent.

[0003] The available power and energy output capacity of pure electric camping vehicles are highly dependent on the current state of charge (SOC), operating temperature, and state of health (SOH). In low temperature environments, the battery internal resistance increases and the chemical reaction activity decreases, resulting in a sharp reduction in output power; and in high load or low SOC conditions, excessive discharge can cause battery protection and even permanent damage. In addition, life power loads are mostly nonlinear and intermittent running devices, and their start-stop process is easy to cause current impact and energy waste, and if there is a lack of fine control, it will significantly shorten the endurance time and battery life. SUMMARY

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present application is to provide a camping vehicle control system and method based on PWM and multi-sensor fusion, which solves the problem of extensive energy management and static single control strategy of the prior art camping vehicle.

[0005] To achieve the above-mentioned purposes and other related purposes, the present application provides the following technical solutions: The camping vehicle control method based on PWM and multi-sensor fusion is characterized in that the method comprises the following steps: S1, acquiring multi-source data collected by a sensor module, the multi-source data including environment data, vehicle state data and user context data; S2, processing the multi-source data by using a multi-level information fusion algorithm to generate an environment situation awareness result with confidence evaluation and a vehicle health state evaluation; S3, based on the environment situation awareness result and the vehicle health state evaluation, combining user instructions and user context data, generating a system control strategy through a dynamic utility optimization model; S4, converting the control strategy into a real-time optimization problem with the goal of maximizing the overall performance of the system, considering multiple constraints of the power limit, remaining power and health state of the lithium battery, and calculating the optimal PWM duty cycle instruction set of each power load through an adaptive solver; and smoothing the instruction set to generate the final PWM control signal of each channel; S5, acquiring the execution result through a feedback sensor, and dynamically adjusting the duty cycle of the output PWM signal based on the deviation between the feedback value and the target value.

[0006] To achieve the above technical solutions, the camping vehicle control method based on PWM and multi-sensor fusion of the present application realizes intelligent management through multi-source data perception and closed-loop optimization control. First, the system collects multi-source information including environment data, vehicle state data and user context data through a sensor module, comprehensively perceiving the internal and external environment in which the camping vehicle runs. Then, the multi-level information fusion algorithm is used to process these heterogeneous data in layers to generate an environment situation awareness result with confidence evaluation and a vehicle health state evaluation, providing a reliable basis for decision-making. On this basis, the system control strategy is generated through a dynamic utility optimization model in combination with user instructions and context information, and is converted into a real-time optimization problem with the goal of maximizing the overall performance of the system, and the optimal PWM duty cycle instruction set of each power load is calculated by an adaptive solver under multiple constraints of lithium battery power, power and health state. After smoothing, the final PWM control signal is output for execution, and the actual execution result is monitored through a feedback sensor, and the PWM duty cycle is dynamically adjusted according to the deviation, forming a closed-loop control to ensure the accuracy and stability of the system response.

[0007] In an embodiment of the present application, S1 includes S1.1, dynamically activating the corresponding sensor subset and adjusting its sampling frequency according to the current running mode of the camping vehicle; S1.2, time stamping and spatial position labeling the raw data from each sensor, and transmitting through the vehicle Ethernet or CAN FD protocol to form a structured data stream with a unified space-time reference; S1.3, calculating the confidence of each sensor data in real time at the data acquisition end, and using a time-space graph neural network-based completion and correction algorithm to reconstruct the low-confidence data marked therein to generate a time-space data set suitable for multi-source information fusion and having time-space consistency.

[0008] To achieve the technical solutions described above, the application improves the efficiency and reliability of multi-source data acquisition by dynamic perception management and spatio-temporal consistency data preprocessing. In the S1 stage, the system dynamically activates the corresponding sensor subset according to the current operating mode of the camping vehicle and adjusts its sampling frequency, realizing on-demand perception, reducing system power consumption and communication load while ensuring data effectiveness; by stamping each sensor data with a timestamp and a spatial position label, and based on the vehicle-mounted Ethernet or CANFD protocol transmission, a structured data stream with a unified spatio-temporal reference is constructed, solving the misalignment problem of multi-source heterogeneous data in time and space; further, the confidence of each sensor data is calculated in real time at the data acquisition end, and the low-confidence data is reconstructed using a spatio-temporal graph neural network-based completion and correction algorithm, effectively dealing with data anomalies caused by sensor failure, signal drift or environmental interference, generating a data set with high spatio-temporal consistency. This mechanism significantly improves the quality and fusion availability of raw data, providing accurate, synchronized and reliable input basis for subsequent environmental situation awareness and intelligent control decision-making, enhancing the robustness and response capability of the system in complex dynamic environments.

[0009] In an embodiment of the application, S2 includes: S2.1, performing data-level fusion, receiving the spatio-temporal data set from S1, and generating a set of spatio-temporal consistent underlying physical quantities through timestamp alignment, coordinate system unification and standardization processing; S2.2, performing feature-level fusion, based on the underlying physical quantities, identifying high-level environmental and vehicle features through a pre-set feature extraction model; S2.3, performing situation-level fusion, fusing and reasoning the high-level environmental and vehicle features with user context data from S1, and outputting environmental situation awareness results and vehicle comprehensive state evaluation with confidence levels; S2.4, performing decision-level fusion, based on the environmental situation awareness results and vehicle comprehensive state evaluation, calculating the real-time demand priority of each electrical load through a multi-attribute decision fusion algorithm.

[0010] The above technical solutions are implemented, and through the multi-level data fusion strategy, accurate perception and efficient management of the camping vehicle environment and vehicle state are realized. First, in the data level fusion step, the system performs timestamp alignment, coordinate system unification and standardization processing on the received spatio-temporal data set, ensuring the consistency and accuracy of the underlying physical quantities. Then, in the feature level fusion step, the pre-set feature extraction model is used to identify high-level environment and vehicle features from these underlying physical quantities, laying the foundation for further analysis. Next, in the situation level fusion step, the above high-level features are combined with user context data, and the environment situation awareness result and vehicle comprehensive state evaluation with confidence level are output through fusion reasoning, providing a basis for decision-making. Finally, in the decision level fusion step, according to the information obtained before, a multi-attribute decision fusion algorithm is used to calculate the real-time demand priority of each electrical load, so as to realize accurate power distribution and optimized user experience. This method not only improves the adaptability and safety of the camping vehicle in complex environments, but also prolongs the service life of the equipment and improves the energy use efficiency through optimized power resource management.

[0011] In an embodiment of the present application, the calculation of the real-time demand priority of each electrical load by the multi-attribute decision fusion algorithm includes constructing a device priority evaluation index system, including safety influence factor, comfort contribution degree, energy consumption level and user preference intensity; introducing domain expert experience, using triangular fuzzy numbers to represent the relative importance between each index, and constructing a fuzzy judgment matrix; performing consistency check on the fuzzy judgment matrix to ensure the logical rationality of expert judgment; when the check fails, feedback correction suggestions and re-collect evaluation data; when the consistency check passes, use the eigenvector method to extract the evaluation weight of each index from the fuzzy judgment matrix; combine the real-time environment situation and vehicle state data to calculate the comprehensive priority score of each device; sort the devices based on the priority score to generate a dynamically adjustable demand priority list.

[0012] To achieve the technical solutions described above, the application realizes intelligent and dynamic sequencing of the priority of the electrical load by using a multi-attribute decision fusion algorithm. When calculating the real-time demand priority, an equipment priority evaluation index system is first constructed, and multi-dimensional factors such as safety influence factor, comfort contribution, energy consumption level and user preference intensity are comprehensively considered to comprehensively reflect the value weight of the load under different operating scenarios. Based on the experience of field experts, the relative importance between each index is quantitatively expressed by using a triangular fuzzy number, a fuzzy judgment matrix is constructed, and the uncertainty and fuzziness in subjective judgment are effectively handled; the consistency of the matrix is checked to ensure the logical rationality of the expert evaluation, and if the check fails, a feedback correction suggestion is fed back and the data is re-collected to ensure the reliability of the decision basis. After the check is passed, the fuzzy judgment matrix is solved by using the characteristic vector method to extract the evaluation weight of each index. The real-time environmental situation awareness result and the vehicle comprehensive state evaluation data output by S2 are combined to dynamically calculate the comprehensive priority score of each electrical load, and a sortable and updatable demand priority list is generated accordingly. This method not only realizes scientific decision-making under multi-objective and multi-constraint conditions, improves the objectivity and robustness of the priority determination, but also enhances the adaptability of the system in complex and uncertain scenarios by introducing fuzzy logic and feedback correction mechanism, effectively balances the relationship between safety, comfort, energy efficiency and user personalized demand, and significantly improves the intelligent level of the camping car energy management and user experience.

[0013] In an embodiment of the application, S3 includes S3.1, constructing a dynamic utility function of each controlled electrical load, the dynamic utility function taking environmental situation awareness results, vehicle state evaluation and user context data as input variables, outputting a quantitative value for representing the comprehensive utility level of the load under the current operating context, the comprehensive utility level reflecting the contribution degree of the load to the overall performance of the system; S3.2, based on user preference configuration or historical behavior data, dynamically adjusting the utility preference weight between different utility dimensions to realize personalized configuration of multi-objective optimization; S3.3, the dynamic utility function and the utility preference weight together constitute the objective function of the optimization problem in step S4, and an executable control instruction set containing the target working parameters and priority constraints of each load is generated.

[0014] To achieve the technical solutions described above, the application realizes the personalization of control targets and the context-adaptive optimization by constructing a dynamic utility function. In S3.1, a dynamic utility function is established for each controlled power consumption load, and the environmental situation awareness result, the vehicle health state evaluation and the user context data are used as input variables to quantize the comprehensive utility level of the load in the current running scene, so as to effectively reflect the contribution degree of the load to the overall performance (such as safety, comfort and energy efficiency) of the system; in S3.2, based on the user preference configuration or historical behavior data, the utility preference weights of each dimension in the utility function are dynamically adjusted, so that the optimization target can be flexibly changed according to the user habits and actual needs, and the personalization configuration of multi-objective optimization is realized; in S3.3, the dynamic utility function and the utility preference weights are integrated into a target function, and an executable control instruction set is generated in combination with the load priority constraint to provide a clear target orientation for subsequent optimization solution. The mechanism breaks through the limitations of the traditional fixed weight control strategy, and significantly improves the adaptability and decision intelligence level of the system in complex and variable scenes.

[0015] The comprehensive utility level is obtained by the following formula, The comprehensive utility level of the i-th load in the current running context is obtained by the following formula, The number of utility dimensions is obtained by the following formula, The utility preference weight of the i-th utility dimension at time t is obtained by the following formula, The state variable of the i-th utility dimension at time t is obtained by the following formula, The state variable of the i-th utility dimension at time t is obtained by the following formula, The nonlinear utility mapping function is obtained by the following formula,

[0016] The comprehensive utility level is obtained by the following formula, ) The i-th dimension reference weight is obtained by the following formula, The modulation gain coefficient is obtained by the following formula, The current environmental situation vector is obtained by the following formula, The current vehicle state vector is obtained by the following formula, The element-wise multiplication is obtained by the following formula, ​​​​​​​​​​​The model achieves a dynamic balance and personalized response to user comfort, energy efficiency, and safety under different operating conditions.

[0017] In one embodiment of the present invention, step S4 includes: S4.1, constructing an optimization model with the goal of maximizing the overall system efficiency. The objective function covers comfort, safety, energy efficiency, and equipment lifespan, and the constraints include the real-time power output capability, remaining power, temperature characteristics, and health status of the lithium battery; S4.2, converting the optimization model into a mixed integer programming method to solve for continuous PWM duty cycle variables and discrete equipment start / stop variables. Based on the objective function and constraints of the optimization model, the optimal control sequence in the future prediction time domain is calculated by an adaptive solver, and the PWM duty cycle command at the current moment is extracted as the optimal solution; S4.3, performing adaptive smoothing processing based on load characteristics on the obtained PWM duty cycle command set to generate PWM control signals for each channel; S4.4, establishing a multi-level quality assurance mechanism including feasibility verification, safety checks, and performance evaluation. The optimal control sequence output by the adaptive solver is verified at three levels. When any verification fails, the system automatically switches to a rule-based degradation control strategy.

[0018] To achieve the above technical solution, this method aims to maximize the overall system efficiency through an optimization model. This model considers multiple factors such as comfort, safety, energy efficiency, and equipment lifespan, and incorporates the real-time status of the lithium battery as a constraint. This method not only improves the system's automation and intelligence levels but also effectively extends equipment lifespan, enhances energy utilization efficiency, and improves user safety and comfort.

[0019] The power output capability of the lithium battery is limited by its current state of charge, temperature, and health. The following power output capability constraint model is adopted: in, express The actual output power at any given time; The reference power factor is the nominal value of the battery's rated power. For a moment The state of charge of the battery; For a moment The internal temperature of the battery; The optimal operating temperature for battery performance; This is the temperature sensitivity coefficient; It is a Gaussian temperature decay factor used to characterize the nonlinear suppression effect of high and low temperatures on battery output power. For a moment The battery health state is determined according to the battery health state estimation model, and the power output capability constraint model is used to realize fine and dynamic modeling of the battery output capability, so that the operation safety is ensured and the energy utilization efficiency is improved.

[0020] In an embodiment of the present application, the S4.4 includes: S4.4.1, three-level verification is performed on the PWM duty cycle instruction set output by the adaptive solver, the three-level verification includes hard constraint satisfaction checking, multi-step forward-looking performance and safety evaluation based on a system model, and logical consistency verification based on a rule base; S4.4.2, when any one of the verifications fails, a corresponding rule-based degradation control strategy is selected and switched to from a preset degradation strategy library according to the type of the exception and the current running scene; S4.4.3, the degradation switching process is smoothed, and the exception and the degradation case are recorded for offline optimization of the system.

[0021] To implement the above technical solution, the S4.4 ensures the safety and reliability of the optimization control instruction by constructing a multi-level and multi-dimensional quality guarantee mechanism. Specifically, after the adaptive solver outputs the PWM duty cycle instruction, the system performs three-level verification in turn: first, the hard constraint satisfaction checking is performed to ensure that the instruction does not exceed the physical boundaries such as the power and temperature of the lithium battery; second, multi-step forward-looking simulation is performed based on the system dynamic model to evaluate the performance trend and potential safety risk in the future time domain; and finally, the logical consistency verification is performed in combination with the preset rule base to prevent unreasonable control combinations such as “the air conditioner and the window being opened at the same time”. When any one of the verifications fails, the system intelligently matches and switches to a corresponding rule-based degradation control strategy according to the type of the exception (such as exceeding, conflict, and prediction violation) and the current running scene (such as parking, driving, and charging), so as to ensure the normal operation of the basic function. At the same time, the switching process is smoothly transitioned to avoid impact caused by sudden changes of the execution mechanism, and the exception event and the degradation response process are completely recorded for subsequent offline analysis and model optimization. The mechanism significantly improves the fault tolerance and functional safety of the system, realizes the organic unification of “optimal control” and “safety bottom line”, and effectively deals with complex and variable actual working conditions.

[0022] In an embodiment of the present application, the adaptive solver is configured to: at the start of each rolling optimization period, call a lightweight digital twin model of the recreational vehicle, based on the current environmental situation and user intention, perform simulation deduction of energy consumption, comfort and battery loss on a plurality of candidate control strategies in parallel; and select a Pareto optimal candidate strategy set as an initial solution or constraint boundary adjustment basis for the mixed integer nonlinear programming solution, so as to improve the solving efficiency and global optimality.

[0023] The adaptive solver first calls the embedded lightweight digital twin model at the start of each rolling optimization cycle, combines the current environmental situation (such as external temperature, light intensity) and user intention (such as setting mode, travel plan), and performs parallel simulation deduction on multiple candidate control strategies to predict the performance of each strategy in terms of energy consumption, cabin comfort, and battery loss. Through multi-objective optimization analysis, the Pareto optimal candidate strategy set is selected to provide high-quality initial solutions for the mixed integer nonlinear programming solver or as the basis for dynamically adjusting the constraint boundary and target weight. This mechanism significantly shortens the convergence time of complex optimization problems, improves the solving efficiency and numerical stability, and enhances the global optimality, avoiding falling into a local suboptimal solution. Through the pre-evaluation of the digital twin model, the system can achieve a better long-term performance balance while ensuring real-time performance, effectively improving energy utilization efficiency, prolonging battery life, and enhancing user comfort experience.

[0024] In an embodiment of the present application, S5 specifically comprises: S5.1, obtaining the actual operation parameters of each actuator through the feedback sensor, comparing the actual operation parameters with the target value to generate a device-level error signal; using a fusion of feedforward compensation and parameter self-tuning feedback control law to dynamically adjust the duty cycle of the PWM signal output to each actuator, realizing device-level control; S5.2, real-time evaluation of system-level tracking deviation of system-level key performance indicators, when the deviation indicates that the current control strategy is mismatched with the environment or system state, triggering the re-execution of steps S2 and S3, dynamically updating the environmental situation awareness result, vehicle state evaluation and control strategy, and then adjusting the PWM instruction generated by S4; S5.3, continuously record and analyze historical error data and system running context, when detecting that the system-level tracking deviation repeatedly occurs in similar environment or running scene, starting the model parameter online calibration mechanism, dynamically adjusting the utility preference weight in the dynamic utility function constructed in step S3 and the constraint boundary parameter in the optimization model solved in step S4, realizing continuous optimization of the control strategy.

[0025] To achieve the technical solutions described above, the S5 realizes continuous improvement of control precision and strategy adaptability by constructing a three-level collaborative mechanism of "device-level feedback + system-level closed loop + online learning". Specifically, at the device level, the actual operating parameters of the actuator are collected by using feedback sensors to generate a device-level error signal, and a feedback control law combining feedforward compensation and parameter self-tuning is used to dynamically adjust the PWM duty cycle, effectively suppressing disturbances and improving response accuracy; at the system level, the tracking effect of key performance indicators (such as cabin temperature deviation and battery SOC change rate) is evaluated in real time, and once significant deviation is detected indicating that the current control strategy is mismatched with the actual working condition, the environmental perception and state evaluation module is triggered to re-execute, dynamically updating the situation awareness and priority decision, and then optimizing the control instruction generation; at the same time, the system continuously records historical errors and operating contexts, and if the deviation repeatedly occurs in similar scenarios, the online calibration mechanism of model parameters is started to automatically adjust the utility preference weight or optimize the model constraint boundary, realizing the autonomous evolution of the control strategy. This method not only guarantees high-precision tracking at the device level, but also endows the system with adaptive and self-optimizing capabilities in complex and variable environments, significantly improving overall control performance, energy efficiency and user experience.

[0026] The system-level tracking deviation is calculated by the following formula: wherein, is the system-level tracking deviation; , , is a weight coefficient, and satisfies ; is the actual measured temperature in the cabin; is the target temperature set by the user; is the maximum allowed temperature difference; is the temperature tracking error term; is the current actual state of charge of the battery; is the predicted value of the battery SOC; is the reference SOC value; is the battery SOC prediction deviation term; is an oscillation indicator function, taking a value of 0 or 1; is the oscillation index term; and a system-level deviation threshold is set. If , execute S2 and S3 and guide S4 to recalculate the optimal PWM duty cycle command.

[0027] The system-level tracking deviation is quantitatively evaluated by integrating temperature control accuracy, battery state prediction consistency and control stability in three dimensions. Realize closed-loop feedback and dynamic re-optimization, effectively improve system robustness and control precision.

[0028] In an embodiment of the present application, a PWM and multi-sensor fusion-based control method and system for a camping vehicle are provided. The system includes a multi-source perception module for collecting multi-source data, including environmental data, vehicle state data, and user context data, and generating a structured data stream with a unified space-time reference; a data processing and fusion module for processing the multi-source data using a multi-level information fusion algorithm to generate an environmental situation awareness result with confidence evaluation and a vehicle health state evaluation; a control strategy generation module for generating a system control strategy based on the environmental situation awareness result and the vehicle health state evaluation, combining user instructions and user context data, and through a dynamic utility optimization model; an optimization solving module for converting the control strategy into a real-time optimization problem with the goal of maximizing system comprehensive performance, considering multiple constraints of power limits, remaining power, and health state of lithium batteries, and calculating an optimal PWM duty cycle instruction set for each power load through an adaptive solver; a signal generation and execution module for smoothing the optimal PWM duty cycle instruction set, generating final PWM control signals for each channel, and outputting to each power load; and a feedback control module for obtaining execution results through feedback sensors and dynamically adjusting the PWM signal duty cycle based on the deviation between the feedback value and the target value.

[0029] To achieve the above technical solutions, the system described in the present application collects environmental, vehicle state, and user context data in real time through a multi-source perception module, and generates a structured data stream with a unified space-time reference to provide comprehensive information support for upper-layer decision-making; a data processing and fusion module uses a multi-level information fusion algorithm to deeply process the data, outputs an environmental situation awareness result with confidence evaluation and a vehicle health state evaluation, and improves cognitive accuracy; a control strategy generation module combines the above evaluation results, user instructions, and context information, constructs a dynamic utility optimization model, and generates a control strategy that takes into account comfort, safety, energy efficiency, and equipment life; an optimization solving module converts the strategy into a real-time optimization problem subject to multiple constraints such as lithium battery power limits, SOC, and SOH, and efficiently solves an optimal PWM duty cycle instruction set for each power load through an adaptive solver; a signal generation and execution module performs load characteristic adaptive smoothing processing on the instruction set, generates smooth PWM control signals, and drives the execution mechanism; and a feedback control module collects actual running states in real time through sensors, dynamically adjusts PWM output based on the deviation, and forms a closed-loop correction. The system realizes intelligent collaboration of the whole link from perception to execution, significantly improves the energy utilization efficiency, control accuracy, and user experience of the camping vehicle, and enhances the robustness and safety of the system under complex working conditions.

[0030] As described above, the camping vehicle control system and method based on PWM and multi-sensor fusion of the present application have the following beneficial effects: The present application provides a highly integrated and adaptive intelligent control solution in particular for pure electric camping vehicles that completely rely on battery power supply and do not use fuel auxiliary energy, significantly improving energy utilization efficiency and system reliability. First, through multi-source sensor fusion technology, the system real-time perceives external environment, cabin state, battery working condition and user behavior, constructs a high-confidence environment situation and vehicle health evaluation model, and provides accurate decision basis for energy scheduling under the pure electric platform. Second, the dynamic utility optimization model is introduced, which quantifies the comfort contribution, energy consumption level and user preference of each power load as a comprehensive utility value, and dynamically adjusts the priority combined with the battery state, to ensure that critical needs are met under limited power conditions, and to avoid "power drop" or "function interruption" caused by blind power supply. The present application fully considers the off-grid characteristics of pure electric camping vehicles, designs a degradation control strategy and fault tolerance mechanism, which can still maintain basic functions when sensors fail or optimization fails, and ensure basic living needs of users. The overall scheme not only greatly improves the energy utilization efficiency and prolongs the service life of the battery, but also significantly enhances the comfort, safety and intelligent experience of users in long-distance travel, wild camping and other scenarios under the condition of no external energy support, with outstanding technical progress and broad market application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 A flowchart showing the method of the present application is shown. DETAILED DESCRIPTION

[0032] The embodiments of the present application will be described in detail by specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in the specification. It should be noted that the following examples and features in the examples can be combined with each other without conflict.

[0033] Please refer to Figure 1The application provides a camping vehicle control method based on PWM and multi-sensor fusion, which comprises the following steps: S1, acquiring multi-source data collected by a sensor module, wherein the multi-source data comprises environmental data, vehicle state data and user context data; S2, processing the multi-source data by using a multi-level information fusion algorithm to generate an environmental situation awareness result with confidence evaluation and vehicle health state evaluation; S3, based on the environmental situation awareness result and vehicle health state evaluation, combining user instructions and user context data, generating a system control strategy through a dynamic utility optimization model; S4, converting the control strategy into a real-time optimization problem with the goal of maximizing the overall performance of the system, considering the multiple constraints of the power limit, residual capacity and health state of the lithium battery, and calculating the optimal PWM duty cycle instruction set of each power load through an adaptive solver; and performing smoothing processing on the instruction set to generate the final PWM control signal of each channel; S5, acquiring the execution result through a feedback sensor, and dynamically adjusting the duty cycle of the output PWM signal based on the deviation between the feedback value and the target value.

[0034] The working principle of the above technical solution is that the camping vehicle control method based on PWM and multi-sensor fusion of the application realizes intelligent management through multi-source data perception and closed-loop optimization control. First, the system collects multi-source information including environmental data, vehicle state data and user context data through a sensor module, comprehensively perceives the internal and external environment of the camping vehicle running. Then, the multi-level information fusion algorithm is used to process these heterogeneous data in layers to generate an environmental situation awareness result with confidence evaluation and vehicle health state evaluation, providing a reliable basis for decision-making. On this basis, the system control strategy is generated through a dynamic utility optimization model by combining user instructions and context information, and is converted into a real-time optimization problem with the goal of maximizing the overall performance of the system. Under the multiple constraints of lithium battery power, capacity and health state, the optimal PWM duty cycle instruction set of each power load is calculated by an adaptive solver. After smoothing processing, the final PWM control signal is output for execution. The actual execution result is monitored through a feedback sensor, and the PWM duty cycle is dynamically adjusted according to the deviation, forming a closed-loop control to ensure the accuracy and stability of the system response.

[0035] The beneficial effects of the above technical solutions are that the multi-level information fusion algorithm integrates the multi-source data of the environment, the vehicle and the user, improves the accuracy and reliability of the environment situation awareness and the vehicle state evaluation, and provides a solid foundation for intelligent decision-making. By constructing a dynamic utility optimization model and combining real-time system constraints, a system control strategy that takes into account comfort, safety and energy efficiency can be automatically generated, realizing personalized and adaptive control. An adaptive solver is used to solve the optimal PWM duty cycle command set under multiple constraints such as lithium battery power limit, remaining power and health state, effectively ensuring the safe operation and service life of the energy system. The PWM command set is smoothed and combined with feedback deviation to dynamically adjust the output, significantly improving the stability and response accuracy of the control process. The overall scheme realizes fine and intelligent PWM control of the electric load of the camping car, taking into account user demand, system safety and energy efficiency in complex and variable use scenarios.

[0036] S1 includes S1.1, according to the current running mode of the camping car, dynamically activating the corresponding sensor subset and adjusting its sampling frequency; S1.2, the original data from each sensor is time-stamped and spatially labeled, and is transmitted through the vehicle-mounted Ethernet or CAN FD protocol to form a structured data stream with a unified space-time reference; S1.3, the confidence of each sensor data is calculated in real time at the data acquisition end, and for the low-confidence data marked therein, a data reconstruction algorithm based on a space-time graph neural network is used for data reconstruction to generate a space-time data set suitable for multi-source information fusion and having space-time consistency.

[0037] The present application improves the efficiency and reliability of multi-source data acquisition through dynamic perception management and space-time consistent data preprocessing. In the S1 stage, the system dynamically activates the corresponding sensor subset and adjusts its sampling frequency according to the current running mode of the camping car, realizes on-demand perception, reduces system power consumption and communication load while ensuring data effectiveness; by time-stamping and spatially labeling each sensor data and transmitting it based on the vehicle-mounted Ethernet or CAN FD protocol, a structured data stream with a unified space-time reference is constructed, solving the misalignment problem of multi-source heterogeneous data in time and space; further, the confidence of each sensor data is calculated in real time at the data acquisition end, and for low-confidence data, a data reconstruction algorithm based on a space-time graph neural network is used for reconstruction, effectively dealing with data anomalies caused by sensor failure, signal drift or environmental interference, generating a data set with high space-time consistency. This mechanism significantly improves the quality and fusion availability of the original data, providing accurate, synchronized and reliable input for subsequent environment situation awareness and intelligent control decisions, enhancing the robustness and response capability of the system in complex dynamic environments.

[0038] The S2 includes: S2.1, performing data-level fusion, receiving the spatio-temporal data set from S1, and generating a set of spatio-temporally consistent underlying physical quantities through timestamp alignment, coordinate system unification and standardization processing; S2.2, performing feature-level fusion, based on the underlying physical quantities, identifying high-level environmental and vehicle features through a pre-set feature extraction model; S2.3, performing situation-level fusion, fusing and reasoning the high-level environmental and vehicle features with the user context data from S1, and outputting the environmental situation awareness result and vehicle comprehensive state evaluation with confidence level; S2.4, performing decision-level fusion, based on the environmental situation awareness result and vehicle comprehensive state evaluation, calculating the real-time demand priority of each electrical load through a multi-attribute decision fusion algorithm.

[0039] Through the multi-level data fusion strategy, the precise perception and efficient management of the camping vehicle environment and vehicle state are realized. First, in the data-level fusion step, the system performs timestamp alignment, coordinate system unification and standardization processing on the received spatio-temporal data set to ensure the consistency and accuracy of the underlying physical quantities. Then, in the feature-level fusion step, the pre-set feature extraction model is used to identify high-level environmental and vehicle features from these underlying physical quantities, laying the foundation for further analysis. Next, in the situation-level fusion step, the above high-level features are combined with user context data to output the environmental situation awareness result and vehicle comprehensive state evaluation with confidence level through fusion reasoning, providing a basis for decision-making. Finally, in the decision-level fusion step, according to the information obtained before, a multi-attribute decision fusion algorithm is used to calculate the real-time demand priority of each electrical load, thereby realizing precise power distribution and optimized user experience. This method not only improves the adaptability and safety of the camping vehicle in complex environments, but also prolongs the service life of the equipment and improves the energy use efficiency through optimized power resource management.

[0040] The calculation of the real-time demand priority of each electrical load through the multi-attribute decision fusion algorithm includes: constructing a device priority evaluation index system, including safety influence factor, comfort contribution degree, energy consumption level and user preference intensity; introducing domain expert experience, using triangular fuzzy numbers to represent the relative importance between indicators, and constructing a fuzzy judgment matrix; performing consistency check on the fuzzy judgment matrix to ensure the logical rationality of expert judgment; when the check fails, feedback correction suggestions and re-collect evaluation data; when the consistency check passes, use the eigenvector method to extract the evaluation weight of each indicator from the fuzzy judgment matrix; combine the real-time environmental situation and vehicle state data to calculate the comprehensive priority score of each device; sort the devices based on the priority score to generate a dynamically adjustable demand priority list.

[0041] The application realizes the intelligent and dynamic sequencing of the priority of the electrical load by a multi-attribute decision fusion algorithm. In the calculation of the real-time demand priority, an equipment priority evaluation index system is first constructed, and multi-dimensional factors such as safety influence factor, comfort contribution, energy consumption level and user preference intensity are comprehensively considered to fully reflect the value weight of the load under different operating scenarios. Based on the experience of field experts, the relative importance between each index is quantitatively expressed by using triangular fuzzy numbers, a fuzzy judgment matrix is constructed, and the uncertainty and fuzziness in subjective judgment are effectively handled; through consistency checking of the matrix, the logical rationality of the expert evaluation is ensured, if the checking fails, feedback correction suggestions are given and data are re-collected to ensure the reliability of the decision basis. After the checking is passed, the fuzzy judgment matrix is solved by using the characteristic vector method to extract the evaluation weight of each index. Combined with the real-time environmental situation awareness results and vehicle comprehensive state evaluation data output by S2, the comprehensive priority score of each electrical load is dynamically calculated, and a sortable and updatable demand priority list is generated accordingly. This method not only realizes scientific decision-making under multi-objective and multi-constraint, improves the objectivity and robustness of the priority determination, but also enhances the adaptability of the system in complex and uncertain scenarios through the introduction of fuzzy logic and feedback correction mechanism, effectively balances the relationship between safety, comfort, energy efficiency and user personalized demand, and significantly improves the intelligent level of the camping car energy management and user experience.

[0042] The S3 includes S3.1, constructing a dynamic utility function of each controlled electrical load, the dynamic utility function taking environmental situation awareness results, vehicle state evaluation and user context data as input variables, outputting a quantitative value for representing the comprehensive utility level of the load under the current operating context, the comprehensive utility level reflecting its contribution to the overall performance of the system; S3.2, based on user preference configuration or historical behavior data, dynamically adjusting the utility preference weight between different utility dimensions to realize personalized configuration of multi-objective optimization; S3.3, the dynamic utility function and the utility preference weight together constitute the objective function of the optimization problem in step S4, and generate an executable control instruction set containing the target working parameters and priority constraints of each load.

[0043] This invention achieves personalized and context-adaptive optimization of control objectives by constructing dynamic utility functions. In S3.1, a dynamic utility function is established for each controlled electrical load, using environmental situational awareness results, vehicle health status assessment, and user context data as input variables to quantify its comprehensive utility level in the current operating scenario, effectively reflecting the load's contribution to the overall system performance (such as safety, comfort, and energy efficiency). In S3.2, based on user preference configurations or historical behavior data, the utility preference weights of each dimension in the utility function are dynamically adjusted, allowing the optimization objective to flexibly change according to user habits and actual needs, achieving personalized configuration for multi-objective optimization. In S3.3, the dynamic utility function and its utility preference weights are integrated into an objective function, and combined with load priority constraints to generate an executable control instruction set, providing a clear objective guide for subsequent optimization solutions. This mechanism overcomes the limitations of traditional fixed-weight control strategies, significantly improving the system's adaptability and decision-making intelligence in complex and changing scenarios.

[0044] The overall utility level is obtained using the following formula. Among them, the For the first The overall utility level of a workload in the current runtime context; The number of utility dimensions; Represented as the first Each utility dimension at time The utility preference weights, and satisfying ; For the first Each utility dimension at time The state variables; This is a predefined nonlinear utility mapping function used to characterize the nonlinear response characteristics of the utility value in this dimension as the state changes.

[0045] The It can be obtained through the following formula. ) in, The table is the first Each dimension has a benchmark weight; This is the modulation gain coefficient, and its value range is... Control the sensitivity of the environment and state to the weights; Represented as the current environmental situation vector; Represented as the current vehicle state vector; Indicates element-wise multiplication; subscript Indicates taking the first The model achieves a dynamic balance and personalized response to user comfort, energy efficiency, and safety under different operating conditions.

[0046] S4 includes: S4.1, constructing an optimization model with the goal of maximizing the overall system efficiency. The objective function covers comfort, safety, energy efficiency, and equipment lifespan. The constraints include the real-time power output capability, remaining power, temperature characteristics, and health status of the lithium battery; S4.2, transforming the optimization model into a mixed integer programming method to solve for continuous PWM duty cycle variables and discrete equipment start-stop variables. Based on the objective function and constraints of the optimization model, the optimal control sequence in the future prediction time domain is calculated through an adaptive solver, and the PWM duty cycle command at the current moment is extracted as the optimal solution. S4.3 Perform adaptive smoothing processing on the obtained PWM duty cycle instruction set based on load characteristics to generate PWM control signals for each channel; S4.4 Establish a multi-level quality assurance mechanism including feasibility verification, safety check and performance evaluation, and perform three-level verification on the optimal control sequence output by the adaptive solver. When any verification fails, automatically switch to a rule-based degradation control strategy.

[0047] This method aims to maximize the overall system performance through an optimization model that considers factors such as comfort, safety, energy efficiency, and equipment lifespan, while incorporating the real-time status of the lithium battery as a constraint. This approach not only enhances the system's automation and intelligence levels but also effectively extends equipment lifespan, improves energy efficiency, and enhances user safety and comfort.

[0048] In one embodiment, the power output capability of the lithium battery is limited by its current state of charge, temperature, and health status. The following power output capability constraint model is adopted: in, express The actual output power at any given time; The reference power factor is the nominal value of the battery's rated power. For a moment The state of charge of the battery; For a moment The internal temperature of the battery; The optimal operating temperature for battery performance; This is the temperature sensitivity coefficient; It is a Gaussian temperature decay factor used to characterize the nonlinear suppression effect of high and low temperatures on battery output power. For a moment The battery health state. The power output capability constraint model realizes fine and dynamic modeling of the battery output capability, guarantees operation safety, and improves energy utilization efficiency.

[0049] The S4.4 includes S4.4.1, three-level verification of the PWM duty cycle instruction set output by the adaptive solver, including hard constraint satisfaction checking, multi-step forward-looking performance and safety evaluation based on a system model, and logical consistency verification based on a rule base; S4.4.2, when any level of verification fails, according to the type of exception and the current running scene, adaptively selecting and switching to the corresponding rule-based degradation control strategy from the preset degradation strategy library; S4.4.3, smoothing the degradation switching process, and recording the exception and degradation cases for offline optimization of the system.

[0050] The S4.4 ensures the safety and reliability of the optimization control instruction by constructing a multi-level and multi-dimensional quality assurance mechanism. Specifically, after the adaptive solver outputs the PWM duty cycle instruction, the system performs three-level verification in turn: first, hard constraint satisfaction checking is performed to ensure that the instruction does not exceed the physical boundaries of the lithium battery power and temperature; second, multi-step forward-looking simulation is performed based on the system dynamic model to evaluate the performance trend and potential safety risks in the future time domain; and finally, logical consistency verification is carried out in combination with the preset rule base to prevent unreasonable control combinations such as "air conditioner and window opening at the same time". When any level of verification fails, the system intelligently matches and switches to the corresponding rule-based degradation control strategy according to the type of exception (such as exceeding, conflict, and prediction violation) and the current running scene (such as parking, driving, and charging), ensuring normal operation of basic functions. At the same time, smooth transition control is implemented for the switching process to avoid shocks caused by sudden changes in the actuator, and the exception events and degradation response process are recorded completely for subsequent offline analysis and model optimization. This mechanism significantly improves the fault tolerance and functional safety of the system, realizes the organic unification of "optimal control" and "safety bottom line", and effectively deals with complex and variable actual working conditions.

[0051] The adaptive solver is configured to: at the start of each rolling optimization period, call the lightweight digital twin model of the camper, based on the current environmental situation and user intention, carry out simulation and deduction of energy consumption, comfort and battery loss for multiple candidate control strategies in parallel; select the Pareto optimal candidate strategy set as the initial solution or constraint boundary adjustment basis for the mixed integer nonlinear programming solution, to improve the solving efficiency and global optimality.

[0052] The adaptive solver first calls the embedded lightweight digital twin model at the start of each rolling optimization cycle, combines the current environmental situation (such as external temperature, light intensity) and user intent (such as setting mode, travel plan), and performs parallel simulation and deduction on multiple candidate control strategies to predict their performance in terms of energy consumption, cabin comfort, and battery loss. Through multi-objective optimization analysis, the Pareto optimal candidate strategy set is selected to provide high-quality initial solutions for the mixed integer nonlinear programming solver or as the basis for dynamically adjusting the constraint boundary and objective weight. This mechanism significantly shortens the convergence time of complex optimization problems, improves the solving efficiency and numerical stability, and enhances the global optimality, avoiding falling into local suboptimal solutions. Through the pre-evaluation of the digital twin model, the system can achieve better long-term performance balance while ensuring real-time performance, effectively improving energy efficiency, prolonging battery life, and enhancing user comfort experience.

[0053] The S5 specifically includes: S5.1, obtaining the actual operation parameters of each actuator through the feedback sensor, comparing the actual operation parameters with the target value to generate a device-level error signal; adopting a fusion of feedforward compensation and parameter self-tuning feedback control law to dynamically adjust the duty cycle of the PWM signal output to each actuator, realizing device-level control; S5.2, real-time evaluation of system-level tracking deviation of system-level key performance indicators, when the deviation indicates that the current control strategy is mismatched with the environment or system state, triggering the re-execution of steps S2 and S3, dynamically updating the environmental situation awareness result, vehicle state evaluation and control strategy, and then adjusting the PWM instruction generated by S4; S5.3, continuously recording and analyzing historical error data and system running context, when detecting that the system-level tracking deviation repeatedly occurs in similar environments or running scenarios, starting the model parameter online calibration mechanism, dynamically adjusting the utility preference weight in the dynamic utility function constructed in step S3 and the constraint boundary parameter in the optimization model solved in step S4, realizing continuous optimization of the control strategy.

[0054] The S5 realizes continuous improvement of control accuracy and strategy adaptability by constructing a three-level collaborative mechanism of "device-level feedback + system-level closed loop + online learning". Specifically, at the device level, the actual operating parameters of the actuator are collected by the feedback sensor to generate a device-level error signal, and a feedback control law combining feedforward compensation and parameter self-tuning is used to dynamically adjust the PWM duty cycle, effectively suppressing disturbances and improving response accuracy; at the system level, the tracking effect of key performance indicators (such as cabin temperature deviation and battery SOC change rate) is evaluated in real time, and once significant deviation is detected indicating that the current control strategy is mismatched with the actual working condition, the environmental perception and state evaluation module is triggered to re-execute, dynamically updating the situation awareness and priority decision, and then optimizing the control instruction generation; at the same time, the system continuously records historical errors and operating contexts, and if the deviation repeatedly occurs in similar scenarios, the online calibration mechanism of model parameters is started to automatically adjust the utility preference weight or optimize the model constraint boundary, realizing the autonomous evolution of the control strategy. This method not only guarantees high-precision tracking at the device level, but also endows the system with adaptive and self-optimizing capabilities in complex and variable environments, significantly improving overall control performance, energy efficiency and user experience.

[0055] The system-level tracking deviation is calculated by the following formula: wherein, is the system-level tracking deviation; , , is the weight coefficient, and satisfies ; is the actual measured temperature in the cabin; is the target temperature set by the user; is the maximum allowed temperature difference; is the temperature tracking error term; is the current actual state of charge of the battery; is the predicted value of the battery SOC; is the reference SOC value; is the battery SOC prediction deviation term; is an oscillation indicator function, taking values of 0 or 1; is the oscillation index term; and set the system-level deviation threshold , if , trigger the execution of S2 and S3 and guide S4 to recalculate the optimal PWM duty cycle command.

[0056] The system-level tracking deviation is quantitatively evaluated by integrating temperature control accuracy, battery state prediction consistency and control stability in three dimensions. Realize closed-loop feedback and dynamic re-optimization, effectively improve system robustness and control accuracy.

[0057] In another embodiment of the present application, a camping car control system based on PWM and multi-sensor fusion is provided, comprising a multi-source perception module for collecting multi-source data, the multi-source data including environmental data, vehicle state data and user context data, and generating a structured data stream with a unified space-time reference; a data processing and fusion module for processing the multi-source data using a multi-level information fusion algorithm to generate an environmental situation awareness result with confidence evaluation and a vehicle health state evaluation; a control strategy generation module for generating a system control strategy based on the environmental situation awareness result and the vehicle health state evaluation, combining user instructions and user context data, and through a dynamic utility optimization model; an optimization solving module for converting the control strategy into a real-time optimization problem with the goal of maximizing system comprehensive performance, considering multiple constraints of power limit, remaining power and health state of lithium battery, and calculating an optimal PWM duty cycle instruction set for each power load through an adaptive solver; a signal generation and execution module for smoothing the optimal PWM duty cycle instruction set, generating final PWM control signals for each channel, and outputting to each power load; a feedback control module for obtaining execution results through feedback sensors, and dynamically adjusting the PWM signal duty cycle based on the deviation between the feedback value and the target value.

[0058] The system described in the present application collects environmental, vehicle state and user context data in real time through a multi-source perception module, and generates a structured data stream with a unified space-time reference, providing comprehensive information support for upper-level decision-making; the data processing and fusion module uses a multi-level information fusion algorithm to deeply process the data, outputs an environmental situation awareness result with confidence evaluation and a vehicle health state evaluation, and improves cognitive accuracy; the control strategy generation module combines the above evaluation results, user instructions and context information, constructs a dynamic utility optimization model, and generates a control strategy that takes into account comfort, safety, energy efficiency and equipment life; the optimization solving module converts the strategy into a real-time optimization problem subject to multiple constraints such as lithium battery power limit, SOC and SOH, and efficiently solves the optimal PWM duty cycle instruction set for each power load through an adaptive solver; the signal generation and execution module performs load characteristic adaptive smoothing processing on the instruction set, generates smooth PWM control signals and drives the execution mechanism; the feedback control module collects actual running state in real time through sensors, dynamically adjusts the PWM output based on the deviation, and forms a closed-loop correction. The system realizes intelligent collaboration of the whole link from perception to execution, significantly improves the energy utilization efficiency, control accuracy and user experience of the camping car, and enhances the robustness and safety of the system under complex working conditions.

[0059] The above embodiments merely illustrate the principles of the application and its efficacy, and are not intended to limit the application. All equivalent modifications or changes made by those with ordinary knowledge in the art without departing from the spirit and technical ideas disclosed in the application shall be covered by the claims of the application.

Claims

1. A camping car control method based on PWM and multi-sensor fusion, characterized in that, The method comprises the following steps: S1, acquiring multi-source data collected by a sensor module, the multi-source data comprising environmental data, vehicle state data and user context data; S2, processing the multi-source data using a multi-level information fusion algorithm to generate an environmental situation awareness result with confidence evaluation and a vehicle health state evaluation; S3, based on the environmental situation awareness result and the vehicle health state evaluation, combining user instructions and user context data, generating a system control strategy through a dynamic utility optimization model; S4, converting the control strategy into a real-time optimization problem with the goal of maximizing system comprehensive performance, considering multiple constraints of power limit, remaining capacity and health state of lithium batteries, and calculating the optimal PWM duty cycle instruction set of each electrical load through an adaptive solver; and smoothing the instruction set to generate the final PWM control signal of each channel; S5, acquiring the execution result through a feedback sensor, and dynamically adjusting the duty cycle of the output PWM signal based on the deviation between the feedback value and the target value.

2. The camping car control method based on PWM and multi-sensor fusion according to claim 1, characterized in that, The S1 comprises, S1.1, according to the current running mode of the camper, dynamically activating the corresponding sensor subset and adjusting its sampling frequency; S1.2, time-stamping and spatial position labeling the raw data from each sensor, and transmitting through the vehicle-mounted Ethernet or CAN FD protocol to form a structured data stream with a unified space-time reference; S1.3, calculating the confidence of each sensor data in real time at the data acquisition end, and using a time-space graph neural network-based completion and correction algorithm to reconstruct the low-confidence data marked therein, to generate a time-space data set suitable for multi-source information fusion and having time-space consistency.

3. The camping car control method based on PWM and multi-sensor fusion according to claim 2, characterized in that, The S2 comprises: S2.1, performing data-level fusion, receiving the time-space data set from S1, and generating a set of time-space consistent underlying physical quantities through time stamp alignment, coordinate system unification and standardization processing; S2.2, performing feature-level fusion, based on the underlying physical quantities, identifying high-level environmental and vehicle features through a pre-set feature extraction model; S2.3, performing situation-level fusion, fusing and reasoning the high-level environmental and vehicle features with the user context data from S1, and outputting the environmental situation awareness result with confidence level and the vehicle comprehensive state evaluation; S2.4, performing decision-level fusion, based on the environmental situation awareness result and the vehicle comprehensive state evaluation, calculating the real-time demand priority of each electrical load through a multi-attribute decision fusion algorithm.

4. The camping car control method based on PWM and multi-sensor fusion according to claim 3, characterized in that, The calculation of the real-time demand priority of each electrical load through the multi-attribute decision fusion algorithm comprises, constructing a device priority evaluation index system, including safety influence factor, comfort contribution degree, energy consumption level and user preference intensity; introducing domain expert experience, using triangular fuzzy numbers to represent the relative importance between indicators, and constructing a fuzzy judgment matrix; performing consistency check on the fuzzy judgment matrix to ensure the logical rationality of expert judgment; when the test fails, feedback correction suggestions and re-collect evaluation data; When the consistency check passes, the eigenvector method is used to extract the evaluation weight of each index from the fuzzy judgment matrix; Combined with real-time environmental situation and vehicle state data, the comprehensive priority score of each device is calculated; Based on the priority score, the devices are sorted to generate a dynamically adjustable demand priority list.

5. The camping car control method based on PWM and multi-sensor fusion according to claim 1, characterized in that, The S3 includes, S3.1, constructing a dynamic utility function of each controlled electrical load, the dynamic utility function taking environmental situation awareness results, vehicle state evaluation and user context data as input variables, and outputting a quantitative value for representing the comprehensive utility level of the load under the current running context, which reflects its contribution to the overall performance of the system; S3.2, based on user preference configuration or historical behavior data, dynamically adjusting the utility preference weight between different utility dimensions to realize personalized configuration of multi-objective optimization; S3.3, the dynamic utility function and the utility preference weight together constitute the objective function of the optimization problem in step S4, and an executable control instruction set containing target working parameters and priority constraints of each load is generated.

6. The PWM and multi-sensor fusion-based control method for a camper vehicle according to claim 1, wherein, The S4 includes, S4.1, constructing an optimization model with the goal of maximizing the overall performance of the system, the objective function covering comfort, safety, energy efficiency and device life, and the constraint conditions including real-time power output capability, remaining power, temperature characteristics and health status of lithium batteries; S4.2, converting the optimization model into a mixed integer programming method to solve continuous PWM duty variable and discrete device start-stop variable, based on the objective function and constraint conditions of the optimization model, the optimal control sequence in the future prediction time domain is calculated through an adaptive solver, and the PWM duty instruction at the current time is extracted as the optimal solution; S4.3, performing adaptive smoothing processing on the PWM duty instruction set obtained by solving based on the characteristics of the load to generate PWM control signals for each channel; S4.4, establishing a multi-quality guarantee mechanism including feasibility verification, safety check and performance evaluation to perform three-level verification on the optimal control sequence output by the adaptive solver, and when any verification fails, automatically switching to a rule-based degradation control strategy.

7. The camping car control method based on PWM and multi-sensor fusion according to claim 6, characterized in that, The S4.4 includes, S4.4.1, performing three-level verification on the PWM duty instruction set output by the adaptive solver, the three-level verification including hard constraint satisfaction check, multi-step forward-looking performance and safety evaluation based on the system model, and logical consistency check based on the rule base; S4.4.2, when any level of verification fails, adaptively selecting and switching to the corresponding rule-based degradation control strategy from the preconfigured degradation strategy library according to the exception type and the current running scenario; S4.4.3, smoothing the degradation switching process and recording the exception and degradation cases for offline optimization of the system.

8. The camping car control method based on PWM and multi-sensor fusion according to claim 6, characterized in that, The adaptive solver is configured to: at the start of each rolling optimization period, call the lightweight digital twin model of the camper, based on the current environmental situation and user intention, to carry out simulation deduction of energy consumption, comfort and battery loss for multiple candidate control strategies in parallel; The Pareto-optimal candidate strategy set is screened out as an initial solution or constraint boundary adjustment basis for solving a mixed integer nonlinear programming, so as to improve the solving efficiency and global optimality.

9. The camping car control method based on PWM and multi-sensor fusion according to claim 5, characterized in that, The S5 specifically includes: S5.1, acquiring actual operation parameters of each actuator through a feedback sensor, comparing the actual operation parameters with target values to generate a device-level error signal, and adopting a fusion of feedforward compensation and parameter self-tuning feedback control law to dynamically adjust the duty cycle of the PWM signal output to each actuator, so as to realize device-level control; S5.2, real-time evaluation of system-level tracking deviation of system-level key performance indicators, when the deviation indicates that the current control strategy is mismatched with the environment or system state, triggering re-execution of steps S2 and S3, dynamically updating the environment situation awareness result, vehicle state evaluation and control strategy, and then adjusting the PWM instruction generated in S4; S5.3, continuously recording and analyzing historical error data and system running context, when it is detected that the system-level tracking deviation repeatedly occurs in similar environments or running scenarios, starting an online calibration mechanism of model parameters, dynamically adjusting the utility preference weight in the dynamic utility function constructed in step S3 and the constraint boundary parameter in the optimization model solved in step S4, and realizing continuous optimization of the control strategy.

10. A system for implementing the method of any one of claims 1 to 9, characterized in that, It includes, A multi-source perception module for collecting multi-source data including environment data, vehicle state data and user context data, and generating a structured data stream with a unified space-time reference; A data processing and fusion module for processing the multi-source data using a multi-level information fusion algorithm to generate an environment situation awareness result with confidence evaluation and a vehicle health state evaluation; A control strategy generation module for generating a system control strategy based on the environment situation awareness result and the vehicle health state evaluation, combining user instructions and user context data, and through a dynamic utility optimization model; An optimization solving module for converting the control strategy into a real-time optimization problem with the goal of maximizing the overall system performance, considering multiple constraints of power limit, remaining capacity and health state of the lithium battery, and calculating an optimal PWM duty cycle instruction set of each power load through an adaptive solver; A signal generation and execution module for smoothing the optimal PWM duty cycle instruction set, generating final PWM control signals of each channel, and outputting to each power load; A feedback control module for acquiring execution results through a feedback sensor, and dynamically adjusting the duty cycle of the output PWM signal based on the deviation between the feedback value and the target value.

Images