Intelligent storage Internet of Things monitoring and control system for tobacco production

Abstract

The invention discloses an intelligent warehousing Internet of Things monitoring and control system for tobacco production, and the system comprises the following steps: building continuous monitoring data through collecting the temperature, humidity, gas components and inventory identification information in a warehousing environment at multiple points; on the basis, an improved horizon estimation method is adopted to carry out comprehensive estimation on the storage environment and the inventory state, and a result reflecting state change is obtained; further constructing a control barrier function based on the estimation result to form a security constraint which is dynamically adjusted along with the state change; an environment regulation and control instruction is generated under the limitation of safety constraints, and regulation and control operation is executed through ventilation equipment, dehumidification equipment, humidification equipment, gas emission equipment and the like; meanwhile, the operation state of the execution equipment is collected, monitoring data are fed back and updated, an automatic closed-loop control process is formed, and continuous monitoring and safe and stable automatic regulation and control of the tobacco storage environment are achieved. The invention belongs to the technical field of industrial automatic control and Internet of Things.

Description

Technical Field

[0001] This invention relates to the fields of industrial automatic control and Internet of Things (IoT) technology, and in particular to an intelligent warehouse IoT monitoring and control system for tobacco production. Background Technology

[0002] As tobacco production processes become increasingly information-based and automated, the demand for refined control over environmental conditions and inventory status in tobacco warehousing is constantly rising. Currently, tobacco warehousing typically monitors the storage environment by deploying temperature and humidity sensors, gas detection devices, and conducting manual inspections. Ventilation, dehumidification, and humidification equipment are controlled based on empirical thresholds or fixed rules to maintain the quality of stored tobacco.

[0003] Existing tobacco storage monitoring and control technologies still have significant shortcomings. On the one hand, existing systems mostly rely on single-point or limited sensor data for environmental assessment, lacking comprehensive utilization of information from multiple locations and time scales. This makes it difficult to reflect the overall changes in the storage environment and inventory status in a timely and accurate manner, easily leading to delayed status perception or judgment bias. On the other hand, existing control methods generally use static thresholds or simple rule triggers, lacking the ability to predict the evolution trend of environmental status. Moreover, control constraints are mostly fixed settings and cannot be dynamically adjusted according to changes in storage status, easily leading to over-regulation or delayed response, increasing the risk to tobacco quality. In addition, the correlation between monitoring, decision-making, and execution in existing systems is weak. The operating status of execution equipment is difficult to be fed back in a timely manner and participate in subsequent control decisions, making it difficult to form a stable and reliable closed-loop automatic control mechanism.

[0004] Therefore, how to provide an intelligent warehousing IoT monitoring and control system for tobacco production is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] One objective of this invention is to propose an intelligent warehouse IoT monitoring and control system for tobacco production. This invention continuously estimates the warehouse status by integrating multi-source environmental and inventory monitoring data, and generates environmental control commands under safety constraints to achieve automatic control of ventilation, dehumidification, humidification and gas emissions. It has the advantages of timely monitoring, forward-looking control and stable operation.

[0006] The intelligent warehousing IoT monitoring and control system for tobacco production according to an embodiment of the present invention includes the following steps: The multi-source sensing module is used to collect temperature, relative humidity, gas composition concentration and inventory carrier identification information at multiple locations in the storage space, and to add the corresponding collection time and collection location to the collected data to generate raw observation data. The association construction module groups and aligns the raw observation data according to the collection location and storage carrier information to generate a continuous observation sequence. The horizon estimation module constructs a state estimation problem within a sliding time window using a continuous observation sequence as input, and performs an improved horizon estimation based on the evolutionary constraints between the continuous observation sequences, generating a state trajectory and a consistency index. The control barrier function construction module constructs a set of control barrier functions based on the state trajectory, and updates the set of control barrier functions according to the state trajectory and consistency index to generate a dynamic safety constraint model. The constraint control module receives the state trajectory and dynamic safety constraint model, and generates environmental regulation instructions within the feasible control domain defined by the set of control barrier functions. The actuator module receives environmental control commands and executes corresponding ventilation, dehumidification, humidification, and gas emission actions. The closed-loop feedback module collects the operating status of the actuator module to form feedback data and sends it back to the association construction module to update the continuous observation sequence and output the warehouse environment control results that meet the dynamic safety constraint model.

[0007] Optional, the multi-source sensing module includes: The temperature and humidity acquisition unit collects ambient temperature data and relative humidity data at multiple different physical locations within the storage space. The temperature at each location is the real-time ambient temperature value at the corresponding acquisition location, and the relative humidity is the real-time ambient relative humidity value at the corresponding acquisition location. A gas composition acquisition unit acquires gas composition concentration data at the multiple acquisition locations, the gas composition concentration data including oxygen concentration and carbon dioxide concentration reflecting the storage environment status; The inventory carrier acquisition unit reads the inventory carrier identification information from the inventory carrier. The inventory carrier identification information includes a unique identification code for distinguishing the inventory carrier and a batch identification for distinguishing the source of the tobacco. The data acquisition and combination unit combines temperature data, relative humidity data, and gas composition concentration data obtained from the same acquisition location with the corresponding read inventory carrier identification information to form the acquired data within the same acquisition cycle. An additional unit is used to attach a collection time to the collected data using a unified time reference, which is used to identify the collection time corresponding to the collected data. The result output unit appends the acquisition location corresponding to the acquired data as an acquisition location identifier to the acquired data with the acquisition time already appended, forming raw observation data containing the acquisition content, acquisition time, and acquisition location.

[0008] Optionally, the association construction module includes: receiving raw observation data and extracting the acquisition location, storage carrier identification information, and acquisition time; using the combination of acquisition location and storage carrier identification information as the grouping key to group the raw observation data and generate a subset of raw observation data; arranging the raw observation data in the order of acquisition time in each subset of raw observation data to form a sequence of raw observation data; performing time-series alignment processing on the raw observation data sequence according to the sampling period based on a preset unified sampling period; when multiple raw observation data exist in the same sampling period, selecting one of them as the alignment data for the sampling period; when no raw observation data exists in a certain sampling period, generating supplementary data based on the raw observation data in adjacent sampling periods; and connecting the data after grouping, sorting, and time-series alignment processing according to the sampling period order to generate a continuous observation sequence.

[0009] Optionally, the horizon estimation module includes: The sequence receiving unit is used to receive the continuous observation sequence output by the correlation construction module; The change assessment unit is used to calculate the degree of change in the raw observation data corresponding to adjacent acquisition times based on the continuous observation sequence, and generate change assessment results. The adaptive horizon generation unit is used to filter raw observation data from the continuous observation sequence based on the change assessment results, and to correct the filtering results by combining the consistency index output from the previous estimation period, thereby generating an adaptive horizon observation set. The window determination unit is used to determine the sliding time window based on the temporal distribution range of the adaptive horizon observation set, and to form the state estimation input with the adaptive horizon observation set covered by the sliding time window; The state variable sequence construction unit is used to define state variables for each sampling time within a sliding time window and form a state variable sequence. The constraint construction unit is used to construct state evolution constraints between adjacent state variables based on the state variable sequence, and to construct observation consistency constraints between state variables and the adaptive horizon observation set based on the state estimation input. The prior introduction unit is used to introduce the state estimation result output at the end of the previous sliding time window as a prior state constraint, and it works together with the state evolution constraint and the observation consistency constraint on the state variable sequence. The optimization and solution unit is used to construct and solve the state estimation optimization problem based on the state variable sequence, state evolution constraints, observation consistency constraints and prior state constraints, generate the state estimation results within the sliding time window and form the state trajectory; The trajectory consistency evaluation unit is used to generate trajectory consistency results based on the state trajectory. The end correction and output unit is used to constrain and correct the state estimation results corresponding to the end of the state trajectory based on the trajectory consistency results, output the corrected state estimation results, and generate a consistency index based on the trajectory consistency results.

[0010] Optionally, the control barrier function building blocks include: An input receiving unit is used to receive the state trajectory and the consistency index corresponding to the state trajectory; The constraint quantity construction unit is used to construct constraint quantities that characterize the relative safety requirements of the state variables based on the state variables corresponding to each sampling time in the state trajectory. The constraint quantities are directly calculated from the state variables. The barrier function construction unit is used to construct control barrier functions with the constraint quantity as input, so that the control barrier functions can be used to characterize the degree to which the state variables satisfy the safety constraints, and to combine the control barrier functions constructed for different constraint quantities to form a control barrier function set. The constraint model generation unit is used to generate a dynamic safety constraint model for constraining control variables based on the set of control barrier functions. The update judgment unit is used to determine the fit of the current state trajectory with respect to the set of control barrier functions based on the consistency index, and to generate an update judgment result when the fit changes. The online update unit is used to update the control barrier function set with the latest state trajectory as input after generating the update judgment result. The update includes recalculating the constraint quantities corresponding to each control barrier function in the control barrier function set, reconstructing the control barrier function set based on the recalculated constraint quantities, updating the dynamic safety constraint model, and outputting it.

[0011] Optionally, the constraint control module includes: The input receiving unit is used to receive the state trajectory and the dynamic safety constraint model; The control variable definition unit is used to define a set of control variables corresponding to environmental control instructions. The set of control variables includes ventilation control quantity, dehumidification control quantity, humidification control quantity, and gas emission control quantity. The feasible control domain construction unit is used to take the dynamic safety constraint model as the constraint input, apply the set of control barrier functions in the dynamic safety constraint model to the set of control variables, generate a feasible range of values ​​for control variables that satisfies the constraint conditions of the set of control barrier functions, and determine the feasible range of values ​​for control variables as the feasible control domain. The candidate control input generation unit is used to generate a set of candidate control variable values ​​with the end state of the state trajectory as input, and provides the set of candidate control variable values ​​to the feasible control domain construction unit for constraint determination; The constraint determination unit is used to match the set of candidate control variable values ​​with the feasible control domain, retain the candidate control variable values ​​that fall into the feasible control domain and generate a set of feasible candidate control variables. The target control variable selection unit is used to select the value of the target control variable from the set of feasible candidate control variables; The instruction generation unit is used to generate environmental control instructions based on the values ​​of the target control variables.

[0012] Optionally, the actuator module includes: The instruction receiving unit is used to receive environmental control instructions and parse the environmental control instructions to generate ventilation control quantities, dehumidification control quantities, humidification control quantities and gas emission control quantities; The control quantity analysis unit is used to extract target parameters corresponding to ventilation control quantity, dehumidification control quantity, humidification control quantity and gas emission control quantity from environmental control instructions, and associate the target parameters with the corresponding actuators respectively; The ventilation actuator is used to drive the ventilation equipment to start or stop or adjust the operating intensity according to the ventilation control quantity, so as to complete the ventilation action. The dehumidification execution unit is used to drive the dehumidification equipment to start or stop or adjust the operating intensity according to the dehumidification control quantity, so as to complete the dehumidification action. The humidification execution unit is used to drive the humidification equipment to start or stop or adjust the humidification output according to the humidification control quantity, so as to complete the humidification action. The gas emission actuator is used to drive the gas emission device to start or stop or adjust the emission intensity according to the gas emission control quantity, so as to complete the gas emission action.

[0013] Optional, the closed-loop feedback module includes: The status acquisition unit is used to collect the operating status of the actuator module, and to obtain the start / stop status and operating intensity parameters of the ventilation actuator, the start / stop status and operating intensity parameters of the dehumidification actuator, the start / stop status and humidification output parameters of the humidification actuator, and the start / stop status and emission intensity parameters of the gas emission actuator, and generate actuator operating status data. A marking unit is used to attach a feedback time to the actuator operating status data and attach the acquisition position corresponding to the actuator operating status data; The feedback unit is used to send the actuator operating status data at the additional feedback time and acquisition location back to the association construction module; The sequence positioning unit is used to determine the continuous observation sequence corresponding to the operating status data of the actuator based on the acquisition location, and to determine the corresponding sampling period of the operating status data of the actuator in the continuous observation sequence based on the feedback time. The sequence update unit is used to write the operating status data of the actuator into a continuous observation sequence of the corresponding sampling period, forming a continuous observation sequence containing the original observation data and the operating status data of the actuator, and outputting the warehouse environment control result that satisfies the dynamic safety constraint model based on the continuous observation sequence containing the original observation data and the operating status data of the actuator.

[0014] The beneficial effects of this invention are: This invention introduces an improved horizon estimation method to continuously fuse and reconstruct multi-source environmental monitoring data and inventory status information, thereby achieving dynamic and reliable perception of the tobacco storage environment and inventory status. This enables the system to accurately grasp the trend of environmental changes under complex and variable storage conditions, overcoming the problems of lagging state perception and insufficient accuracy caused by relying on single-point monitoring or static analysis in the prior art, thus providing a stable and reliable state basis for subsequent control.

[0015] This invention transforms the safety and quality requirements in tobacco storage into calculable and constrainable safety boundaries by constructing and dynamically updating control barrier functions. It also constrains control commands in real time during the control process, ensuring that environmental control is always carried out within a safe range. This avoids the over-adjustment or response lag problems that are prone to occur under traditional fixed threshold or empirical rule control, thereby improving the safety and stability of storage environment control.

[0016] This invention organically integrates monitoring, state estimation, constraint control, execution, and feedback to form a complete closed-loop automatic control system. This system enables the operating status of the actuators to be fed back in real time and participate in subsequent decision updates, reducing reliance on manual intervention and achieving continuous automatic regulation of the tobacco storage environment. This significantly improves the automation level and operational efficiency of storage management and effectively reduces tobacco quality risks. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 The flowchart is shown below for the intelligent warehouse IoT monitoring and control system for tobacco production proposed in this invention. Figure 2 This is a schematic diagram of the warehouse status reconstruction process based on improved horizon estimation as described in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figures 1-2 The intelligent warehousing IoT monitoring and control system for tobacco production includes the following steps: The multi-source sensing module is used to collect temperature, relative humidity, gas composition concentration and inventory carrier identification information at multiple locations in the storage space, and to add the corresponding collection time and collection location to the collected data to generate raw observation data. The association construction module groups and aligns the raw observation data according to the collection location and storage carrier information to generate a continuous observation sequence. The horizon estimation module constructs a state estimation problem within a sliding time window using a continuous observation sequence as input, and performs an improved horizon estimation based on the evolutionary constraints between the continuous observation sequences, generating a state trajectory and a consistency index. The control barrier function construction module constructs a set of control barrier functions based on the state trajectory, and updates the set of control barrier functions according to the state trajectory and consistency index to generate a dynamic safety constraint model. The constraint control module receives the state trajectory and dynamic safety constraint model, and generates environmental regulation instructions within the feasible control domain defined by the set of control barrier functions. The actuator module receives environmental control commands and executes corresponding ventilation, dehumidification, humidification, and gas emission actions. The closed-loop feedback module collects the operating status of the actuator module to form feedback data and sends it back to the association construction module to update the continuous observation sequence and output the warehouse environment control results that meet the dynamic safety constraint model.

[0020] In this embodiment, the multi-source sensing module includes: The temperature and humidity acquisition unit collects ambient temperature data and relative humidity data at multiple different physical locations within the storage space. The temperature at each location is the real-time ambient temperature value at the corresponding acquisition location, and the relative humidity is the real-time ambient relative humidity value at the corresponding acquisition location. A gas composition acquisition unit acquires gas composition concentration data at the multiple acquisition locations, the gas composition concentration data including oxygen concentration and carbon dioxide concentration reflecting the storage environment status; The inventory carrier acquisition unit reads the inventory carrier identification information from the inventory carrier. The inventory carrier identification information includes a unique identification code for distinguishing the inventory carrier and a batch identification for distinguishing the source of the tobacco. The storage carrier is a physical unit used to hold tobacco products, including pallets, turnover boxes or containers. The storage carrier has a unique identifier to correspond to the storage location and tobacco batch information, and moves within the storage space as the inventory moves. The data acquisition and combination unit combines temperature data, relative humidity data, and gas composition concentration data obtained from the same acquisition location with the corresponding read inventory carrier identification information to form the acquired data within the same acquisition cycle. An additional unit is used to attach a collection time to the collected data using a unified time reference, which is used to identify the collection time corresponding to the collected data. A unified time reference is a standard timing source used within the system to generate consistent time signatures for each acquisition cycle. This timing source is provided by the system clock or a synchronization server and remains consistent across all data acquisition modules. The result output unit appends the acquisition location corresponding to the acquired data as an acquisition location identifier to the acquired data with the acquisition time already appended, forming raw observation data containing the acquisition content, acquisition time, and acquisition location.

[0021] In this embodiment, the association construction module includes: receiving raw observation data and extracting the acquisition location, storage carrier identification information and acquisition time; using the combination of acquisition location and storage carrier identification information as the grouping key to group the raw observation data and generate a subset of raw observation data; arranging the raw observation data in each subset of raw observation data according to the order of acquisition time to form a sequence of raw observation data; performing time-series alignment processing on the raw observation data sequence according to the sampling period based on a preset unified sampling period; when there are multiple raw observation data in the same sampling period, selecting one of them as the alignment data for the sampling period; when there is no raw observation data in a certain sampling period, generating supplementary data based on the raw observation data in adjacent sampling periods; and connecting the data after grouping, sorting and time-series alignment processing according to the sampling period order to generate a continuous observation sequence. The chronological order of data collection refers to arranging multiple raw observation data according to the chronological relationship of the data collection time markers attached to the raw observation data, so as to reflect the time order in which the data were generated. The preset uniform sampling period is a fixed time interval set in the system and used as a time reference for timing alignment. This time interval is configured by the system and remains consistent during data processing. Time alignment processing based on sampling period refers to dividing time intervals with a unified sampling period and assigning the original observation data to each time interval accordingly, so that the data form time alignment results within the same time interval.

[0022] In this embodiment, the horizon estimation module includes: The sequence receiving unit is used to receive the continuous observation sequence output by the correlation construction module; The change assessment unit is used to calculate the degree of change in the raw observation data corresponding to adjacent acquisition times based on the continuous observation sequence, and generate change assessment results. Calculating the degree of change of the raw observation data corresponding to adjacent acquisition times refers to comparing the raw observation data at two consecutive acquisition times and generating a change assessment result to characterize the magnitude of change based on the difference in the corresponding acquisition data values. The adaptive horizon generation unit is used to filter raw observation data from the continuous observation sequence based on the change assessment results, and to correct the filtering results by combining the consistency index output from the previous estimation period, thereby generating an adaptive horizon observation set. Screening raw observation data refers to selecting and processing raw observation data based on the change assessment results, retaining only the raw observation data whose degree of change meets the preset conditions; Correcting the screening results refers to adjusting the original observation data that has been screened by combining the consistency index output from the previous estimation period, in order to correct any possible biases in the screening results. The window determination unit is used to determine the sliding time window based on the temporal distribution range of the adaptive horizon observation set, and to form the state estimation input with the adaptive horizon observation set covered by the sliding time window; The time distribution range refers to the time interval covered by the earliest and latest times of the corresponding acquisition times of each original observation data in the adaptive horizon observation set, and is used to determine the start and end times of the sliding time window; The state variable sequence construction unit is used to define state variables for each sampling time within a sliding time window and form a state variable sequence. Defining a state variable refers to setting a variable to characterize the state of the inventory carrier at each sampling time within a sliding time window, and using this variable as the object for subsequent state estimation and constraint construction; The constraint construction unit is used to construct state evolution constraints between adjacent state variables based on the state variable sequence, and to construct observation consistency constraints between state variables and the adaptive horizon observation set based on the state estimation input. Constructing state evolution constraints between adjacent state variables refers to establishing continuous change constraints between state variables at adjacent sampling times, which are used to limit the relationship between state variables as they change over time. The observation consistency constraint refers to establishing a matching relationship between the observation data in the adaptive horizon observation set and the state variables at the corresponding sampling time, in order to restrict the consistency between the observations generated by the state variables and the actual observation data. The prior introduction unit is used to introduce the state estimation result output at the end of the previous sliding time window as a prior state constraint, and it works together with the state evolution constraint and the observation consistency constraint on the state variable sequence. The optimization and solution unit is used to construct and solve the state estimation optimization problem based on the state variable sequence, state evolution constraints, observation consistency constraints and prior state constraints, generate the state estimation results within the sliding time window and form the state trajectory; Constructing and solving a state estimation optimization problem refers to taking the sequence of state variables as the object to be solved, incorporating state evolution constraints, observation consistency constraints, and prior state constraints into the solution process, and obtaining the state estimation result. The trajectory consistency evaluation unit is used to generate trajectory consistency results based on the state trajectory. The end correction and output unit is used to constrain and correct the state estimation results corresponding to the end of the state trajectory based on the trajectory consistency results, output the corrected state estimation results, and generate a consistency index based on the trajectory consistency results. Constraint correction refers to adjusting the state estimation results at the end of the state trajectory based on the trajectory consistency results, so that the end state is consistent with the overall change of the state trajectory, and is used as the output state.

[0023] In this embodiment, the control barrier function construction module includes: An input receiving unit is used to receive the state trajectory and the consistency index corresponding to the state trajectory; The constraint quantity construction unit is used to construct constraint quantities that characterize the relative safety requirements of the state variables based on the state variables corresponding to each sampling time in the state trajectory. The constraint quantities are directly calculated from the state variables. The barrier function construction unit is used to construct control barrier functions with the constraint quantity as input, so that the control barrier functions can be used to characterize the degree to which the state variables satisfy the safety constraints, and to combine the control barrier functions constructed for different constraint quantities to form a control barrier function set. Constructing a control barrier function refers to mapping the constraint quantities calculated from the state variables into a functional form that characterizes whether the state meets the safety requirements, so that the function can describe the safety constraint relationship of the state variables. The constructed control barrier function combination refers to the aggregation of multiple control barrier functions generated for different constraint quantities to form a set of control barrier functions for simultaneously constraining multiple state variables. The constraint model generation unit is used to generate a dynamic safety constraint model for constraining control variables based on the set of control barrier functions. The dynamic safety constraint model refers to the set of constraint conditions generated by the set of control barrier functions, which is used to apply corresponding safety constraints to the control variables as the state trajectory changes; The update judgment unit is used to determine the fit of the current state trajectory with respect to the set of control barrier functions based on the consistency index, and to generate an update judgment result when the fit changes. Determining the fit refers to substituting the current state trajectory into each control barrier function in the control barrier function set, and evaluating the degree to which the state trajectory satisfies the control barrier function set based on the function output results; A change in the fit refers to a difference in the degree to which the set of control barrier functions satisfies the state trajectory compared to the previous estimation period, which is used to characterize the change in the relationship between safety constraints and the state trajectory. An online update unit is used to update the control barrier function set with the latest state trajectory as input after generating the update judgment result. The update includes recalculating the constraint quantities corresponding to each control barrier function in the control barrier function set, reconstructing the control barrier function set based on the recalculated constraint quantities, updating the dynamic safety constraint model, and outputting it. Recalculation refers to recalculating the constraints used to construct the control barrier function based on the latest state trajectory to reflect the latest changes in the state variables; The recalculated constraint reconstructing control barrier function set refers to updating or replacing the original control barrier functions based on the recalculated constraint quantities, thereby forming a control barrier function set corresponding to the latest state trajectory.

[0024] In this embodiment, the constraint control module includes: The input receiving unit is used to receive the state trajectory and the dynamic safety constraint model; The control variable definition unit is used to define a set of control variables corresponding to environmental control instructions. The set of control variables includes ventilation control quantity, dehumidification control quantity, humidification control quantity, and gas emission control quantity. The feasible control domain construction unit is used to take the dynamic safety constraint model as the constraint input, apply the set of control barrier functions in the dynamic safety constraint model to the set of control variables, generate a feasible range of values ​​for control variables that satisfies the constraint conditions of the set of control barrier functions, and determine the feasible range of values ​​for control variables as the feasible control domain. The candidate control input generation unit is used to generate a set of candidate control variable values ​​with the end state of the state trajectory as input, and provides the set of candidate control variable values ​​to the feasible control domain construction unit for constraint determination; The terminal state of the state trajectory refers to the state estimation result corresponding to the latest sampling time in the state trajectory obtained by horizon estimation within the sliding time window, which is used to characterize the system state at the current time. Constraint determination refers to substituting the values ​​of candidate control variables into the constraint conditions corresponding to the dynamic safety constraint model, and determining whether the values ​​satisfy the constraint requirements of the control barrier function set. The constraint determination unit is used to match the set of candidate control variable values ​​with the feasible control domain, retain the candidate control variable values ​​that fall into the feasible control domain and generate a set of feasible candidate control variables. Matching determination refers to comparing the values ​​of candidate control variables with the range of values ​​of feasible control domains to determine whether the values ​​of candidate control variables fall within the feasible control domains. The target control variable selection unit is used to select the value of the target control variable from the set of feasible candidate control variables; The instruction generation unit is used to generate environmental control instructions based on the values ​​of the target control variables.

[0025] In this embodiment, the actuator module includes: The instruction receiving unit is used to receive environmental control instructions and parse the environmental control instructions to generate ventilation control quantities, dehumidification control quantities, humidification control quantities and gas emission control quantities; The control quantity analysis unit is used to extract target parameters corresponding to ventilation control quantity, dehumidification control quantity, humidification control quantity and gas emission control quantity from environmental control instructions, and associate the target parameters with the corresponding actuators respectively; The ventilation actuator is used to drive the ventilation equipment to start or stop or adjust the operating intensity according to the ventilation control quantity, so as to complete the ventilation action. Executing start-stop or adjusting operating intensity refers to controlling the start-up or stop status of the equipment according to environmental control instructions, or adjusting its operating power, speed or output level while the equipment is running; The dehumidification execution unit is used to drive the dehumidification equipment to start or stop or adjust the operating intensity according to the dehumidification control quantity, so as to complete the dehumidification action. The humidification execution unit is used to drive the humidification equipment to start or stop or adjust the humidification output according to the humidification control quantity, so as to complete the humidification action. The gas emission actuator is used to drive the gas emission device to start or stop or adjust the emission intensity according to the gas emission control quantity, so as to complete the gas emission action. The action coordination unit is used to coordinate the actions of the ventilation execution unit, dehumidification execution unit, humidification execution unit and gas emission execution unit within the same control cycle, so that each execution action is completed synchronously according to the environmental control instructions.

[0026] In this embodiment, the closed-loop feedback module includes: The status acquisition unit is used to collect the operating status of the actuator module, and to obtain the start / stop status and operating intensity parameters of the ventilation actuator, the start / stop status and operating intensity parameters of the dehumidification actuator, the start / stop status and humidification output parameters of the humidification actuator, and the start / stop status and emission intensity parameters of the gas emission actuator, and generate actuator operating status data. A marking unit is used to attach a feedback time to the actuator operating status data and attach the acquisition position corresponding to the actuator operating status data; Feedback time refers to the time when the operating status of the actuator is collected and feedback data is formed. This time marker is generated by a unified time base and is used to characterize the collection time corresponding to the feedback data. The feedback unit is used to send the actuator operating status data at the additional feedback time and acquisition location back to the association construction module; The sequence positioning unit is used to determine the continuous observation sequence corresponding to the operating status data of the actuator based on the acquisition location, and to determine the corresponding sampling period of the operating status data of the actuator in the continuous observation sequence based on the feedback time. The sequence update unit is used to write the operating status data of the actuator into a continuous observation sequence of the corresponding sampling period, forming a continuous observation sequence containing the original observation data and the operating status data of the actuator, and outputting the warehouse environment control result that satisfies the dynamic safety constraint model based on the continuous observation sequence containing the original observation data and the operating status data of the actuator.

[0027] Example 1: To verify the feasibility of this invention in practice, it was applied to the finished tobacco storage scenario of a tobacco production enterprise. This type of storage space is typically a closed structure, with a single warehouse area exceeding several thousand square meters. It contains multiple storage areas for tobacco products stored on pallets and in crates. The storage cycle generally lasts from several weeks to several months, and the storage is highly sensitive to changes in ambient temperature, relative humidity, and the composition of gases within the warehouse. Under traditional management methods, temperature and humidity sensors are usually only installed in a few locations within the warehouse. Environmental control mainly relies on fixed threshold triggers and manual judgment, making it difficult to reflect local environmental changes in a timely manner, and easily leading to problems such as delayed environmental control and excessive adjustments.

[0028] After deploying the intelligent warehousing IoT monitoring and control system described in this invention in this scenario, environmental data acquisition devices are set up at multiple different physical locations within the warehouse space. The number of acquisition points has increased from less than 10 to more than 30, covering the main storage areas. The system continuously collects environmental temperature, relative humidity, and oxygen and carbon dioxide concentration data at each location and binds them to the inventory carrier identification information to form raw observation data with the acquisition time and location. The acquisition cycle is maintained at the minute level, enabling the system to obtain high temporal resolution environmental change information.

[0029] After grouping and aligning the original observation data according to time sequence using the association construction module, a continuous observation sequence is formed. In the initial statistical analysis of the system, it can be observed that the relative humidity difference between different cargo areas can reach as high as 6% to 8%, whereas under traditional monitoring methods, this difference often cannot be identified in a timely manner due to the sparseness of sensor points. With the system of this invention, the continuous observation sequence can stably reflect the environmental change trends in each area, providing reliable input for subsequent state estimation.

[0030] In the horizon estimation module, the system performs improved horizon estimation based on a continuous observation sequence. By evaluating the degree of change in observation data at adjacent sampling times, the system dynamically adjusts the sliding time window length, enabling the state estimation to both smooth short-term fluctuations and reflect environmental change trends in a timely manner. In actual operation data, the fluctuation amplitude of the state estimation results is reduced by approximately 40% compared to the original observation data, while the response lag time to continuous changing trends is reduced by approximately 30%. When the humidity in a local area shows a continuous upward trend, the system can identify risks in advance before the humidity exceeds the traditional control threshold.

[0031] The control barrier function, built based on state trajectories, transforms warehouse safety and quality requirements into dynamic safety constraints. Statistical data shows that during system operation, the feasible range of control variables before the generation of environmental control commands automatically adjusts with state changes. Compared to fixed threshold control methods, the feasible control domain is reduced by approximately 20%, but it always covers safety control requirements. This makes the control process more precise, avoiding excessive ventilation or dehumidification.

[0032] In the constraint control module, the system generates environmental control commands within the feasible control domain. Comparison of operational data reveals that, under the same storage conditions, the number of start-ups and shutdowns of ventilation and dehumidification equipment is reduced by approximately 25% compared to traditional control methods, and the average magnitude of a single control action is reduced by approximately 15%. Despite the reduced control magnitude, the stability of the storage environment is significantly improved, with the daily fluctuation range of relative humidity decreasing from ±4% to within ±2.5%.

[0033] The actuator module drives the ventilation, dehumidification, humidification, and gas emission equipment according to environmental control commands. The closed-loop feedback module synchronously collects the start-stop status and operating intensity of the actuators and writes the operating status data into a continuous observation sequence. Continuous operation statistics show that when the control effect deviates from expectations due to operating limitations of certain equipment, the system can automatically correct the state trajectory and control strategy in subsequent estimation cycles, gradually bringing environmental indicators back to a safe range.

[0034] Comprehensive operational data shows that after adopting the system of this invention, the long-term average deviation of temperature and relative humidity in the storage space is significantly reduced, the duration of abnormal humidity is shortened by about 50%, and quality fluctuations in stored tobacco caused by environmental issues are significantly reduced. Because the control process is more proactive and gentle, the cumulative operating energy consumption of the equipment is reduced by about 10% to 15%. The above data demonstrate that this invention can achieve stable, reliable, and efficient automatic monitoring and control in a real tobacco storage environment, significantly improving the problems of untimely monitoring, inaccurate control, and insufficient closed-loop capability in existing technologies.

[0035] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An intelligent warehousing IoT monitoring and control system for tobacco production, characterized in that, Includes the following modules: The multi-source sensing module is used to collect temperature, relative humidity, gas composition concentration and inventory carrier identification information at multiple locations in the storage space, and to add the corresponding collection time and collection location to the collected data to generate raw observation data. The association construction module groups and aligns the raw observation data according to the collection location and storage carrier information to generate a continuous observation sequence. The horizon estimation module constructs a state estimation problem within a sliding time window using a continuous observation sequence as input, and performs an improved horizon estimation based on the evolutionary constraints between the continuous observation sequences, generating a state trajectory and a consistency index. The control barrier function construction module constructs a set of control barrier functions based on the state trajectory, and updates the set of control barrier functions according to the state trajectory and consistency index to generate a dynamic safety constraint model. The constraint control module receives the state trajectory and dynamic safety constraint model, and generates environmental regulation instructions within the feasible control domain defined by the set of control barrier functions. The actuator module receives environmental control commands and executes corresponding ventilation, dehumidification, humidification, and gas emission actions. The closed-loop feedback module collects the operating status of the actuator module to form feedback data and sends it back to the association construction module to update the continuous observation sequence and output the warehouse environment control results that meet the dynamic safety constraint model.

2. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The multi-source sensing module includes: The temperature and humidity acquisition unit collects ambient temperature data and relative humidity data at multiple different physical locations within the storage space. The temperature at each location is the real-time ambient temperature value at the corresponding acquisition location, and the relative humidity is the real-time ambient relative humidity value at the corresponding acquisition location. A gas composition acquisition unit acquires gas composition concentration data at the multiple acquisition locations, the gas composition concentration data including oxygen concentration and carbon dioxide concentration reflecting the storage environment status; The inventory carrier acquisition unit reads the inventory carrier identification information from the inventory carrier. The inventory carrier identification information includes a unique identification code for distinguishing the inventory carrier and a batch identification for distinguishing the source of the tobacco. The data acquisition and combination unit combines temperature data, relative humidity data, and gas composition concentration data obtained from the same acquisition location with the corresponding read inventory carrier identification information to form the acquired data within the same acquisition cycle. The additional unit adds the acquisition time to the acquired data using a unified time reference; The result output unit appends the acquisition location corresponding to the acquired data as an acquisition location identifier to the acquired data with the acquisition time already appended, forming raw observation data containing the acquisition content, acquisition time, and acquisition location.

3. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The association construction module includes: receiving raw observation data and extracting the acquisition location, storage carrier identification information, and acquisition time; using the combination of acquisition location and storage carrier identification information as the grouping key to group the raw observation data, generating a subset of raw observation data; arranging the raw observation data in the order of acquisition time in each subset of raw observation data to form a raw observation data sequence; performing time-series alignment processing on the raw observation data sequence according to the preset unified sampling period; when multiple raw observation data exist in the same sampling period, selecting one as the alignment data for the sampling period; when no raw observation data exists in a certain sampling period, generating supplementary data based on the raw observation data in adjacent sampling periods; and connecting the grouped, sorted, and time-series aligned data in the order of the sampling periods to generate a continuous observation sequence.

4. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The horizon estimation module includes: The sequence receiving unit is used to receive the continuous observation sequence output by the correlation construction module; The change assessment unit is used to calculate the degree of change in the raw observation data corresponding to adjacent acquisition times based on the continuous observation sequence, and generate change assessment results. The adaptive horizon generation unit is used to filter raw observation data from the continuous observation sequence based on the change assessment results, and to correct the filtering results by combining the consistency index output from the previous estimation period, thereby generating an adaptive horizon observation set. The window determination unit is used to determine the sliding time window based on the temporal distribution range of the adaptive horizon observation set, and to form the state estimation input with the adaptive horizon observation set covered by the sliding time window; The state variable sequence construction unit is used to define state variables for each sampling time within a sliding time window and form a state variable sequence. The constraint construction unit is used to construct state evolution constraints between adjacent state variables based on the state variable sequence, and to construct observation consistency constraints between state variables and the adaptive horizon observation set based on the state estimation input. The prior introduction unit is used to introduce the state estimation result output at the end of the previous sliding time window as a prior state constraint, and it works together with the state evolution constraint and the observation consistency constraint on the state variable sequence. The optimization and solution unit is used to construct and solve the state estimation optimization problem based on the state variable sequence, state evolution constraints, observation consistency constraints and prior state constraints, generate the state estimation results within the sliding time window and form the state trajectory; The trajectory consistency evaluation unit is used to generate trajectory consistency results based on the state trajectory. The end correction and output unit is used to constrain and correct the state estimation results corresponding to the end of the state trajectory based on the trajectory consistency results, output the corrected state estimation results, and generate a consistency index based on the trajectory consistency results.

5. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The control barrier function building module includes: An input receiving unit is used to receive the state trajectory and the consistency index corresponding to the state trajectory; The constraint quantity construction unit is used to construct constraint quantities that characterize the relative safety requirements of state variables based on the state variables corresponding to each sampling time in the state trajectory. The barrier function construction unit is used to construct control barrier functions with the constraint quantity as input, and to combine the control barrier functions constructed for different constraint quantities to form a control barrier function set; The constraint model generation unit is used to generate a dynamic safety constraint model for constraining control variables based on the set of control barrier functions. The update judgment unit is used to determine the fit of the current state trajectory with respect to the set of control barrier functions based on the consistency index, and to generate an update judgment result when the fit changes. The online update unit is used to update the control barrier function set with the latest state trajectory as input after generating the update judgment result. The update includes recalculating the constraint quantities corresponding to each control barrier function in the control barrier function set, reconstructing the control barrier function set based on the recalculated constraint quantities, updating the dynamic safety constraint model, and outputting it.

6. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The constraint control module includes: The input receiving unit is used to receive the state trajectory and the dynamic safety constraint model; A control variable definition unit is used to define a set of control variables, which includes ventilation control variables, dehumidification control variables, humidification control variables, and gas emission control variables. The feasible control domain construction unit is used to take the dynamic safety constraint model as the constraint input, apply the set of control barrier functions in the dynamic safety constraint model to the set of control variables, generate a feasible range of values ​​for control variables that satisfies the constraint conditions of the set of control barrier functions, and determine the feasible range of values ​​for control variables as the feasible control domain. The candidate control input generation unit is used to generate a set of candidate control variable values ​​with the end state of the state trajectory as input, and provides the set of candidate control variable values ​​to the feasible control domain construction unit for constraint determination; The constraint determination unit is used to match the set of candidate control variable values ​​with the feasible control domain, retain the candidate control variable values ​​that fall into the feasible control domain and generate a set of feasible candidate control variables. The target control variable selection unit is used to select the value of the target control variable from the set of feasible candidate control variables; The instruction generation unit is used to generate environmental control instructions based on the values ​​of the target control variables.

7. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The actuator module includes: The instruction receiving unit is used to receive environmental control instructions and parse the environmental control instructions to generate ventilation control quantities, dehumidification control quantities, humidification control quantities and gas emission control quantities. The control quantity analysis unit is used to extract target parameters corresponding to ventilation control quantity, dehumidification control quantity, humidification control quantity and gas emission control quantity from environmental control instructions, and associate the target parameters with the corresponding actuators respectively; The ventilation actuator is used to drive the ventilation equipment to start or stop or adjust the operating intensity according to the ventilation control quantity, so as to complete the ventilation action. The dehumidification execution unit is used to drive the dehumidification equipment to start or stop or adjust the operating intensity according to the dehumidification control quantity, so as to complete the dehumidification action. The humidification execution unit is used to drive the humidification equipment to start or stop or adjust the humidification output according to the humidification control quantity, so as to complete the humidification action. The gas emission actuator is used to drive the gas emission device to start or stop or adjust the emission intensity according to the gas emission control quantity, so as to complete the gas emission action.

8. The intelligent warehousing IoT monitoring and control system for tobacco production according to claim 1, characterized in that, The closed-loop feedback module includes: The status acquisition unit is used to collect the operating status of the actuator module, and to obtain the start / stop status and operating intensity parameters of the ventilation actuator, the start / stop status and operating intensity parameters of the dehumidification actuator, the start / stop status and humidification output parameters of the humidification actuator, and the start / stop status and emission intensity parameters of the gas emission actuator, and generate actuator operating status data. A marking unit is used to attach a feedback time to the actuator operating status data and attach the acquisition position corresponding to the actuator operating status data; The feedback unit is used to send the actuator operating status data at the additional feedback time and acquisition location back to the association construction module; The sequence positioning unit is used to determine the continuous observation sequence corresponding to the operating status data of the actuator based on the acquisition location, and to determine the corresponding sampling period of the operating status data of the actuator in the continuous observation sequence based on the feedback time. The sequence update unit is used to write the operating status data of the actuator into a continuous observation sequence of the corresponding sampling period, forming a continuous observation sequence containing the original observation data and the operating status data of the actuator, and outputting the warehouse environment control result that satisfies the dynamic safety constraint model based on the continuous observation sequence containing the original observation data and the operating status data of the actuator.

Images