Integrated marketing advertisement release and management system based on Internet of Things

Abstract

The invention relates to the technical field of Internet advertisement services, and discloses an integrated marketing advertisement publishing and management system based on the Internet of Things, and the system comprises a state monitoring unit which monitors a terminal input stream in real time to obtain the interaction frequency and the interface depth; a load calculation unit which generates a load increment based on the weighted product, accumulates the load increment to a load counter, and periodically deducts a count value at a preset rate; the threshold value generation unit is used for mapping the real-time count value into a dynamic admission base price; according to the invention, a dynamic pricing mechanism based on leaky bucket points is constructed, and the inertial attenuation characteristic of a load counter is utilized to quantify the nonlinear regression process of the cognitive load of a user, so that the problems of mistaken advertisement touch and thinking flow interruption caused by instantaneous state detection are solved, and the accuracy of advertisement display is improved. And intelligent dynamic balance between commercial resource configuration and user experience protection is realized.

Description

Technical Field

[0001] This invention relates to an integrated marketing advertising publishing and management system based on the Internet of Things, belonging to the field of Internet advertising service technology. Background Technology

[0002] The current internet advertising distribution architecture follows the mobile terminal traffic monetization logic, taking the device's online status and screen activation time as the core marketing resources, and distributing promotional content in real time according to preset bidding strategies and exposure metrics. Unlike the information browsing priority of mobile terminals, the primary attribute of IoT devices such as in-vehicle central control or industrial tablets is physical world task execution tools.

[0003] Under the existing distribution mechanism, the system only monitors the availability of network channels and screen layers, lacking the ability to perceive and quantify the criticality of the physical tasks currently being carried out by the device. This extensive delivery based on a single traffic perspective results in advertising information being forcibly displayed when users are performing high-load tasks such as parameter fine-tuning, emergency communication, or real-time monitoring, thus cutting off the user's physical operation link. For example, Chinese invention patent CN118115200A discloses a big data-based digital integrated marketing service system. This system focuses on collecting users' historical shopping records and social media account information to construct static user profile tags, and then distributes these tags to various terminals. Pushing product information to users' devices, based on static matching mechanisms using identity tags and historical preferences, is essentially a crude approach to traditional traffic-driven advertising. Its core flaw lies in the fact that the system only focuses on the relevance between advertising content and user interests, ignoring the exclusivity and uniqueness of the device's current physical task in the IoT context. This technology cannot analyze the instantaneous cognitive load of the user's device operation in real time. The push logic is executed as soon as the network is smooth and the tag is matched. It lacks a dynamic value assessment mechanism for the timing of operations and is prone to forcibly popping up when users are performing high-risk or high-load tasks such as reversing a vehicle or fine-tuning precise industrial control parameters, thus cutting off the user's physical operation link.

[0004] Therefore, how to construct a method that maps the physical task load of IoT devices to advertising access barriers in real time, and achieves dynamic distribution of high-value information while ensuring that the critical task flow is not disturbed, has become the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: An integrated marketing advertising publishing and management system based on the Internet of Things, the system comprising: The status monitoring unit is used to monitor the input command flow of IoT terminals in real time, obtain the frequency of interaction commands per unit time as the interaction frequency value, and obtain the layer depth of the current display interface as the interface depth value. The load calculation unit is used to maintain a load counter in memory. When an input instruction stream is received, a load increment is generated based on the weighted product of the interaction frequency value and the interface depth value. The load increment is accumulated to the current value of the load counter. At the same time, the current value of the load counter is periodically deducted at a preset decay rate using timing logic until the value reaches zero. The threshold generation unit is used to read the current count value in the load counter; and convert the current count value into a dynamic entry price for advertising bidding using a preset mapping rule. The bidding processing unit is used to receive advertising requests and their real-time bids for IoT terminals; compare the real-time bids with the dynamic admission threshold; and generate a rejection instruction or a downgrade instruction if the real-time bid is lower than the dynamic admission threshold.

[0006] Preferably, when the load calculation unit performs the deduction operation, if the current value of the load counter is greater than zero and less than the decay step size of a single deduction cycle, the load counter is directly reset to zero. The status monitoring unit also includes a service identifier extraction module, which is used to obtain the service class identifier of the currently running process. The load calculation unit also includes a locking logic module, which ignores the current state of the input instruction stream when the service class identifier matches the preset list of highly sensitive services, forcibly locks the value of the load counter to the preset maximum threshold, and suspends the deduction operation of the timed logic.

[0007] Preferably, the threshold generation unit includes a trend response module for monitoring the rate of change of the value in the load counter; when the rate of change indicates that the current count value is in an upward trend, the current count value is directly output for calculating the dynamic admission floor price; when the rate of change indicates that the current count value is in a downward trend, a state hold timer is started, and the peak value before the decline is maintained within the preset window period of the state hold timer until the preset window period expires and no reverse upward trend is detected.

[0008] Preferably, the bidding processing unit is also used to execute asynchronous distribution logic that separates creative materials from licenses; the bidding processing unit includes: a preloading module, used to send the encrypted creative material data package of the candidate advertising object to the local storage area of ​​the IoT terminal in advance while the dynamic admission threshold price is maintained below a preset idle threshold; a key distribution module, used to generate a display authorization key for the target advertising object when the real-time bid is higher than the dynamic admission threshold price, and send the display authorization key as a distribution instruction to the IoT terminal; the display authorization key is used to trigger the IoT terminal to decrypt and render the corresponding encrypted creative material data package in the local storage area.

[0009] Preferably, the threshold generation unit further includes an idle compensation module, used to continuously monitor the duration for which the current count value is below a preset low threshold; and to calculate the threshold compensation amount using the following formula. : ,in, The preset compensation coefficient, This is the current value of the duration. The preset base time, It is a unit step function; the threshold generation unit adds the threshold compensation amount to the dynamic access floor price. After the duration exceeds the benchmark time, the dynamic access floor price shows a monotonically increasing trend with the increase of the duration.

[0010] Preferably, the bidding processing unit includes a display mode adjustment module for storing a table showing the correspondence between different advertising display modes and price ranges. When the real-time bid is lower than the dynamic entry threshold but higher than the preset base threshold, the display mode adjustment module retrieves the corresponding table and generates a downgraded display instruction that instructs the IoT terminal to display the advertising content in a non-intrusive notification bar format. The non-intrusive notification bar format is limited to a preset key operation area that does not obscure the current functional interface of the IoT terminal.

[0011] Preferably, the attenuation rate in the load calculation unit is configured as a variable value; the load calculation unit includes an adaptive rate adjustment module, which is used to count the frequency of the load counter value reaching the maximum threshold within a preset historical period; when the frequency exceeds the preset frequency threshold, the attenuation rate value is reduced to extend the retention time of the load counter value.

[0012] Preferably, when the status monitoring unit obtains the interface depth value, it performs the following operations: traverses the interface tree structure of the current active window of the IoT terminal; identifies the hierarchical index of the current focused element in the interface tree structure; normalizes the value of the hierarchical index to an integer within a preset range as the interface depth value; and the load calculation unit uses the interface depth value as a weighting coefficient to perform weighted calculation on the interaction frequency value to generate the load increment.

[0013] Preferably, the bidding processing unit is also used to record the source identifier of the rejected ad delivery request after generating the rejection display instruction; if the same source identifier triggers the rejection display instruction more than the preset limit within a preset time window, a blocking instruction is generated for the source identifier, and all ad delivery requests from the source identifier are directly discarded within the subsequent preset cooldown period without comparison.

[0014] Preferably, the system is deployed in a distributed architecture including a cloud server and edge terminals; the status monitoring unit and the load calculation unit reside on the edge terminal, which are used to maintain the load counter locally and calculate the current count value; the edge terminal sends a status update request to the cloud server only when the current count value changes beyond a preset threshold; the threshold generation unit and the bidding processing unit reside on the cloud server, which are used to update the dynamic admission floor price in response to the status update request and process advertising placement requests from third-party advertising exchange platforms.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In integrated marketing for the Internet of Things (IoT), a dynamic access mechanism based on terminal operating parameters is constructed. A value adjustment interface is established between the physical functional logic of the device and the commercial distribution logic of the advertising system. The depth value and interaction heat value of the current task attribute function are collected and mapped to quantitative indicators representing the exclusivity and mutual exclusion of the task. Based on this, a dynamic fluctuation access threshold is generated. This abandons the traditional solution of relying solely on the device's network status or screen being on as a usable resource, without increasing hardware perception costs, and accurately capturing the user's cognitive surplus time. When the bidding for external advertising information is sufficient to penetrate the cost barrier built by the current physical task, the information is allowed to be sent, ensuring that the core tool attributes of the device are not interfered with, eliminating ineffective adversarial exposure, and improving the attention and return on investment of the final advertising slot.

[0016] 2. By introducing a service identifier-based logic short-circuit and forced locking mechanism, the blind spots of the single behavior statistical model in handling observational tasks are overcome. Before the mutual exclusion index calculation module executes the regular calculation process, it compares the service class identifiers of the foreground running process with the list of highly sensitive services. When low-interaction-heat but high-cognitive-load silent high-risk tasks such as reversing camera, navigation guidance, or industrial monitoring are detected, the bypass is based on the regular judgment of click frequency, and the task mutual exclusion entropy is directly locked as the blocking threshold. By utilizing the existing process management information at the operating system level, the risk of idle misjudgment that may occur in general algorithms in scenarios where users only watch and do not move is eliminated. This ensures that the advertising distribution system maintains zero intrusion in scenarios involving personal safety or critical operations, and meets equipment engineering safety standards.

[0017] 3. By utilizing asymmetric time-delay filtering logic, the system addresses the issues of bidding signal oscillation and thought process interruption caused by transient user operation jitter. For the upward trend of task mutual exclusion entropy, a direct response strategy is adopted, increasing the entry threshold at millisecond speeds to ensure immediate protection against sudden physical tasks. For the downward trend, a state-holding window is introduced, forcibly extending the high threshold lock-in time to filter out brief pauses or hesitation periods during user operations. Time-domain nonlinear processing logic is used to financialize pricing based on user cognitive inertia, preventing the system from accidentally blocking high-value advertisements and avoiding prematurely sending interfering information before the user has fully finished thinking about the task, ensuring the stable commercial usability of idle resource signals output to the server. Attached Figure Description

[0018] Figure 1 This is a flowchart of the cognitive load quantification and dynamic access bidding decision-making process of this invention; Figure 2 This is a time-domain comparison of the anti-jitter characteristics of the access threshold price of this invention under intermittent operation; Figure 3 This is an architecture diagram of the cloud-edge collaborative distributed advertising publishing system of the present invention. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0020] This invention provides an integrated marketing advertising publishing and management system based on the Internet of Things (IoT). It includes a status monitoring unit and a load calculation unit residing at the edge of the IoT terminal, and a threshold generation unit and a bidding processing unit residing on a cloud server. This system architecture, through the collaboration of edge computing and cloud decision-making, achieves real-time quantification of the physical task load of the terminal and dynamic adjustment of advertising access barriers. The status monitoring unit monitors the input command flow of the IoT terminal in real time to quantify the user's operational load, targeting the interaction frequency value, i.e. This unit sets a sliding time window to count the number of interrupt signals generated by the input subsystem within the window. This count includes touchscreen click events, physical button trigger events, and valid voice command events, based on the interface depth value. This unit retrieves the node object of the currently active window in the graphical user interface tree structure, traces back up the parent linked list of the node object to the root node, counts the number of nodes in the backtracking path, and maps this number to a normalized integer. In specific engineering settings, the interface depth value of the standby interface is... Defined as The interface depth value of the first-level menu interface Defined as The interface depth value of the specific application function execution interface Defined as greater than or equal to The integer; the load calculation unit maintains a load counter in memory, i.e. This is used to characterize the user's cognitive load, responding to the input command stream detected by the state monitoring unit, which bases its response on the interaction frequency value. With interface depth value The weighted product generates the load increment, which is then added to the current value of the load counter. This accumulation process follows the formula... ,in The load equivalent of a single basic operation. For depth-weighted coefficients, At the current moment, this unit utilizes timing logic to decay at a preset rate. The current value of the load counter is periodically decremented until it reaches zero. This decrementing process follows the formula... ,in To reduce the timeout period, this mechanism ensures that the load counter maintains a non-zero value for a preset time after the physical operation stops, protecting the user's mental inertia. The timing logic is anchored to the SoC's high-precision hardware counter interrupt. In the interrupt service routine, the operating system scheduling layer is bypassed, and an atomic subtraction instruction is directly executed on the counter address mapped to SRAM, controlling the time-dimensional decay jitter within a certain range. Within the clock cycle, ensure the temporal fidelity of cognitive load quantization.

[0021] The status monitoring unit further obtains the service class identifier of the currently running process. The load calculation unit compares this service class identifier with a preset list of highly sensitive services, including process characteristic strings representing functions such as reversing camera, emergency call, or industrial alarm. When the service class identifier matches any item in the list of highly sensitive services, the load calculation unit ignores the current state of the input command stream, locks the value of the load counter to a preset maximum threshold, and suspends the deduction operation of the timing logic until the service class identifier no longer matches the list. The threshold generation unit reads the current count value in the load counter and uses a mapping rule to convert the current count value into a dynamic entry threshold price for advertising bidding. The calculation formula is as follows: ,in The minimum price for dynamic entry is [price]. The preset base price for the ad space. To mitigate the interference penalty coefficient, this unit monitors the rate of change of the value in the load counter. When the rate of change indicates that the current count value is trending upward, the unit directly outputs the current count value to calculate the dynamic admission floor price. When the rate of change indicates that the current count value is trending downward, the unit starts a state hold timer and maintains the output of the peak value before the decline within a preset window period of the state hold timer until the preset window period expires and no reverse upward trend is detected. The threshold generation unit continuously monitors the duration for which the current count value is below a preset low threshold. When the duration exceeds the preset baseline time At that time, this unit uses the formula Calculate the threshold compensation amount ,in As the time value drift coefficient, this unit adds the threshold compensation amount to the dynamic access floor price, so that the dynamic access floor price shows a monotonically increasing trend as the idle time increases.

[0022] The bidding processing unit receives advertising requests and their real-time bids from IoT terminals. This unit executes asynchronous distribution logic that separates creative content from the license. When the dynamic admission threshold remains below a preset idle threshold, the unit controls the pre-sending of encrypted creative data packets of candidate advertising objects to the local storage area of ​​the IoT terminal. When the real-time bid is higher than the dynamic admission threshold, the unit generates a display authorization key for the target advertising object and sends this key as a distribution instruction to the IoT terminal, triggering the immediate decryption and rendering of the corresponding encrypted creative data packets in the local storage area. When the real-time bid is higher than the dynamic admission threshold, the bidding processing unit generates an allow display instruction; when the real-time bid is lower than the dynamic admission threshold, the unit generates an allow display instruction. When the entry price is higher than the preset base price, the unit generates a downgrade display instruction, controlling the IoT terminal to display the advertising content in a non-intrusive notification bar format. This non-intrusive notification bar format does not obscure the preset key operation area of ​​the IoT terminal's current functional interface. When the real-time bid is lower than the base price, the unit generates a reject display instruction. The reject display instruction drives the GPU command queue overwrite logic, fills the pre-allocated advertising rendering time slot with a hexadecimal 0x90 null opcode, synchronously lowers the display controller overlay enable pin level, physically cuts off the direct memory access channel from the texture unit to the frame buffer, and forces the rendering pipeline into a computation graph collapse state to avoid computing power leakage and heat accumulation caused by background silent rendering.

[0023] Example 1: In a typical vehicle-to-everything (V2X) application scenario, the terminal device is under high load for assisted driving. The currently active window is a composite display interface of the vehicle's reversing image and radar ranging. Under this condition, the user's operation behavior exhibits pulse-like interaction characteristics, including frequent parameter fine-tuning clicks and silent observation confirmation actions. This results in the input command stream appearing as an alternating sequence of high-frequency triggers and zero inputs on the timeline. Existing advertising distribution logic that relies solely on instantaneous input status to determine idle resources faces the technical risk of misjudging the device as idle due to the detection of zero input, thus issuing full-screen advertisements that obscure the reversing image. The status monitoring unit collects the terminal's operating parameters in real time. The system identifies that the current active window is located in a deep functional area within the graphical user interface tree structure and measures the interface depth value. for For the specific function execution page, at the moment the user performs a fine-tuning operation, the status monitoring unit counts the interruption signals within the sliding time window and measures the interaction frequency value. It is at a high level; measured at the instant the user stops operating and begins observation. The load calculation unit is reset to zero according to the formula. Perform the calculation, where, As a gain coefficient, amplification For load counter The contribution weight is adjusted so that the load increment generated by a single click operation is higher than that generated by an operation on a shallow menu page.

[0024] When users enter a silent observation period, the real-time interaction frequency value The load is reduced to zero, and the load calculation unit is based on a preset attenuation rate. right Periodic deductions are implemented due to The calibration value makes The zeroing period is longer than the user's usual observation and confirmation period, therefore, within a short window after the user stops operating, Maintaining the value in the high-order bits (non-zero), the threshold generation unit reads the high-order bits. Numerical values, using formulas The system calculates the current dynamic access threshold price. This calculation mechanism quantifies the user's cognitive occupancy during physical operation intervals into an extremely high advertising bidding threshold. When the bidding processing unit receives an external advertising request, it compares the real-time bid with this dynamic access threshold price. Since the real-time bid is lower than the dynamic access threshold price, the system generates a rejection instruction. Without relying on image recognition or complex user behavior analysis algorithms, the system effectively protects the silent attention period in the critical task flow through only low-computing-power numerical calculations. The status monitoring unit extracts the service class identifier of the current foreground process as com.auto.reverse. The load calculation unit compares this identifier with a pre-set list of highly sensitive services and determines whether it matches the feature string of the reversing camera service. In response to this matching result, the system triggers a forced locking logic, ignoring the current... and The calculation results will Set the bit to the maximum value allowed by the system and pause the timed deduction logic. At this time, the dynamic admission floor price is locked at the theoretical maximum value, ensuring that even if... Even if the image quality naturally degrades to zero due to prolonged inactivity, the reversing camera interface, which is related to driving safety, still retains its ability to resist advertisements.

[0025] Example 2: This example verifies the technical effectiveness of the leaky bucket integral-based dynamic pricing mechanism in combating transient jitter interference and compares it with conventional instantaneous state monitoring technology. The experiment is conducted on a controlled simulation test bench, which is equipped with a simulated IoT terminal running a state monitoring daemon and a server instance deploying threshold generation logic. The simulated terminal generates a predefined sequence of input command streams through a script to simulate the interaction behavior of real users. This experiment sets up two control groups: control group (conventional technology): using traditional instantaneous state pricing logic, i.e., when an interaction event is detected ( When there are no interactive events, increase the minimum bid price for ad access; when there are no interactive events ( When the load is reset to the base value, the experimental group (the scheme of this invention) adopts load calculation based on leaky bucket integration and asymmetric time delay filtering logic. The key parameters are configured as follows: base single operation load equivalent. The interface depth value is fixed. decay rate Unit / second, status hold timer window set to Second.

[0026] To simulate user hesitation and pauses during operation, a total duration of [duration to be specified] is designed. The test input sequence for seconds: Second (high-frequency operation period): per A click event is generated every second to simulate continuous user actions. Seconds (brief or paused): No input event occurs, simulating user thinking or hesitation. Seconds (reoperation period): per A click event is generated every second to simulate a user's recovery action. Seconds (end period): No input event, start simulation test, and record the dynamic admission floor price of the two systems under the above input sequence. The change curve sets the base price. Unit, interference penalty coefficient The experimental data are recorded in the table below (sampling interval). Second).

[0027] Table 1: Experimental Data Recording Table

[0028] exist Instant The brief pause of a second, due to the disappearance of the input event, caused the floor price of the control group to instantly drop from... Fall to (Base price), when users only briefly consider... Within the one-second window, the system mistakenly judges the space as idle, releasing an opportunity for advertisers to infiltrate with low prices. If a bid is placed at this time... An ad request will be allowed to be displayed, thus interrupting the user's workflow. (Starting in seconds); At the second mark, although the physical input stopped, the load counter of the experimental group... Only deduct The value remains at At a high level, combined with the state-maintaining timer logic, the experimental group's floor price remained consistently high throughout the entire suspension period. The above high positions, in the same Within a two-second interval, the minimum admission price for the experimental group is [missing value] of the control group. More than twice.

[0029] Example 3: This example combines Figures 1 to 3 This describes an integrated marketing advertising delivery and management system based on the Internet of Things (IoT), such as... Figure 1 As shown, the core processing flow begins with the input stream from the IoT terminal, which includes user interaction commands and operational behaviors. The status monitoring unit is responsible for receiving this input stream and acquiring the interaction frequency value and interface depth value in real time as key operating parameters. The load calculation unit generates a weighted load increment based on the above parameters and performs a periodic deduction operation to output the current count value. The threshold generation unit receives the current count value and generates a dynamic admission floor price through mapping logic to quantify the cognitive load barrier. The bidding processing unit receives the advertising placement request containing the real-time bid, compares the real-time bid with the dynamic admission floor price, and makes a decision. If the real-time bid is higher than the floor price, the system generates a display instruction to allow the advertisement to be displayed. If the real-time bid is lower than the floor price, the system generates a rejection or downgrade instruction to terminate the current high-level display process.

[0030] like Figure 2 As shown, the horizontal axis represents time in seconds, and the vertical axis represents the minimum entry price. The dashed line represents the control group (traditional solution), and the solid line represents the experimental group (this invention). During the continuous click phase from 0 to 3.0 seconds, both curves show an upward trend. During the input stop interval from 3.5 to 5.0 seconds, the control group curve rapidly drops to near the baseline of 0, while the experimental group curve, relying on the inertial decay characteristic of the load value, remains consistently above the high threshold range of 25, continuing to climb after the input event is detected again at 5.5 seconds. Figure 3 As shown, IoT terminal A, taking an in-vehicle central control unit as an example, and IoT terminal B, taking an industrial tablet as an example, are both edge devices. Both integrate a status monitoring unit and a load computing unit, and are configured with a local storage area for storing pre-loaded encrypted materials. These terminals establish a two-way data channel with the cloud server cluster through the Internet or a dedicated network, mainly sending status update requests. The cloud server cluster, as the decision-making core, is equipped with a threshold generation unit, a bidding processing unit, and a strategy database that stores historical logs and rules. This cluster is responsible for responding to terminal requests to dynamically adjust the admission floor price and directly connecting to a third-party advertising exchange platform to process advertising placement requests and execute subsequent key distribution operations.

[0031] Example 4: This example focuses on eliminating ambiguity in implementation details in the following key areas: In the load calculation logic, the weighting mechanism of interface depth value and interaction frequency value lacks specific quantitative model support; in the threshold generation logic, the setting of interference penalty coefficient lacks a theoretical basis; and in the idle compensation mechanism, the determination of time value drift coefficient lacks engineering optimization procedures. Addressing the black box of the quantitative model in the load calculation logic, this example constructs a weighting coefficient calibration procedure based on user cognitive load theory and defines a set of standardized test tasks, covering browsing from shallow menus (depth value...). To deep parameter configuration (depth value) The typical operational path involves inviting a group of users to perform the above tasks in a controlled experimental environment, while simultaneously monitoring their physiological indicators such as pupil diameter change rate and heart rate variability as objective representations of cognitive load. Through regression analysis of the experimental data, an interface depth value is established. The nonlinear mapping relationship between cognitive load increment and the load calculation formula is then determined. Depth weighting coefficients in Experimental results show that when Set as When the calculated load increment has the best correlation with the actual cognitive load, it provides a solid physiological and statistical basis for the selection of this parameter.

[0032] This embodiment further elaborates on the interference penalty coefficient, addressing the black box nature of parameter setting in the threshold generation logic. The adaptive optimization logic, based on a dual-objective optimization model of ad click-through rate and task interruption rate, will be used in the initial deployment phase of the system. The value is set to a wider search range, such as In actual operation, the system continuously collects user feedback data after ad placement: on the one hand, it calculates the click-through rate (CTR) to represent commercial value; on the other hand, it calculates user feedback within a preset time after ad display. The frequency with which ads are manually closed or the current task is exited per second is used as a proxy indicator of task interruption rate. The system uses a gradient descent algorithm to dynamically adjust... The goal is to find a balance point where the task interruption rate is controlled within a preset tolerance threshold, such as... Under the following premise, to maximize ad click-through rate, after multiple rounds of iteration, the following was determined. For the optimal parameter configuration in the current application scenario, and to address the black box of engineering optimization in the idle compensation mechanism, this embodiment further describes the time value drift coefficient. The offline calibration process quantifies the correlation between device idle time and the decay of user attention value. In the experiment, a group of devices with different idle times (from...) were selected. minutes to A standardized set of test advertisements was displayed on terminal devices (within 24 hours) and user responses, such as whether they clicked or woke up the device, were recorded. As idle time increased, the user response rate to the advertisements decreased exponentially. Based on this pattern, a compensation formula was obtained using the least squares method. coefficients in The calibration results show that when Values When the unit is per minute, the generated dynamic admission floor price can effectively filter out responses with a response rate lower than a preset threshold, such as... This reduces low-value ad requests, thereby protecting network bandwidth while ensuring the accuracy of ad delivery.

[0033] Example 5: In commercial complexes with a large number of self-service terminals such as vending machines and public information kiosks, the system faces the problem of ineffective traffic surges during long-tail idle periods. These devices are often in a long-term unattended state late at night or during off-peak hours. At this time, the interaction frequency value measured by the status monitoring unit is... The interface depth value remains zero. Staying on the standby page ( Based on the basic load calculation logic, the load counter It will decay to zero, resulting in a dynamic access floor price. Falling back to the lowest base price If external advertising platforms, based on a low-price-first-served strategy, push a large number of low-quality and untargeted filler ads, not only will they fail to attract effective user attention, but they will also needlessly consume valuable IoT bandwidth resources and even increase hardware wear and tear on device screens. To address these challenges, this embodiment enables an idle value attenuation compensation mechanism, and the threshold generation unit continuously monitors the load counter. The real-time value, when If it remains below the preset low threshold, such as When this occurs, the system starts an idle timer to accumulate and record the duration. ,when Exceeding the preset baseline time If set as Minutes later, the system determines that the device has entered the invalid attention period. At this time, the threshold generation unit uses the formula... Calculate threshold compensation amount ,in, This is the time value drift factor, and its value is set to a positive value, such as... Unit: per minute.

[0034] As equipment idle time continues to increase ( (Continuously increasing) The system exhibits a linear growth trend, and this compensation amount is added to the base price, resulting in a final dynamic entry price adjustment. This process causes advertising access prices to rise rather than fall during periods of equipment downtime, when the downtime reaches [a certain value]. Hourly, accumulated This can cause the entry price to reach several times the base price. The bidding processing unit filters ad requests based on this high price, with the vast majority of low-bid fill ads being directly rejected because they cannot meet the price requirement. Only a very small number of high-value ads with extremely high bids, such as visually impactful brand campaigns designed to engage users, are granted display rights. This mechanism, through a reverse pricing strategy where idle devices become more expensive, effectively cleans up invalid traffic on long-term inactive devices, causing ad placement to converge towards high-value and highly engaging content. This protects network bandwidth resources while maintaining the overall ecological value of the system. Once a user physically operates the device again (…), ), load counter Upon detecting this sudden jump, the system will pause the idle timer. Reset to zero, threshold compensation amount Once the system is reset to zero, it immediately reverts to the load-based dynamic pricing model.

[0035] Example 6: This example constructs a pre-deployment calibration and adaptive calibration procedure for thousands of heterogeneous IoT terminals in an industrial manufacturing base. For devices distributed across different workshops and undertaking tasks of varying criticality, such as security monitoring, AGV scheduling, and production line control, customized benchmark parameters for load calculation and threshold generation are provided to ensure stable execution of physical tasks even after integration with the advertising distribution system. The system performs benchmark load calibration, and when the device is idle with no user interaction and stable background tasks, the status monitoring unit continuously collects data for a preset duration, such as... Interaction frequency value within minutes With interface depth value Statistical analysis of load counters during this period The numerical distribution of the expression is calculated, and its mean is determined. with standard deviation Based on these statistical results, the idle threshold of the device is dynamically set to... This is to adapt to the background noise levels of different devices and avoid misjudgments caused by individual differences in equipment.

[0036] Secondly, considering the varying sensitivities of different devices to task interruptions, the system adaptively optimizes the interference penalty coefficient. In the initial stage, the interference penalty coefficient is... Set to a conservative initial value, the system delivers a set of test ads with different bid tiers to the device (using a non-intrusive logging mode, without actual display). Utilizing the device's historical operation logs, it simulates user interaction behavior under different load conditions. If during simulated high-load periods ( If the bid for a virtual ad exceeds the set threshold and falls below the dynamic minimum bid price, it is considered a potential interference event, and the coefficient is adjusted accordingly. If the ad fails to display during periods of low load, adjust the settings accordingly. The simulated annealing algorithm is used for multiple iterations until a method is found that reduces the potential interference rate below a preset safety limit. And the optimal method to maximize ad fill rate The system then performs time-domain calibration of the idle compensation parameters, monitors the equipment's active time distribution over the past week, identifies typical long-tail idle periods such as nighttime shutdowns, and dynamically adjusts the base time in the idle compensation formula based on the duration distribution characteristics of these idle periods. With time value drift coefficient ,Will Set as the average operating interval of the device times, will The dynamic entry threshold price is set to smoothly rise to the base price before the end of a typical idle period. More than twice.

[0037] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An integrated marketing advertising publishing and management system based on the Internet of Things, characterized in that, The system includes: The status monitoring unit is used to monitor the input command flow of IoT terminals in real time, obtain the frequency of interaction commands per unit time as the interaction frequency value, and obtain the layer depth of the current display interface as the interface depth value. The load calculation unit is used to maintain a load counter in memory. When an input instruction stream is received, a load increment is generated based on the weighted product of the interaction frequency value and the interface depth value. The load increment is accumulated to the current value of the load counter. At the same time, the current value of the load counter is periodically deducted at a preset decay rate using timing logic until the value reaches zero. The threshold generation unit is used to read the current count value in the load counter; and convert the current count value into a dynamic entry price for advertising bidding using a preset mapping rule. The bidding processing unit is used to receive advertising requests and their real-time bids for IoT terminals; compare the real-time bids with the dynamic admission threshold; and generate a rejection instruction or a downgrade instruction if the real-time bid is lower than the dynamic admission threshold.

2. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, When the load calculation unit performs the deduction operation, if the current value of the load counter is greater than zero and less than the decay step size of a single deduction cycle, the load counter is directly reset to zero. The status monitoring unit also includes a service identifier extraction module, which is used to obtain the service class identifier of the currently running process. The load calculation unit also includes a locking logic module, which ignores the current state of the input instruction stream when the service class identifier matches the preset list of highly sensitive services, forcibly locks the value of the load counter to the preset maximum threshold, and suspends the deduction operation of the timed logic.

3. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The threshold generation unit includes a trend response module for monitoring the rate of change of values ​​in the load counter; When the rate of change indicates that the current count value is on an upward trend, the current count value is directly output for calculating the dynamic entry floor price. When the rate of change indicates that the current count value is in a downward trend, the state hold timer is started. Within the preset window period of the state hold timer, the peak value before the decline is maintained until the preset window period expires and no reverse upward trend is detected.

4. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The bidding processing unit is also used to execute asynchronous distribution logic that separates creative materials from licenses. The bidding processing unit includes: a preloading module, used to send encrypted creative material data packets of candidate advertising objects to the local storage area of ​​the IoT terminal while the dynamic admission threshold is maintained below a preset idle threshold; a key distribution module, used to generate a display authorization key for the target advertising object when the real-time bid is higher than the dynamic admission threshold, and send the display authorization key as a distribution instruction to the IoT terminal; the display authorization key is used to trigger the IoT terminal to decrypt and render the corresponding encrypted creative material data packets in the local storage area.

5. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The threshold generation unit also includes an idle compensation module, which is used to continuously monitor the duration for which the current count value is lower than the preset low threshold. Calculate the threshold compensation amount using the following formula. : ,in, The preset compensation coefficient, This is the current value of the duration. The preset base time, It is a unit step function; the threshold generation unit adds the threshold compensation amount to the dynamic access floor price. After the duration exceeds the benchmark time, the dynamic access floor price shows a monotonically increasing trend with the increase of the duration.

6. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The bidding processing unit includes a display mode adjustment module, which stores a table showing the correspondence between different ad display modes and price ranges. When the real-time bid is lower than the dynamic entry threshold but higher than the preset base price, the display mode adjustment module retrieves the corresponding table and generates a downgraded display instruction that instructs the IoT terminal to display the ad content in a non-intrusive notification bar format. The non-intrusive notification bar format is limited to a preset key operation area that does not obscure the current functional interface of the IoT terminal.

7. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The attenuation rate in the load calculation unit is configured as a variable value; The load calculation unit includes an adaptive rate adjustment module, which is used to count the frequency of the load counter value reaching the maximum threshold within a preset historical period; when the frequency exceeds the preset frequency threshold, the decay rate value is reduced to extend the retention time of the load counter value.

8. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, When acquiring the interface depth value, the status monitoring unit performs the following operations: traverses the interface tree structure of the currently active window of the IoT terminal; identifies the hierarchical index of the currently focused element in the interface tree structure; The values ​​of the hierarchical indexes are normalized to integers within a preset range and used as the interface depth values. The load calculation unit uses the interface depth value as a weighting coefficient to generate the load increment after weighting the interaction frequency value.

9. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The bidding processing unit is also used to record the source identifier of the rejected ad delivery request after generating the rejection display instruction; if the same source identifier triggers the rejection display instruction more than the preset limit within a preset time window, a blocking instruction is generated for the source identifier, and all ad delivery requests from the source identifier are directly discarded without comparison during the subsequent preset cool-down period.

10. The integrated marketing advertising publishing and management system based on the Internet of Things according to claim 1, characterized in that, The system is deployed in a distributed architecture that includes cloud servers and edge terminals. The status monitoring unit and load calculation unit reside on the edge terminal and are used to maintain the load counter locally and calculate the current count value. The edge terminal only sends a status update request to the cloud server when the current count value changes beyond a preset threshold. The threshold generation unit and bidding processing unit reside on the cloud server and are used to update the dynamic admission price in response to the status update request and process advertising placement requests from third-party advertising exchange platforms.

Images