Smart farm-oriented agricultural machine operation data secure transmission method

Abstract

The present application relates to the field of data processing, especially to a kind of intelligent farm-oriented agricultural machinery operation data security transmission method, the method comprises: determining high load channel;Determine that data transmission abnormal event occurs;Determine abnormal agricultural machinery;Determine composite abnormal agricultural machinery;Determine agricultural machinery adjustment instruction and adjustment threshold, the present application filters high load channel by load crowded index;Determine transmission abnormal event based on collision rate and retransmission rate;Subsequently, by sending success rate and signal receiving power, construct agricultural machinery operation feature vector and identify outlier agricultural machinery;Again, accurately locate composite abnormal agricultural machinery that is suppressed by long distance and frequently switched;Finally, according to the spatial position of composite abnormal agricultural machinery and cause-effect chain relationship, generate adjustment instruction, and correct threshold according to the change of transmission abnormal index, effectively solve the problem of abnormal identification lag and low data transmission accuracy due to the inability to adapt to complex and changeable operation conditions.

Description

Technical Field

[0001] This invention relates to the field of data processing, and in particular to a method for secure transmission of agricultural machinery operation data for smart farms. Background Technology

[0002] With the rapid development of smart agriculture, large-scale plain farms commonly employ intelligent equipment such as unmanned agricultural machinery and drones for precision operations. During these operations, real-time transmission of multi-dimensional data, including location, operating conditions, yield, and video, is required to the management center for remote monitoring, scheduling optimization, and precise decision-making. However, large farms cover vast areas (typically thousands or even tens of thousands of acres) and have numerous agricultural machines (tens to hundreds). Furthermore, the constant switching of communication channels by each machine during operations, driven by movement and coverage requirements, results in highly dynamic, high-density, and high-concurrency communication networks. Ensuring the secure and reliable transmission of agricultural machinery operation data in this complex and ever-changing channel environment has become a core technical challenge in the construction of smart farms.

[0003] Currently, data transmission for agricultural machinery operations mainly employs technologies such as cellular mobile communication, Wi-Fi ad hoc networks, or dedicated short-range communication, utilizing access selection strategies based on signal strength or channel occupancy, collision avoidance mechanisms based on random backoff, and data encryption transmission schemes based on authentication. Some research has also introduced edge computing nodes for data aggregation and forwarding to alleviate bandwidth pressure on the backhaul link. Regarding channel resource management, existing solutions determine channel load status by monitoring channel occupancy or the number of agricultural machines connected, and mitigate congestion by adjusting transmission power, switching channels, or reducing data transmission rates.

[0004] However, existing methods have the following shortcomings when dealing with the parallel operation of multiple agricultural machines in large-scale plain farms: they cannot comprehensively reflect the actual congestion caused by the combined effect of quantity and load when multiple agricultural machines share the same channel, and are prone to being judged as congested before the channel is saturated or being severely congested but not identified; they are difficult to distinguish between short-term fluctuations caused by instantaneous interference and continuous performance degradation caused by long-term overload, which may trigger unnecessary adjustment measures or delay the handling of real anomalies; they ignore the inherent correlation between the indicators of various agricultural machines in the same channel, which leads to some agricultural machines in the normal fluctuation range being misjudged as abnormal, while agricultural machines with seriously degraded performance are missed because a single indicator does not exceed the threshold; for complex scenarios with multiple types of transmission anomalies at the same time, existing methods usually only take uniform adjustment measures for a certain type of explicit anomaly, without distinguishing the coupling relationship between different anomalies, and the adjustment effect is limited. Summary of the Invention

[0005] To address this, the present invention provides a method for secure transmission of agricultural machinery operation data for smart farms. This method overcomes the problems of delayed anomaly identification and low data transmission accuracy caused by the inability to adapt to complex and ever-changing operating conditions in existing technologies through multi-dimensional data analysis and dynamic adjustment mechanisms.

[0006] To achieve the above objectives, the present invention provides a method for secure transmission of agricultural machinery operation data for smart farms, comprising: In a large-scale plain farm, several agricultural machines under test operate in parallel and switch communication channels in real time. The number of agricultural machines and the instantaneous occupancy rate of each channel under test are collected in real time. Based on the comparison between the load congestion index and the preset congestion index threshold, several high-load channels are identified. The load congestion index is determined based on the number of agricultural machines and the instantaneous occupancy rate. The collision rate and retransmission rate of each high-load channel are collected in real time. The transmission anomaly index is compared with the preset anomaly index threshold to determine that a data transmission anomaly event has occurred in the high-load channel. The transmission anomaly index is determined based on the collision rate and retransmission rate. Based on the data transmission anomaly event, the transmission success rate and signal reception power of each of the agricultural machines under test in the high-load channel are collected. Several abnormal agricultural machines are identified according to the outlier of the agricultural machine operation feature vector, wherein the agricultural machine operation feature vector is constructed based on the transmission success rate and signal reception power. The instantaneous speed and switching frequency of the abnormal agricultural machinery are collected, and the composite abnormal agricultural machinery is determined based on the near-far effect characteristic value and the moving switching characteristic value. The near-far effect characteristic value is determined based on the signal receiving power and spatial position relationship, and the moving switching characteristic value is determined based on the instantaneous speed and switching frequency. The adjustment instructions for the agricultural machinery are determined based on the spatial location and causal chain location of the composite abnormal agricultural machinery. Based on the agricultural machinery adjustment command, the preset congestion index threshold is adjusted according to the change characteristics of the transmission anomaly index within a preset observation period.

[0007] Furthermore, when the load congestion index is greater than the preset congestion index threshold, the channel under test is determined to be a high-load channel, thus obtaining several high-load channels.

[0008] Furthermore, the process of determining the load congestion index based on the number of agricultural machines and instantaneous occupancy rate includes: The load congestion index is obtained by nonlinear coupling calculation based on the number of agricultural machines, the preset maximum number of agricultural machines to be supported, the instantaneous occupancy rate, and the preset channel saturation occupancy rate threshold.

[0009] Furthermore, when the transmission anomaly index is greater than the preset anomaly index threshold, it is determined that a data transmission anomaly event has occurred on the high-load channel.

[0010] Furthermore, the process of determining the transmission anomaly index based on the collision rate and retransmission rate includes: The collision fluctuation value is calculated based on the collision rate within a preset time period; Calculate the retransmission rate fluctuation value based on the retransmission rate within the preset time period; The transmission anomaly index is calculated by weighting the collision fluctuation value and the retransmission rate fluctuation value.

[0011] Furthermore, several abnormal agricultural machines are identified based on the outlier of the agricultural machinery operation feature vector. The process of constructing the agricultural machinery operation feature vector based on the transmission success rate and signal reception power includes: The agricultural machinery operation feature vector is constructed based on the transmission success rate and the signal reception power. The outlier degree is obtained by calculating the local outlier factor based on the agricultural machinery operation feature vector of each agricultural machinery under test and all agricultural machinery under test in the high-load channel; When the outlier degree is greater than the preset outlier threshold, the agricultural machine under test is determined to be the abnormal agricultural machine.

[0012] Furthermore, when the near-far effect feature value is greater than a preset first feature threshold and the movement switching feature value is greater than a preset second feature threshold, the abnormal type of the abnormal agricultural machine is determined to be a near-far movement composite abnormality, and the abnormal agricultural machine is determined to be a composite abnormal agricultural machine.

[0013] Furthermore, the near-far effect characteristic value is determined based on the product of the difference in signal received power between the abnormal agricultural machine and the agricultural machine with the strongest signal in the channel and the spatial collinearity of the two relative to the gateway. The spatial collinearity is calculated based on the absolute value of the cosine of the angle between the abnormal agricultural machine and the agricultural machine with the strongest signal and the gateway. The mobile handover characteristic value is determined based on the product of the instantaneous speed of the abnormal agricultural machine and the handover frequency.

[0014] Furthermore, the process of determining the agricultural machinery adjustment command based on the spatial location and causal chain location of the composite abnormal agricultural machinery includes: Based on the spatial location of the composite abnormal agricultural machine, a temporary communication channel with a load congestion index less than a preset congestion index threshold is determined, and the composite abnormal agricultural machine is moved to the temporary communication channel to obtain an agricultural machine adjustment command. Adjusting power and switching parameters based on the causal chain position of the complex abnormal agricultural machinery yields the agricultural machinery adjustment command.

[0015] Furthermore, the process of adjusting the preset congestion index threshold based on the changing characteristics of the transmission anomaly index within a preset observation period includes: The rate of change of the index is calculated based on all the transmission anomaly indices within the preset observation period; The preset congestion index threshold is adjusted based on the comparison between the rate of change of the index and the preset change threshold.

[0016] Compared with existing technologies, the beneficial effects of this invention lie in establishing a causal chain between secure transmission of agricultural machinery operation data and anomalies in the physical layer of wireless communication through a hierarchical diagnostic logic that ranges from overall channel load to individual agricultural machinery behavior. First, a load congestion index is constructed using the number of agricultural machines and instantaneous occupancy rate to filter out high-load channels, as channel congestion is the fundamental cause of collisions and retransmissions. Then, transmission anomaly events are determined based on collision rate and retransmission rate, transforming macroscopic channel deterioration into quantifiable anomaly signals. Subsequently, within the anomaly channel, an agricultural machinery operation feature vector is constructed using transmission success rate and signal reception power to identify outlier agricultural machines, because low success rate and weak reception power directly correlate with... This approach reflects local interference or black hole coverage; by combining instantaneous velocity and switching frequency, it characterizes near-far effect and mobile switching oscillation, thereby accurately locating agricultural machinery with complex anomalies that are both suppressed from a distance and frequently switched, revealing coupled faults that cannot be explained by a single indicator; finally, it generates adjustment instructions based on the spatial location and causal chain relationship of the complex anomaly agricultural machinery, and reversely corrects the congestion index threshold based on the change of the transmission anomaly index within the observation period, significantly improving the reliability, security, and adaptability of agricultural machinery data transmission in the complex dynamic environment of large plain farms, effectively solving the problems of delayed anomaly identification and low data transmission accuracy caused by the inability to adapt to complex and changing operating conditions.

[0017] Furthermore, if the occupancy rate is close to saturation even when the number of agricultural machines has not reached its limit, it indicates high-volume operations or long data packet transmissions, which can also lead to a high risk of collisions. Conversely, if the occupancy rate is low but the number of agricultural machines is extremely large, dense random access can also cause congestion. Therefore, by introducing a preset maximum number of agricultural machines and a preset channel saturation occupancy rate threshold to calculate the load congestion index, the positive correlation between the collision probability and the two factors under multi-terminal shared channel conditions is simulated. When the proportion of agricultural machines exceeds the capacity threshold or the instantaneous occupancy rate exceeds the saturation threshold, the index exhibits superlinear growth, thus more sensitively reflecting the critical state of high load. When both are below the threshold, the index remains at a low level, avoiding misjudgments and preventing missed detections or false alarms caused by looking only at the number or occupancy rate, thereby improving the robustness and accuracy of channel assessment in complex smart farm operation scenarios.

[0018] Furthermore, the intensity of competition for channel resources is jointly determined by the number of agricultural machines and the instantaneous occupancy rate. These two factors are not linearly superimposed but rather coupled. As the number of agricultural machines increases, the rising occupancy rate significantly amplifies the probability of collisions. Therefore, a load congestion index determined based on both factors is used as a single quantitative indicator and compared with a preset threshold. Essentially, this constructs a competition-risk discrimination threshold. Only when the load congestion index exceeds the threshold is the channel classified as a high-load channel. Channels below the threshold, even with high local occupancy, have a relatively small number of agricultural machines, and their actual collision risk remains within a controllable range, eliminating the need to trigger subsequent complex anomaly detection procedures. Only when both a large number of machines and a high occupancy rate are simultaneously met does a channel truly enter the high-collision-risk zone. This avoids unnecessary collision rate and retransmission rate collection and calculation on low-risk channels, significantly reducing the system's real-time processing overhead. Furthermore, it provides a precise input range for judging subsequent data transmission anomalies, allowing diagnostic resources to focus on high-load channels where anomalies are truly likely, thereby improving the real-time performance and engineering feasibility of the entire secure transmission method.

[0019] Furthermore, the collision fluctuation value reflects the irregularity of competition among multiple agricultural machines. For example, sudden large-volume operations cause the collision rate to fluctuate, while the retransmission rate fluctuation value characterizes the retransmission delay jitter caused by signal attenuation or hidden terminals at the link layer. The two are physically complementary. Simple collision fluctuations may be mitigated by rapid backoff, while simple retransmission fluctuations may originate from coverage changes caused by the movement of individual agricultural machines. However, when both fluctuate significantly, the weighted transmission anomaly index will increase significantly, accurately identifying that the channel has entered a coupled deterioration stage that cannot be resolved by competition backoff or repaired by single-point retransmission. Compared to directly using the instantaneous average of the collision rate or retransmission rate, this weighted method based on fluctuation values ​​effectively suppresses false alarms and is more sensitive to continuously deteriorating channel responses. This provides statistically stable and physically causally clear feature inputs for subsequent judgment of abnormal data transmission events, significantly improving the reliability and engineering practicality of anomaly detection in the complex dynamic environment of smart farms.

[0020] Furthermore, by ensuring that the transmission anomaly index does not exceed the threshold when the collision rate occasionally increases but the retransmission rate remains low, overreaction to transient interference is avoided. Conversely, if the collision rate remains high and the retransmission rate rises simultaneously, the index quickly breaks through the threshold, accurately identifying that the channel has entered an abnormal state that cannot be self-healed. This not only filters out normal short-term jitter in the channel but also ensures that only persistent anomalies that truly threaten the secure transmission of data will activate the subsequent agricultural machinery-level positioning and adjustment process. Thus, in the highly dynamic and parallel operating environment of smart farms, the optimal balance between the sensitivity and robustness of anomaly detection is achieved.

[0021] Furthermore, by employing a local outlier factor to calculate outlier degree, essentially comparing the density deviation of each agricultural machine within its local neighborhood, it can automatically adapt to the distribution differences of different density regions in a multi-dimensional feature space. It does not require a preset global threshold. While the product or combination of transmission success rate and signal reception power is not linearly separable, the local outlier factor, through reachability distance and local reachability density, can sensitively capture individuals whose surrounding neighbors are normal but who are clearly different. Even if the absolute values ​​of these abnormal agricultural machines do not exceed conventional thresholds, for example, a slightly lower success rate in an overall low-power region may make them outliers. This effectively overcomes the insufficient adaptability of fixed thresholds under dynamic channel conditions, accurately screening out candidate abnormal agricultural machines that require further analysis of near-far effects and mobile handover, significantly improving the accuracy and robustness of faulty agricultural machine location in the complex electromagnetic environment of large-scale smart farms.

[0022] Furthermore, by recognizing that instantaneous speed and switching frequency are physically positively correlated but not perfectly linear, the product index can distinguish between low-speed high-frequency and high-speed high-frequency. Meanwhile, the spatial collinearity in the near-far effect feature introduces an angle-sensitive factor, ensuring that even with large power differences, spatial orthogonality will not lead to misjudgment as a composite anomaly. This approach captures the coupling effect of mobility in the power domain, spatial domain, and temporal domain in a concise mathematical form, providing a physically meaningful and sensitively controllable quantitative feature for subsequent dual-threshold determination. This significantly improves the accuracy and interpretability of identifying composite anomaly agricultural machinery in the complex dynamic environment of smart farms.

[0023] Furthermore, the near-far effect eigenvalues ​​characterize the suppression intensity of the agricultural machinery in the power domain, while the movement switching eigenvalues ​​characterize the turbulence intensity of the agricultural machinery in the connectivity domain. Only when both exceed their respective preset thresholds is the agricultural machinery identified as having a composite abnormality. This avoids misjudging agricultural machinery with only a single abnormality as having a composite fault requiring special handling, thereby accurately identifying those agricultural machinery that are both suppressed and frequently fluctuating, reducing unnecessary intervention operations, and enabling subsequent agricultural machinery adjustment commands to perform migration channels and parameter optimization for real coupled fault scenarios. This significantly improves the diagnostic accuracy and resource utilization efficiency of the data transmission security method in the high-dynamic operation environment of smart farms.

[0024] Furthermore, spatial location is directly linked to the feasibility of temporary channels and the continuity of communication coverage; causal chain location is inferred through time-series analysis of historical power differences, collinearity, speed and switching frequency to determine the role of composite anomalies in the causal network, thereby achieving dual intervention of migration avoidance and parameter fine-tuning. This not only immediately relieves the congestion pressure of the current channel, but also prevents the recurrence of similar anomalies in temporary channels, ultimately enabling the data transmission of the entire agricultural machinery fleet to gradually return to a stable state, significantly improving the control effectiveness and system convergence speed in the complex dynamic environment of smart farms.

[0025] Furthermore, by introducing the rate of change of the transmission anomaly index as a feedback signal, the transmission anomaly index of all data transmission anomaly events that have occurred is continuously calculated within a preset observation period, and its rate of change is extracted. This rate of change directly reflects whether the collision and retransmission fluctuations in the high-load channel tend to intensify, alleviate, or remain stable. When the rate of change of the index exceeds the preset change threshold, it indicates that the degree of anomaly is rapidly deteriorating. Even if the current congestion index has not yet generally exceeded the standard, the congestion index threshold needs to be lowered so that more channels enter the high-load judgment range in advance and preventive intervention is implemented. This makes the congestion index threshold no longer a static empirical value, but a dynamic and continuous optimization with the actual operation of the farm. It realizes intelligent iteration throughout the entire process and significantly improves the robustness, self-consistency, and scenario generalization ability of the method in long-term operation. Attached Figure Description

[0026] Figure 1 This is a flowchart of the secure data transmission method for agricultural machinery operations in smart farms, as described in this embodiment. Figure 2 This is the logic diagram for determining the high-load channel in this embodiment; Figure 3 This is a logic diagram for determining the occurrence of a data transmission anomaly event in this embodiment; Figure 4 This is a logic diagram for determining abnormal agricultural machinery in this embodiment. Detailed Implementation

[0027] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0028] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0029] Please see Figure 1 The diagram shown is a flowchart of a method for secure data transmission of agricultural machinery operations in smart farms, as described in this embodiment. This embodiment provides a method for secure data transmission of agricultural machinery operations in smart farms, including: In a large-scale plain farm, several agricultural machines under test operate in parallel and switch communication channels in real time. The number of agricultural machines and the instantaneous occupancy rate of each channel under test are collected in real time. Based on the comparison between the load congestion index and the preset congestion index threshold, several high-load channels are identified. The load congestion index is determined based on the number of agricultural machines and the instantaneous occupancy rate. The collision rate and retransmission rate of each high-load channel are collected in real time. The transmission anomaly index is compared with the preset anomaly index threshold to determine that a data transmission anomaly event has occurred in the high-load channel. The transmission anomaly index is determined based on the collision rate and retransmission rate. Based on the data transmission anomaly event, the transmission success rate and signal reception power of each of the agricultural machines under test in the high-load channel are collected. Several abnormal agricultural machines are identified according to the outlier of the agricultural machine operation feature vector, wherein the agricultural machine operation feature vector is constructed based on the transmission success rate and signal reception power. The instantaneous speed and switching frequency of the abnormal agricultural machinery are collected, and the composite abnormal agricultural machinery is determined based on the near-far effect characteristic value and the moving switching characteristic value. The near-far effect characteristic value is determined based on the signal receiving power and spatial position relationship, and the moving switching characteristic value is determined based on the instantaneous speed and switching frequency. The adjustment instructions for the agricultural machinery are determined based on the spatial location and causal chain location of the composite abnormal agricultural machinery. Based on the agricultural machinery adjustment command, the preset congestion index threshold is adjusted according to the change characteristics of the transmission anomaly index within a preset observation period.

[0030] In this embodiment, the number of agricultural machines refers to the total number of agricultural machines currently connected to the same communication channel, obtained by the edge gateway or base station based on the registration / connection status of each agricultural machine. Instantaneous occupancy rate refers to the proportion of time the communication channel is occupied at the current moment, measured by the air interface probe or the physical resource block utilization rate of the base station. Collision rate refers to the proportion of times multiple agricultural machines simultaneously transmit data, causing data packet collisions, out of the total number of transmissions, obtained by the MAC layer collision counter of the base station or edge gateway. Retransmission rate refers to the proportion of the number of data packets that need to be retransmitted due to transmission failure to the total number of transmitted data packets, calculated by the automatic retransmission request log of the terminal or gateway node. Transmission success rate refers to the proportion of the number of data packets successfully transmitted by the agricultural machine terminal to the total number of data packets attempted to be transmitted, obtained by the MAC layer confirmation feedback of the agricultural machine vehicle terminal. Signal reception power refers to the reference signal reception power received by the edge gateway or base station from the agricultural machine's uplink signal, obtained by the base station physical layer measurement report. Instantaneous speed refers to the moving speed of the agricultural machine at the current moment, calculated in real time by the agricultural machine vehicle's GPS / BeiDou positioning module. Handover frequency refers to the number of times an agricultural machine switches between different communication channels or cells per unit time, obtained through statistics from base station handover management logs or terminal measurement reports. Causal chain location refers to the role of the agricultural machine in the near-far effect and mobile handover anomalies (strong signal end or weak signal end, handover oscillation zone or edge coverage zone) and its spatial correlation with surrounding agricultural machines and gateways, determined by analyzing the agricultural machine's signal reception power ranking, instantaneous speed, handover frequency, and the spatial geometric relationship between GPS coordinates and gateway location.

[0031] The preset congestion index threshold is a key benchmark for determining whether a channel has entered a high-load state. It depends on the concurrent density of farm operations and the required detection sensitivity, and is usually set between 0.6 and 1.5. In this embodiment, it is set to 1.2, which balances the timeliness of detection with the false alarm rate. It can provide early warning of potential congestion without frequently triggering the abnormal diagnosis process due to an excessively low threshold, thereby reducing system overhead.

[0032] The preset anomaly index threshold is a decision threshold that distinguishes between normal channel fluctuations and truly abnormal states. It depends on the fluctuation distribution of historical channel statistics and the reliability requirements of the service, and is usually set between 0.4 and 0.8. In this embodiment, it is set to 0.65, which ensures that anomalies are not triggered when the channel is slightly congested, avoiding ineffective adjustments; while responding quickly when the channel enters a stage of continuous deterioration, achieving a balance between sensitivity and stability.

[0033] By employing a hierarchical diagnostic logic that considers both overall channel load and individual agricultural machinery behavior, a causal chain is established between secure transmission of agricultural machinery operation data and anomalies in the physical layer of wireless communication. First, a load congestion index is constructed using the number of agricultural machines and instantaneous occupancy rate to identify high-load channels, as channel congestion is the root cause of collisions and retransmissions. Then, based on collision rate and retransmission rate, transmission anomaly events are determined, transforming macroscopic channel deterioration into quantifiable anomaly signals. Subsequently, within the anomaly channels, agricultural machinery operation feature vectors are constructed using transmission success rate and signal reception power to identify outlier agricultural machines, since low success rate and weak reception power directly reflect local interference or coverage issues. Black holes; by combining instantaneous velocity and switching frequency, the near-far effect and the oscillation of movement switching are characterized respectively, so as to accurately locate the complex abnormal agricultural machinery that is both suppressed by long distance and frequently switches, revealing the coupled fault that cannot be explained by a single indicator; finally, adjustment instructions are generated according to the spatial position and causal chain relationship of the complex abnormal agricultural machinery, and the congestion index threshold is reversed according to the change of the transmission anomaly index within the observation period. This significantly improves the reliability, security and adaptability of agricultural machinery data transmission in the complex dynamic environment of large plain farms, and effectively solves the problems of lag in anomaly identification and low data transmission accuracy caused by the inability to adapt to complex and changing operating conditions.

[0034] Specifically, the process of determining the load congestion index based on the number of agricultural machines and the instantaneous occupancy rate includes: The load congestion index is obtained by nonlinear coupling calculation based on the number of agricultural machines, the preset maximum number of agricultural machines to be supported, the instantaneous occupancy rate, and the preset channel saturation occupancy rate threshold.

[0035] The formula for calculating the load congestion index is: ; Where C is the load congestion index, N is the number of agricultural machines, Nmax is the preset maximum number of agricultural machines, U is the instantaneous occupancy rate, Umax is the preset channel saturation occupancy rate threshold, and k is the preset adjustment factor.

[0036] The preset maximum number of agricultural machines to be supported refers to the maximum number of agricultural machines that a single communication channel can support simultaneously while ensuring basic transmission quality. It is used to characterize the upper limit of the terminal capacity that the channel can accommodate. It depends on the channel bandwidth, modulation and coding method, agricultural machine data reporting frequency, and single packet data volume. It is usually set between 10 and 50 machines. In this embodiment, it is set to 30 machines, which can meet the routine operation needs of a large farm with a single channel and provide a reasonable normalization benchmark for the channel congestion index, avoiding the congestion index being too low for a long time due to an excessively high upper limit, thus failing to trigger detection.

[0037] The preset channel saturation threshold refers to the critical point at which further increasing the load will lead to a significant increase in collision rate and retransmission rate after the channel resource utilization reaches this value. It is used to characterize the saturation threshold of physical transmission resources and depends on the characteristics of the channel physical layer (such as carrier sense mechanism and backoff algorithm efficiency) and the service's sensitivity to latency. It is usually set between 70% and 85%. In this embodiment, it is set to 80%, which can provide early warning before the channel is completely congested, allowing response time for subsequent anomaly detection and adjustment measures, while avoiding frequent misjudgments due to an excessively low threshold.

[0038] The preset adjustment factor is an adjustable coefficient used to control the nonlinear amplification of the load congestion index by the product of the number of agricultural machines and the channel occupancy rate. It depends on the dynamic nature of the farm operation scenario (the speed of agricultural machines and the frequency of channel switching) and the requirements for congestion sensitivity. It is usually set between 2 and 5. The larger the value, the stronger the penalty for high load conditions. In this embodiment, it is set to 3, so that when both the number of agricultural machines and the channel occupancy rate are close to the upper limit, the index term can amplify the congestion index by about 20 times, thereby significantly distinguishing between medium load and high load congestion, and improving the sensitivity and robustness of high load channel determination.

[0039] If the occupancy rate is close to saturation even when the number of agricultural machines is below the upper limit, it indicates high-volume operations or long data packet transmissions, which can also lead to a high risk of collisions. Conversely, if the occupancy rate is low but the number of agricultural machines is extremely large, dense random access can also cause congestion. Therefore, by introducing a preset maximum number of agricultural machines and a preset channel saturation occupancy rate threshold to calculate the load congestion index, the positive correlation between the collision probability and the two factors under multi-terminal shared channel conditions is simulated. When the proportion of agricultural machines exceeds the capacity threshold or the instantaneous occupancy rate exceeds the saturation threshold, the index exhibits superlinear growth, thus more sensitively reflecting the critical state of high load. When both are below the threshold, the index remains at a low level, avoiding misjudgments and preventing missed detections or false alarms caused by looking only at the number or occupancy rate, thereby improving the robustness and accuracy of channel assessment in complex smart farm operation scenarios.

[0040] Please see Figure 2 As shown, it is the determination logic diagram for determining high-load channels in this embodiment. In this embodiment, when the load congestion index is greater than the preset congestion index threshold, the channel under test is determined to be a high-load channel, and several high-load channels are obtained.

[0041] The intensity of competition for channel resources is determined by both the number of agricultural machines and their instantaneous occupancy rate. These two factors are not linearly additive but rather coupled. As the number of agricultural machines increases, the rising occupancy rate significantly amplifies the probability of collisions. Therefore, a load congestion index, determined based on both factors, is used as a single quantitative indicator and compared with a preset threshold. Essentially, this constructs a competition-risk discrimination threshold. Only when the load congestion index exceeds the threshold is the channel classified as a high-load channel. Channels below the threshold, even with high local occupancy, have a relatively small number of agricultural machines, and their actual collision risk remains within a controllable range, eliminating the need to trigger subsequent complex anomaly detection procedures. Only when both a large number of machines and a high occupancy rate are simultaneously met does a channel truly enter the high-collision-risk zone. This avoids unnecessary collision rate and retransmission rate collection and calculation on low-risk channels, significantly reducing the system's real-time processing overhead. Furthermore, it provides a precise input range for judging subsequent data transmission anomalies, allowing diagnostic resources to focus on high-load channels where anomalies are truly likely, thereby improving the real-time performance and engineering feasibility of the entire secure transmission method.

[0042] Specifically, the process of determining the transmission anomaly index based on collision rate and retransmission rate includes: Calculate the standard deviation of the collision rate within a preset time period to obtain the collision fluctuation value; Calculate the standard deviation of the retransmission rate within the preset time period to obtain the retransmission rate fluctuation value; The transmission anomaly index is obtained by weighting the collision fluctuation value, the retransmission rate fluctuation value, the preset collision weight, and the preset retransmission weight.

[0043] The preset collision weight and preset retransmission weight are weighted coefficients used to measure the relative importance of collision rate fluctuations and retransmission rate fluctuations when calculating the transmission anomaly index. They reflect the degree of influence of collision rate fluctuations and retransmission rate fluctuations on channel anomaly judgment, respectively. These values ​​depend on the actual needs of collision sensitivity and retransmission sensitivity in the smart farm, as well as the dominant factors of channel anomaly types. They are typically set between 0.2 and 0.8, with a total sum of 1. In this embodiment, the collision weight is set to 0.6 and the retransmission weight to 0.4. Collisions are the primary manifestation of channel congestion; assigning higher weights allows the transmission anomaly index to respond quickly to collision rate fluctuations. Simultaneously, the retransmission weight serves as an auxiliary factor to avoid misjudgments caused by fluctuations in a single indicator, improving the accuracy and robustness of anomaly detection.

[0044] In this embodiment, the instantaneous values ​​of collision rate and retransmission rate are easily affected by occasional brief interference or random backoff in the channel, resulting in spikes. However, such instantaneous jitter does not mean that the channel has entered a continuous abnormal state. By using the collision fluctuation value and retransmission rate fluctuation value within a preset time period, it is essentially to extract the degree of instability of the two indicators in the time domain. The larger the fluctuation value, the more uneven the distribution of collision or retransmission behavior on the time axis and the more violent the oscillation. This instability is precisely a typical precursor feature of the channel evolving from mild congestion to severe conflict.

[0045] Collision fluctuation values ​​reflect the irregularity of competition among multiple agricultural machines accessing the network. For example, sudden surges in operations can cause collision rates to fluctuate wildly. Retransmission rate fluctuation values, on the other hand, characterize retransmission delay jitter at the link layer caused by signal attenuation or hidden terminals. Physically, these two factors are complementary. Collision fluctuations alone can be mitigated by rapid backoff, while retransmission fluctuations may stem from coverage changes caused by the movement of individual agricultural machines. However, when both fluctuate significantly, the weighted transmission anomaly index rises dramatically, accurately indicating that the channel has entered a deteriorating coupling phase that cannot be resolved through contention backoff or repaired by single-point retransmissions. Compared to directly using the instantaneous average of collision rate or retransmission rate, this fluctuation-based weighted method effectively suppresses false alarms and is more sensitive to continuously deteriorating channel responses. This provides statistically stable and physically causally clear feature inputs for subsequent determination of abnormal data transmission events, significantly improving the reliability and engineering practicality of anomaly detection in the complex dynamic environment of smart farms.

[0046] Please see Figure 3 As shown, this is a logic diagram for determining the occurrence of a data transmission anomaly event in this embodiment. In this embodiment, a data transmission anomaly event is determined to have occurred on the high-load channel when the transmission anomaly index is greater than the preset anomaly index threshold.

[0047] In this embodiment, collision rate and retransmission rate characterize data transmission quality from two causally coupled dimensions: channel access conflict and link recovery capability. Collision rate directly reflects the probability of physical layer collisions caused by multiple agricultural machines transmitting simultaneously, while retransmission rate reflects the degree of recovery failure after the data link layer triggers the ARQ mechanism due to collision or signal attenuation. The two are not independent; an increase in collision rate usually pushes up the retransmission rate, but the retransmission rate may also be affected by nonlinear factors such as weak coverage and hidden terminals. Therefore, merging the two into a transmission anomaly index essentially constructs a joint collision-recovery deterioration threshold: only when this index is greater than a preset anomaly index threshold is it determined that a data transmission anomaly event has occurred on a high-load channel, rather than triggering intervention based solely on a single fluctuation in collision rate or retransmission rate.

[0048] By ensuring that the transmission anomaly index does not exceed the threshold when the collision rate occasionally increases but the retransmission rate remains low, overreaction to transient interference is avoided. Conversely, if the collision rate remains high and the retransmission rate rises simultaneously, the index quickly exceeds the threshold, accurately identifying that the channel has entered an abnormal state that cannot be healed. This not only filters out normal short-term jitter in the channel but also ensures that only persistent anomalies that truly threaten the secure transmission of data will activate the subsequent agricultural machinery-level positioning and adjustment process. Thus, in the highly dynamic and parallel operating environment of smart farms, the optimal balance between the sensitivity and robustness of anomaly detection is achieved.

[0049] Please see Figure 4 As shown, this is the logic diagram for determining abnormal agricultural machinery in this embodiment. In this embodiment, several abnormal agricultural machines are determined based on the outlier of the agricultural machinery operation feature vector. The process of constructing the agricultural machinery operation feature vector based on the transmission success rate and signal reception power includes: The agricultural machinery operation feature vector is constructed based on the transmission success rate and the signal reception power. The outlier degree is obtained by calculating the local outlier factor based on the agricultural machinery operation feature vector of each agricultural machinery under test and all agricultural machinery under test in the high-load channel; When the outlier degree is greater than the preset outlier threshold, the agricultural machine under test is determined to be the abnormal agricultural machine.

[0050] The preset outlier threshold is a critical value for identifying abnormal agricultural machinery based on local outlier factors. It depends on the inherent dispersion of the agricultural machinery feature vectors within the channel and the service's tolerance for false detections, and is typically set between 1.2 and 2.0. In this embodiment, it is set to 1.5, so that when the transmission success rate or signal reception power of an agricultural machinery deviates from that of most agricultural machinery in the channel by a moderate or greater degree, it is judged as abnormal. This avoids frequent adjustments triggered by small fluctuations, while ensuring timely intervention for agricultural machinery with significantly deteriorated performance, achieving a good balance between detection efficiency and stability.

[0051] In this embodiment, in a high-load channel, the transmission success rate and signal reception power of most agricultural machines will exhibit a certain joint distribution pattern. Agricultural machines operating normally usually have a high transmission success rate and a signal reception power within a reasonable range, neither too weak to cause incorrect demodulation nor too strong to cause nonlinear distortion or suppress neighboring agricultural machines. However, a few abnormal agricultural machines may exhibit low success rate and low power, coverage blind spots or excessive distance, low success rate with medium to high power, hidden terminals, collisions leading to retransmission exhaustion or transmitter failure, or high success rate with abnormal power, or potential interference sources.

[0052] By using a local outlier factor to calculate outlier degree, it essentially compares the density deviation of each agricultural machine in its local neighborhood. It can automatically adapt to the distribution differences of different density areas in a multi-dimensional feature space without the need to preset a global threshold. Although the product or combination of the transmission success rate and the signal reception power is not linearly separable, the local outlier factor can sensitively capture individuals that are obviously different from others even though their surrounding neighbors are normal, through reachable distance and local reachable density. Even if the absolute value of these abnormal agricultural machines does not exceed the conventional threshold, such as a slightly lower success rate in an overall low-power area, they may become outliers. This effectively overcomes the lack of adaptability of fixed thresholds under dynamic channel conditions, accurately screens out candidate abnormal agricultural machines that need further analysis of near-far effect and mobile handover, and significantly improves the accuracy and robustness of locating faulty agricultural machines in the complex electromagnetic environment of large smart farms.

[0053] Specifically, the near-far effect characteristic value is determined based on the product of the difference in signal received power between the abnormal agricultural machine and the agricultural machine with the strongest signal in the channel and the spatial collinearity of the two relative to the gateway. The spatial collinearity is calculated based on the absolute value of the cosine of the angle between the abnormal agricultural machine and the agricultural machine with the strongest signal and the gateway. The mobile handover characteristic value is determined based on the product of the instantaneous speed of the abnormal agricultural machine and the handover frequency.

[0054] In this embodiment, the calculation method of the near-far effect characteristic value F is as follows: First, within the channel determined to be high-load, the reference signal received power R of each agricultural machine is obtained, and the agricultural machine with the largest R value is selected and denoted as the strongest signal agricultural machine. Let the reference signal received power of the abnormal agricultural machine be Pi, and the reference signal received power of the strongest signal agricultural machine be P0. The power difference ΔP = P0 - Pi is calculated. Second, the spatial collinearity θ of the abnormal agricultural machine and the strongest signal agricultural machine relative to the edge gateway (base station) is calculated. Let the gateway position be point O, the abnormal agricultural machine position be point A, and the strongest signal agricultural machine position be point B. The GPS coordinates of the three are obtained respectively, and the absolute value of the cosine of the angle between vectors OA and OB is calculated: θ = |cos∠AOB| = |(OA·OB)| / (|OA|·|OB|). The value range of θ is [0,1]. The closer θ is to 1, the closer the three are to collinearity, and the more significant the near-far effect is. Finally, the near-far effect characteristic value F = ΔP × θ, where F has the dimension of dB. The larger the value, the more severe the near-far effect suppression is on the abnormal agricultural machinery.

[0055] In this embodiment, the instantaneous speed and handover frequency are subjected to maximum-minimum normalization during the calculation of the mobile handover feature value. The mobile handover feature value is obtained by calculating the product of the normalized instantaneous speed and handover frequency.

[0056] In this embodiment, the near-far effect is not simply determined by the power difference, but also strongly depends on the spatial geometric relationship between the two agricultural machines and the gateway. If the angle between the lines connecting the two machines and the gateway is very small, i.e., they are almost on the same ray, the strong-signal agricultural machine will physically block the uplink data of the weak-signal agricultural machine, causing the weak-signal agricultural machine to continuously retreat or fail to retransmit. Therefore, by using the product of the power difference and the spatial collinearity, and taking the degree of geometric collinearity as a weighting coefficient, the pseudo-near-far effect in non-collinear scenarios is effectively suppressed, so that the characteristic value only increases significantly under actual conditions where there is a strong risk of suppression. The mobile handover characteristic value is the product of instantaneous speed and handover frequency. At low speeds, even if the handover frequency is high, it is often a normal handover caused by overly detailed channel planning and does not constitute an anomaly. At high speeds, if the handover frequency is also high, it means that the agricultural machine is crossing multiple cell boundaries at a relatively fast speed, which can easily cause ping-pong handover, synchronization loss, and signaling storms. The product of the two amplifies the destructive power of this high-speed fluctuation on data transmission.

[0057] By recognizing that instantaneous speed and switching frequency are physically positively correlated but not perfectly linear, the product index can distinguish between low-speed high-frequency and high-speed high-frequency. Meanwhile, the spatial collinearity in the near-far effect feature introduces an angle-sensitive factor, ensuring that even with large power differences, spatial orthogonality will not lead to misjudgment as a composite anomaly. This approach captures the coupling effect of mobility in the power domain, spatial domain, and temporal domain in a concise mathematical form, providing a physically meaningful and sensitively controllable quantitative feature for subsequent dual-threshold determination. This significantly improves the accuracy and interpretability of identifying composite anomaly agricultural machinery in the complex dynamic environment of smart farms.

[0058] Specifically, when the near-far effect feature value is greater than a preset first feature threshold and the movement switching feature value is greater than a preset second feature threshold, the abnormal type of the abnormal agricultural machine is determined to be a near-far movement composite abnormality, and the abnormal agricultural machine is determined to be a composite abnormal agricultural machine.

[0059] The preset first feature threshold is a critical value used to determine whether abnormal agricultural machinery is significantly affected by the near-far effect. It depends on the signal receiving power distribution range of normal agricultural machinery in the channel, the acceptable degree of power difference, and the sensitivity requirements of spatial collinearity. It is usually set between 3dB and 10dB. In this embodiment, it is set to 6dB, which can effectively identify the obvious suppression of weak signal agricultural machinery by strong signal agricultural machinery, and avoid misjudgment caused by slight power difference or low collinearity.

[0060] The preset second feature threshold is a critical value used to determine whether abnormal agricultural machinery is causing frequent handover anomalies due to high-speed movement. It depends on the distribution of normal agricultural machinery movement speed within the farm, the design redundancy of the base station coverage handover area, and the tolerance for handover jitter. It is usually set between 0.2 and 0.6. In this embodiment, it is set to 0.4, which can effectively distinguish between normal movement and ping-pong handover scenarios caused by high-speed movement, avoid frequently triggering handover parameter adjustments for normally moving agricultural machinery, and at the same time ensure timely intervention for agricultural machinery whose communication is truly abnormal due to mobility.

[0061] In this embodiment, the near-far effect and mobile handover are two fault modes with different physical mechanisms that cause abnormal agricultural machinery data transmission. The near-far effect stems from the excessive difference in signal power between multiple agricultural machines and the gateway, with strong-signal agricultural machines suppressing weak-signal ones, which is a spatial power imbalance. Mobile handover, on the other hand, stems from the frequent switching between different channels when agricultural machines move at high speed, leading to increased signaling overhead, data interruption, or ping-pong effect, which is a temporal connectivity oscillation. When the near-far effect occurs alone, it can be partially mitigated by power control or dynamic adjustment of the contention window. When mobile handover anomalies occur alone, the handover hysteresis threshold or hold time can be optimized. However, when both occur simultaneously and each exceeds its characteristic threshold, a strong positive feedback coupling is formed: the weak-signal agricultural machines suppressed by the near-far effect are more likely to lose synchronization during handover, and frequent handover makes power control unable to converge, further exacerbating the near-far effect.

[0062] The near-farm effect eigenvalues ​​characterize the suppression intensity of agricultural machinery in the power domain, while the movement switching eigenvalues ​​characterize the turbulence intensity of agricultural machinery in the connectivity domain. Only when both exceed their respective preset thresholds is the agricultural machinery identified as having a composite abnormality. This avoids misjudging agricultural machinery with only a single abnormality as having a composite fault requiring special handling, thereby accurately identifying agricultural machinery that is both suppressed and frequently fluctuating, reducing unnecessary intervention operations, and enabling subsequent agricultural machinery adjustment commands to perform migration channels and parameter optimization for real coupled fault scenarios. This significantly improves the diagnostic accuracy and resource utilization efficiency of data transmission security methods in the high-dynamic operation environment of smart farms.

[0063] The process of determining agricultural machinery adjustment instructions based on the spatial location and causal chain location of the composite abnormal agricultural machinery includes: Based on the spatial location of the composite abnormal agricultural machine, the list of available communication channels is scanned, and candidate channels with a load congestion index less than a preset congestion index threshold are selected. If a candidate channel exists, the composite abnormal agricultural machine is moved to the candidate channel with the smallest load congestion index, and an adjustment instruction is obtained. If no candidate channel exists, the adjustment command for the agricultural machine is determined based on the causal chain position of the composite abnormal agricultural machine. When the composite abnormal agricultural machine is a strong signal terminal and is in the switching oscillation zone, an adjustment command is generated to reduce its transmission power and increase its switching hysteresis parameter. When the composite abnormal agricultural machine is a weak signal terminal and is in the edge coverage zone, an adjustment command is generated to increase its transmission power, and a nearby low-mobility agricultural machine is dispatched as a mobile relay. At the same time, an adjustment command is generated to reduce the data reporting frequency of the relay agricultural machine. The generated adjustment instructions are integrated and output as agricultural machinery adjustment instructions.

[0064] In this embodiment, the composite abnormal agricultural machine is simultaneously affected by the near-far effect and the oscillation of mobile handover. If it continues to remain in the original high-load channel, power control will be difficult to converge, and handover conflicts will continue to worsen, forming a positive feedback loop. Therefore, it is necessary to temporarily isolate it from the current congested and abnormal channel and migrate it to a clean channel with a lower load to break the original vicious coupling chain. The real-time spatial coordinates of the agricultural machine determine the coverage area of ​​the adjacent channels available around it. Selecting a temporary channel based on its spatial location can ensure that the agricultural machine can still maintain a basic communication connection after migration, avoiding disconnection due to channel coverage blind spots. At the same time, migration alone is not enough to solve the problem. It is also necessary to adjust the power and handover parameters according to the causal chain position: the causal chain position refers to whether the composite abnormal agricultural machine is suppressed or actively suppressing in the original channel, such as its relative position to the agricultural machine with the strongest signal, its distance from the gateway, and whether it frequently initiates handover or passively responds in the handover chain. Its transmission power and handover parameters are adjusted accordingly. If it is a weak signal agricultural machine, the power is appropriately increased; if it is a strong signal agricultural machine, the maximum power is limited, which can alleviate the near-far effect and ping-pong handover from the source.

[0065] The spatial location is directly linked to the feasibility of temporary channels and the continuity of communication coverage; the causal chain location, through time-series analysis of historical power differences, collinearity, speed and switching frequency, infers the role of composite anomalies in the causal network, thereby achieving dual intervention of migration avoidance and parameter fine-tuning. This not only immediately relieves the congestion pressure of the current channel, but also prevents the recurrence of similar anomalies in temporary channels, ultimately enabling the data transmission of the entire agricultural machinery fleet to gradually return to a stable state, significantly improving the control effectiveness and system convergence speed in the complex dynamic environment of smart farms.

[0066] The process of adjusting the preset congestion index threshold based on the change characteristics of the transmission anomaly index within a preset observation period includes: Calculate the rate of change of all the transmission anomaly indices within the preset observation period to obtain the rate of change of the indices; When the rate of change of the exponential index is greater than the preset change threshold, the preset congestion index threshold is reduced according to the relative deviation between the rate of change of the exponential index and the preset change threshold and the preset adjustment coefficient, where M'=M×[1-a×(|G-G'| / G')], M' is the adjusted preset congestion index threshold, M is the preset congestion index threshold, a is the preset adjustment coefficient, G is the rate of change of the exponential index, and G' is the preset change threshold.

[0067] The preset observation duration is the time window for analyzing the characteristics of transmission anomaly index changes. It depends on the typical cycle of changes in farm channel status and the requirements for adjustment response speed. It is usually set between 10 and 60 seconds. In this embodiment, it is set to 30 seconds to ensure that the trend can be identified and the threshold adaptive adjustment can be triggered within about half a minute when the channel is continuously congested (such as multiple farm machines occupying it for a long time), thus achieving a balance between stability and sensitivity.

[0068] The preset change threshold is a critical value used to determine whether the change in the transmission anomaly index is significant. It depends on the range of random fluctuations in the index under normal channel conditions and the system's conservatism in adjusting the threshold. It is usually set between 0.05 seconds and 0.20 seconds. In this embodiment, it is set to 0.10 seconds, which can eliminate noise interference caused by occasional collisions or instantaneous retransmissions, and ensure that the preset congestion index threshold is adaptively reduced only when the channel continues to deteriorate, thereby improving the robustness of the diagnostic system.

[0069] The preset adjustment coefficient is a factor used to control the adjustment range of the preset congestion index threshold. It depends on the system's trade-off between threshold convergence speed and stability, as well as the need to suppress oscillations that may be caused by excessive adjustment. It is usually set between 0.1 and 0.5. In this embodiment, it is set to 0.3, which can respond quickly to changes in channel congestion and avoid drastic threshold fluctuations caused by excessive adjustment in a single instance, thus ensuring smooth system convergence.

[0070] In this embodiment, if the preset congestion index threshold is fixed, it will not be able to adapt to the long-term impact of dynamic factors such as the scale of operations, crop growth shading, and weather changes on channel capacity in smart farms. When the overall workload of the farm or the electromagnetic environment changes in a trend, the originally appropriate threshold may become too sensitive (leading to frequent misjudgments of high-load channels) or too insensitive (missing to detect real congestion), thereby affecting the stability of the entire secure transmission method.

[0071] By introducing the rate of change of the transmission anomaly index as a feedback signal, the transmission anomaly index of all data transmission anomaly events that have occurred is continuously calculated within a preset observation period, and its rate of change is extracted. This rate of change directly reflects whether the collision and retransmission fluctuations in high-load channels tend to intensify, alleviate, or remain stable. When the rate of change of the index exceeds the preset change threshold, it indicates that the degree of anomaly is rapidly deteriorating. Even if the current congestion index has not yet generally exceeded the standard, the congestion index threshold needs to be lowered so that more channels enter the high-load judgment range in advance and preventive intervention is implemented. This makes the congestion index threshold no longer a static empirical value, but a dynamic and continuous optimization based on the actual operation of the farm. It realizes intelligent iteration throughout the entire process and significantly improves the robustness, self-consistency, and scenario generalization ability of the method in long-term operation.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for secure data transmission of agricultural machinery operations in smart farms, characterized in that, include: In a large-scale plain farm, several agricultural machines under test operate in parallel and switch communication channels in real time. The number of agricultural machines and the instantaneous occupancy rate of each channel under test are collected in real time. Based on the comparison between the load congestion index and the preset congestion index threshold, several high-load channels are identified. The load congestion index is determined based on the number of agricultural machines and the instantaneous occupancy rate. The collision rate and retransmission rate of each high-load channel are collected in real time. The transmission anomaly index is compared with the preset anomaly index threshold to determine that a data transmission anomaly event has occurred in the high-load channel. The transmission anomaly index is determined based on the collision rate and retransmission rate. Based on the data transmission anomaly event, the transmission success rate and signal reception power of each of the agricultural machines under test in the high-load channel are collected. Several abnormal agricultural machines are identified according to the outlier of the agricultural machine operation feature vector, wherein the agricultural machine operation feature vector is constructed based on the transmission success rate and signal reception power. The instantaneous speed and switching frequency of the abnormal agricultural machinery are collected, and the composite abnormal agricultural machinery is determined based on the near-far effect characteristic value and the moving switching characteristic value. The near-far effect characteristic value is determined based on the signal receiving power and spatial position relationship, and the moving switching characteristic value is determined based on the instantaneous speed and switching frequency. The adjustment instructions for the agricultural machinery are determined based on the spatial location and causal chain location of the composite abnormal agricultural machinery. Based on the agricultural machinery adjustment command, the preset congestion index threshold is adjusted according to the change characteristics of the transmission anomaly index within a preset observation period.

2. The method for secure data transmission of agricultural machinery operations for smart farms according to claim 1, characterized in that, When the load congestion index is greater than the preset congestion index threshold, the channel under test is determined to be a high-load channel, and several high-load channels are obtained.

3. The method for secure data transmission of agricultural machinery operations in smart farms according to claim 2, characterized in that, The process of determining the load congestion index based on the number of agricultural machines and instantaneous occupancy rate includes: The load congestion index is obtained by nonlinear coupling calculation based on the number of agricultural machines, the preset maximum number of agricultural machines to be supported, the instantaneous occupancy rate, and the preset channel saturation occupancy rate threshold.

4. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 3, characterized in that, When the transmission anomaly index is greater than the preset anomaly index threshold, it is determined that a data transmission anomaly event has occurred on the high-load channel.

5. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 4, characterized in that, The process of determining the transmission anomaly index based on collision rate and retransmission rate includes: The collision fluctuation value is calculated based on the collision rate within a preset time period; Calculate the retransmission rate fluctuation value based on the retransmission rate within the preset time period; The transmission anomaly index is calculated by weighting the collision fluctuation value and the retransmission rate fluctuation value.

6. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 5, characterized in that, Several abnormal agricultural machines are identified based on the outlier of their operational feature vectors. The process of constructing the agricultural machine operational feature vectors based on transmission success rate and signal reception power includes: The agricultural machinery operation feature vector is constructed based on the transmission success rate and the signal reception power. The outlier degree is obtained by calculating the local outlier factor based on the agricultural machinery operation feature vector of each agricultural machinery under test and all agricultural machinery under test in the high-load channel; When the outlier degree is greater than the preset outlier threshold, the agricultural machine under test is determined to be the abnormal agricultural machine.

7. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 6, characterized in that, When the near-far effect feature value is greater than a preset first feature threshold and the movement switching feature value is greater than a preset second feature threshold, the abnormal type of the abnormal agricultural machine is determined to be a near-far movement composite abnormality, and the abnormal agricultural machine is determined to be a composite abnormal agricultural machine.

8. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 7, characterized in that, The near-far effect characteristic value is determined based on the difference in signal received power between the abnormal agricultural machine and the agricultural machine with the strongest signal in the channel, and the product of their spatial collinearity relative to the gateway. The spatial collinearity is calculated based on the absolute value of the cosine of the angle between the abnormal agricultural machine and the agricultural machine with the strongest signal and the gateway. The mobile handover characteristic value is determined based on the product of the instantaneous speed of the abnormal agricultural machine and the handover frequency.

9. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 8, characterized in that, The process of determining agricultural machinery adjustment instructions based on the spatial location and causal chain location of the composite abnormal agricultural machinery includes: Based on the spatial location of the composite abnormal agricultural machine, a temporary communication channel with a load congestion index less than a preset congestion index threshold is determined, and the composite abnormal agricultural machine is moved to the temporary communication channel to obtain an agricultural machine adjustment command. Adjusting power and switching parameters based on the causal chain position of the complex abnormal agricultural machinery yields the agricultural machinery adjustment command.

10. The method for secure transmission of agricultural machinery operation data for smart farms according to claim 9, characterized in that, The process of adjusting the preset congestion index threshold based on the change characteristics of the transmission anomaly index within a preset observation period includes: The rate of change of the index is calculated based on all the transmission anomaly indices within the preset observation period; The preset congestion index threshold is adjusted based on the comparison between the rate of change of the index and the preset change threshold.

Images