Low-altitude helicopter communication navigation signal enhancement method and system based on multiple channels

Abstract

The application discloses a low-altitude helicopter communication navigation signal enhancement method and system based on multiple channels, which comprises the following steps: extracting feature parameters from a first signal set, performing preliminary enhancement processing on the signal through a Kalman filtering algorithm, superimposing a preset attenuation compensation model for multipath effects caused by terrain complexity, and obtaining an enhanced second signal set; analyzing terrain complexity characteristics by using a fast Fourier transform algorithm according to a first geographic data set, combining real-time requirements, and adopting a parallel computing framework to accelerate processing, so as to obtain a second geographic data set containing terrain characteristics; extracting obstacle data from the second geographic data set, generating a preliminary three-dimensional model through a voxel grid method, superimposing a volume interpolation algorithm for the spatial distribution characteristics of the obstacle data, and obtaining an optimized first three-dimensional model.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of information technology, and particularly relates to a multi-channel-based low-altitude helicopter communication navigation signal enhancement method and system. BACKGROUND

[0002] The technical progress in the field of low-altitude helicopter communication navigation is of decisive significance to flight safety and mission efficiency. With the opening of low-altitude airspace and the increasing application of helicopters, the reliability and accuracy of communication navigation systems have become key issues that cannot be ignored. Especially in complex terrain and severe weather conditions, how to ensure stable signal transmission and accurate data processing directly relates to the success or failure of flight operations. Research in this field not only promotes the innovation of aviation technology, but also provides important support for low-altitude traffic management.

[0003] However, the current communication navigation enhancement method still has significant limitations. Traditional single-channel signal processing technology often has difficulty maintaining stable performance in areas with severe signal attenuation or multipath interference, such as mountainous areas and urban canyons. Existing solutions rely on a single data source and lack the ability to integrate diverse geographic information, resulting in decreased navigation accuracy in dynamic environments and even signal loss. These defects expose the shortcomings of existing technology in adapting to complex low-altitude environments.

[0004] Specifically, the core challenges in this field focus on the optimization of multi-channel signal processing technology and the effective integration of geographic information data. First, multi-channel technology needs to address data synchronization and interference suppression between different channels, otherwise the effectiveness of signal enhancement will be greatly reduced. Second, multi-source fusion analysis of geographic information faces the challenges of data heterogeneity and high real-time requirements, especially in areas with weak signals or interference. How to quickly integrate terrain, obstacles, and weather information to form a reliable three-dimensional model is the key to technological breakthrough. These unresolved technical factors make it difficult to guarantee the reliability of navigation systems in complex environments, thereby deriving unique difficulties in improving low-altitude flight safety.

[0005] Therefore, how to realize real-time fusion of geographic information data and signal enhancement through multi-channel technology to overcome interference and data processing bottlenecks in complex terrain environments has become a key issue in improving the performance of low-altitude helicopter communication navigation systems. SUMMARY

[0006] The technical problem to be solved by the present application is to provide a multi-channel-based low-altitude helicopter communication navigation signal enhancement method and system.

[0007] To achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows:

[0008] A low-altitude helicopter communication navigation signal enhancement method based on multi-channel, comprising:

[0009] According to the original signal data collected by the multi-channel technology, a first signal set with synchronization and low interference is obtained;

[0010] From the first signal set, the feature parameters are extracted, the signal is preliminarily enhanced by Kalman filtering algorithm, and a preset attenuation compensation model is superimposed for the multipath effect caused by terrain complexity, to obtain a second signal set after enhancement;

[0011] According to the terrain data, obstacle data and meteorological information flow in the geographic information source, a first geographic data set is obtained;

[0012] According to the first geographic data set, a second geographic data set containing terrain features is obtained;

[0013] Obstacle data is extracted from the second geographic data set, a preliminary three-dimensional model is generated by the voxel grid method, and a visual interpolation algorithm is superimposed for the spatial distribution characteristics of the obstacle data, to obtain an optimized first three-dimensional model;

[0014] Obtain real-time meteorological information flow to obtain a second three-dimensional model matching the current environment;

[0015] According to the second three-dimensional model and the second signal set, a particle filtering algorithm is used to realize the fusion processing of signal and geographic information, the model parameters are adjusted to optimize the signal coverage range for low-altitude navigation requirements, and a third signal set is obtained after fusion enhancement.

[0016] As an optimization, the original signal data collected by the multi-channel technology is obtained, the time alignment of each channel data is performed by a preset time domain synchronization algorithm, and an adaptive filter is used to suppress the cross interference between channels for signal synchronization requirements, to obtain a first signal set with synchronization and low interference.

[0017] As an optimization, the terrain data, obstacle data and meteorological information flow in the geographic information source are obtained, the heterogeneous data is converted to a unified format by a data preprocessing module, and a standardized mapping method is used to solve the data heterogeneity problem, to obtain a first geographic data set.

[0018] As an optimization, according to the first geographic data set, the terrain complexity characteristics are analyzed by using the fast Fourier transform algorithm, and the parallel computing framework is used to speed up the processing according to the real-time requirements, to obtain a second geographic data set containing terrain features.

[0019] As an optimization, real-time meteorological information flow is obtained, the influence of meteorology on signal transmission is calculated by fluid dynamics simulation algorithm, and a time series prediction method is used to solve the dynamic changes of meteorological information flow, to obtain a second three-dimensional model matching the current environment.

[0020] The application also provides a multi-channel-based low-altitude helicopter communication navigation signal enhancement system, comprising:

[0021] A first processing module is configured to obtain a first signal set with synchronization and low interference according to original signal data collected by a multi-channel technology.

[0022] A second processing module is configured to extract feature parameters from the first signal set, perform preliminary enhancement processing on the signal by using a Kalman filtering algorithm, superimpose a preset attenuation compensation model for multipath effects caused by terrain complexity, and obtain a second signal set after enhancement.

[0023] A third processing module is configured to obtain a first geographic data set with consistency according to terrain data, obstacle data and meteorological information flow in a geographic information source.

[0024] A fourth processing module is configured to obtain a second geographic data set containing terrain features according to the first geographic data set.

[0025] A fifth processing module is configured to extract obstacle data from the second geographic data set, generate a preliminary three-dimensional model by using a voxel grid method, superimpose a volume interpolation algorithm for spatial distribution characteristics of the obstacle data, and obtain an optimized first three-dimensional model.

[0026] A sixth processing module is configured to obtain real-time meteorological information flow and obtain a second three-dimensional model matched with the current environment.

[0027] A seventh processing module is configured to perform fusion processing of the signal and the geographic information by using a particle filtering algorithm according to the second three-dimensional model and the second signal set, adjust model parameters to optimize the signal coverage range for low-altitude navigation requirements, and obtain a third signal set after fusion enhancement.

[0028] Preferably, the first processing module obtains original signal data collected by a multi-channel technology, performs time alignment on channel data by using a preset time domain synchronization algorithm, suppresses cross-interference between channels by using an adaptive filter for signal synchronization requirements, and obtains a first signal set with synchronization and low interference.

[0029] Preferably, the third processing module obtains terrain data, obstacle data and meteorological information flow in a geographic information source, converts heterogeneous data into a unified format by using a data preprocessing module, and obtains a first geographic data set with consistency by using a standardized mapping method for data heterogeneity problems.

[0030] Preferably, the fourth processing module analyzes terrain complexity characteristics by using a fast Fourier transform algorithm according to the first geographic data set, combines real-time requirements, and uses a parallel computing framework to speed up processing, and obtains a second geographic data set containing terrain features.

[0031] As preferred, the sixth processing module acquires a real-time meteorological information stream, calculates the influence of meteorology on signal transmission through a fluid dynamics simulation algorithm, adopts a time series prediction method for dynamic changes in the meteorological information stream, and obtains a second three-dimensional model matched with the current environment.

[0032] The original signal data is collected by the multi-channel technology, and the time alignment and cross-interference suppression between channels are realized by using the time domain synchronization algorithm and the adaptive filter. Subsequently, the signal is enhanced by using the Kalman filter and the attenuation compensation model to cope with the influence of multipath effect. Meanwhile, the geographic information such as terrain, obstacles and weather is fused, the terrain features are analyzed by using the fast Fourier transform and the parallel computing framework, and the optimized three-dimensional model is generated by using the voxel grid method and the volume visualization interpolation algorithm. Finally, the particle filtering algorithm is used to realize the fusion processing of the signal and the geographic information, and the signal coverage range is optimized according to the low-altitude navigation requirements. The application effectively solves the problem of signal transmission interference in complex terrain environment, and improves the accuracy and reliability of low-altitude navigation. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 The flowchart of the low-altitude helicopter communication navigation signal enhancement method based on multi-channel of the application. DETAILED DESCRIPTION

[0034] In order to further understand the content of the application, the application will be described in detail in conjunction with the drawings and examples. It can be understood that the specific examples described herein are only used to explain the related application, and not to limit the application. In addition, it should be noted that, in order to facilitate the description, only the parts related to the application are shown in the drawings.

[0035] Example 1:

[0036] As shown in the figure, the low-altitude helicopter communication navigation signal enhancement method based on multi-channel of the application includes: Figure 1

[0037] S101, the original signal data collected by the multi-channel technology is obtained, the time alignment of each channel data is carried out by using the preset time domain synchronization algorithm, the adaptive filter is used to suppress the cross-interference between channels according to the signal synchronization requirement, and the first signal set with synchronization and low interference is obtained.

[0038] ​The original signal data is obtained from the acquisition device through multi-channel technology, and each channel data is processed by using a preset time domain synchronization algorithm to obtain a second signal set with time alignment. According to the second signal set, an adaptive filter is used to analyze the data between channels to suppress cross interference, and a first signal set with synchronization and low interference is obtained. The data of each channel in the first signal set is obtained, and it is judged whether there is residual interference. If there is, the interference region is smoothed by linear interpolation method to obtain a third signal set. For the third signal set, the frequency domain characteristics are analyzed by using fast Fourier transform to determine whether the signal synchronization meets the preset threshold, and a fourth signal set with frequency domain synchronization is obtained. The time domain characteristics are extracted from the fourth signal set, and the time domain data is optimized again by using an adaptive filter to obtain a fifth signal set. According to the fifth signal set, it is judged whether the phase difference of each channel data is consistent. If not, the phase is adjusted by a phase correction algorithm to obtain a sixth signal set. Through the sixth signal set, the signal is smoothed by using mean filtering to obtain a final seventh signal set.

[0039] Specifically, the original signal data is obtained from the acquisition device through multi-channel technology, aiming to capture signals from multiple channels at the same time.

[0040] For example, in the field of audio processing, a microphone array may receive sound signals from different directions, and the original data may be multiple time domain waveforms with a sampling rate of 48kHz, containing noise and delay differences. Each channel data is processed by using a preset time domain synchronization algorithm, aiming to eliminate time deviation.

[0041] Specifically, the starting point of each channel signal can be detected, such as using the peak position, to control the time offset within 1ms, and obtain a second signal set with time alignment. This alignment lays a foundation for subsequent processing and avoids analysis errors caused by time misalignment. According to the second signal set, an adaptive filter is used to suppress cross interference.

[0042] For example, in a microphone array, a certain channel may be disturbed by the sound waves of adjacent channels. The adaptive filter adjusts the filter coefficient in real time, reduces the interference amplitude by about 20dB based on the least mean square error principle, and obtains a first signal set with synchronization and low interference.

[0043] Preferably, this method can dynamically adapt to environmental changes and improve signal purity. After obtaining the data of each channel in the first signal set, it is judged whether there is residual interference.

[0044] For example, if the channel still has periodic noise, the peak value exceeding the normal range can be found through spectral analysis. At this time, the interference region is smoothed by linear interpolation method, assuming that the interference segment length is 10 ms, the transition of the interpolated signal is natural, and the third signal set is obtained. This step effectively reduces the impact of sudden noise on subsequent analysis. For the third signal set, the frequency domain characteristics are analyzed by using fast Fourier transform.

[0045] For example, the time domain signal is converted to the frequency domain, the main frequency component in the range of 0-5 kHz is observed, and if the peak position deviation of the main frequency of each channel is less than 10 Hz, it is considered that the synchronization meets the threshold, and the fourth signal set is obtained. This process directly reflects the frequency domain consistency of the signal, providing a guarantee for high-precision applications. Time domain characteristics such as waveform amplitude and period are extracted from the fourth signal set, and are optimized by an adaptive filter.

[0046] For example, for the region with large amplitude fluctuations, the filter can further weaken the background noise, and the signal-to-noise ratio after optimization is improved to 30 dB, and the fifth signal set is obtained. This step enhances the time domain stability of the signal. According to the fifth signal set, it is judged whether the phase difference of each channel is consistent.

[0047] For example, if the phase of a certain channel is 15 degrees ahead, the phase can be adjusted based on the inverse tangent function to reduce the deviation to within 2 degrees by using a phase correction algorithm, and the sixth signal set is obtained.

[0048] It should be noted that the phase consistency is crucial for multi-channel signal synthesis and directly affects the quality of the final output. Through the sixth signal set, mean filtering is used for smoothing.

[0049] For example, the mean value is calculated with a 5-point sliding window to eliminate slight jitter, and the seventh signal set is finally obtained. This smoothing process significantly improves the smoothness and readability of the signal.

[0050] For example, the above process is particularly useful in the speech enhancement scene. The original signal may be distorted due to the distance between microphones or environmental echo. After time domain synchronization, interference suppression, frequency domain analysis and phase correction, the final signal clarity is improved by about 40%, which is suitable for downstream tasks such as speech recognition.

[0051] In one possible implementation, if the number of channels is increased to 8, the computational complexity of the adaptive filter increases, but the interference suppression effect is better, and the signal-to-noise ratio can be further improved by 5 dB.

[0052] It can be understood that each technical means is progressive, ensuring continuous optimization of signal quality and providing an efficient solution for multi-channel signal processing.

[0053] S102, extract the feature parameters from the first signal set, perform preliminary enhancement processing on the signal through the Kalman filtering algorithm, superimpose a preset attenuation compensation model for the multipath effect caused by the complexity of the terrain, and obtain an enhanced second signal set.

[0054] Obtain time domain data from the second signal set, smooth the signal by using mean filtering to obtain a third signal set. Extract frequency domain features from the third signal set, analyze the data by using fast Fourier transform to obtain a fourth signal set. For the fourth signal set, if the frequency domain features deviate from the preset threshold, adjust the abnormal area by using linear interpolation method to obtain a fifth signal set. According to the fifth signal set, the Kalman filter is used for secondary optimization of the time domain data to obtain a sixth signal set. Extract phase information from the sixth signal set, and if the phase difference exceeds the preset range, adjust it by using the phase correction algorithm to obtain a seventh signal set. Through the seventh signal set, the signal is finally smoothed by using mean filtering to obtain an eighth signal set.

[0055] Specifically, time domain data is obtained from the second signal set, and the signal is smoothed by using mean filtering. This process aims to reduce the small fluctuations in the signal.

[0056] For example, in audio signal processing, the collection device may introduce high-frequency noise due to environmental jitter. Mean filtering can effectively weaken such interference and make the waveform smoother, which is suitable for subsequent analysis. Fast Fourier transform is commonly used to convert time domain signals to frequency domain views.

[0057] Specifically, assuming that the sampling rate is 44.1 kHz, the spectrum of 0-10 kHz can be observed after transformation, and the main frequency peak position and amplitude information are extracted to provide a basis for signal consistency judgment. For the fourth signal set, if the frequency domain features deviate from the preset threshold, linear interpolation method can be used to adjust the abnormal area.

[0058] For example, if the peak value of a certain frequency band exceeds 10%, it may be caused by interference between devices. By taking the values before and after the abnormal point to smooth the transition, for example, interpolating the abnormal segment of 20ms, the naturalness of the frequency spectrum can be restored. This method is simple and efficient, and is particularly suitable for real-time processing scenarios.

[0059] It should be noted that the adjusted signal is more consistent with the expected model, which is convenient for subsequent optimization. According to the fifth signal set, the Kalman filter is used for secondary optimization of the time domain data.

[0060] In one possible implementation, the signal may be affected by dynamic noise, such as random fluctuations caused by wind noise in a microphone array. The Kalman filter gradually corrects the signal trajectory to make it closer to the true value through prediction and update steps, combining historical data and current observations.

[0061] Preferably, this method is particularly effective for non-stationary signals, and can improve the stability of the data. Phase information is extracted from the sixth signal set, and if the phase difference exceeds the preset range, the phase correction algorithm can be used.

[0062] For example, if a channel phase is 20 degrees ahead, the offset can be calculated by arctangent and adjusted so that the phase difference of each channel is controlled within 5 degrees.

[0063] It can be understood that phase consistency is crucial for the synthesis of multi-channel signals and directly affects the output quality.

[0064] In one embodiment, if the speech signal is processed, the superimposed corrected signal is more natural and reduces the reverberation effect. Through the seventh signal set, mean filtering is used for final smoothing processing.

[0065] For example, calculating the mean value with a 7-point sliding window can further eliminate residual jitter and make the waveform more readable.

[0066] Specifically, this processing is particularly useful in visual analysis, and the smoothed signal is easier to extract features.

[0067] In one possible implementation, if the number of channels increases to 6, the mean filtering can still maintain low computational cost while improving the overall signal quality.

[0068] Exemplarily, this smoothing process lays the foundation for subsequent high-precision applications, especially in scenarios that require long-term stable output. The above process is progressive, from time domain smoothing to frequency domain adjustment, and then to phase optimization, each step is around signal quality improvement.

[0069] For example, in a single scene of a microphone array, whether it is to suppress noise or to correct phase, the final result can make the signal more suitable for downstream task processing.

[0070] It can be understood that this method has strong adaptability in practical application, and the technical effect is also more stable due to the mutual support between steps.

[0071] S103, obtain terrain data, obstacle data and meteorological information flow in the geographic information source, convert heterogeneous data into a unified format through a data preprocessing module, adopt a standardized mapping method for data heterogeneity problems, and obtain a consistent first geographic data set.

[0072] The terrain data, obstacle data and weather information are obtained from geographic information sources, the heterogeneous data are converted by a data preprocessing module, a standardized mapping method is used to obtain a consistent first geographic data set. The spatial distribution characteristics of the terrain data and the obstacle data are extracted from the first geographic data set by the data preprocessing module, and a grid division method is used to obtain a second geographic data set. According to the spatial distribution characteristics in the second geographic data set, if the height difference between the terrain data and the obstacle data exceeds a preset threshold, the abnormal area is adjusted by a linear interpolation method to obtain a third geographic data set. The dynamic change trend of the weather information is obtained from the third geographic data set, and the mean filter is used to smooth the time series data to obtain a fourth geographic data set. For the fourth geographic data set, if the time series fluctuation of the weather information exceeds a preset range, the data is optimized and adjusted by Kalman filtering to obtain a fifth geographic data set. The correlation characteristics between the terrain data and the weather information are extracted from the fifth geographic data set, and the data is grouped by clustering analysis method to obtain a sixth geographic data set. According to the grouping result of the sixth geographic data set, the matching degree of the obstacle data and the weather information is judged, and if the matching degree is lower than a preset threshold, the data is adjusted by data fusion method to obtain a seventh geographic data set.

[0073] Specifically, when obtaining the terrain data, obstacle data and weather information from the geographic information sources.

[0074] Exemplarily, the terrain data such as mountain height and river direction can be extracted from satellite remote sensing map, the obstacle data such as building distribution and road network can be obtained from city planning database, and the real-time weather information such as rainfall and wind speed can be obtained from weather station.

[0075] In a possible implementation, the data preprocessing module will convert these heterogeneous data into a unified format, such as converting the terrain height unit to meters and the wind speed to meters per second, and through the standardized mapping method, the data range is normalized to 0 to 1 to form the first geographic data set. This helps to eliminate the dimensional difference in subsequent analysis. When extracting the spatial distribution characteristics by the data preprocessing module.

[0076] Specifically, the slope change in the terrain data and the density distribution of the obstacle data can be taken as key indicators. A grid division method is used, for example, the area is divided into 10 meters by 10 meters grid units, each unit is labeled with average height and obstacle number to form the second geographic data set. This division can more carefully reflect the spatial heterogeneity. If the height difference exceeds the preset threshold, such as the height difference of adjacent grids exceeds 50 meters, the abnormal area is adjusted by the linear interpolation method.

[0077] For example, at the junction of a valley and a ridge, the height mutation can be smoothed after interpolation to form a third geographical data set, which helps to reduce data noise interference. When obtaining the dynamic change trend of meteorological information from the third geographical data set.

[0078] It can be understood that the rainfall may present periodic fluctuations over time. The time series data is smoothed by mean filtering, such as calculating the average rainfall with a 5-hour window to form a fourth geographical data set, which can effectively weaken the short-time abnormal fluctuations. If the fluctuation of meteorological information exceeds the preset range, for example, the wind speed suddenly increases from 5 meters per second to 15 meters per second within 1 hour, then it is adjusted by Kalman filtering optimization. This method generates a more stable fifth geographical data set by combining prediction and observation, enhancing the reliability of the data. When extracting the correlation features of terrain data and meteorological information.

[0079] In an embodiment, the correlation between mountain height and rainfall can be analyzed, such as the rainfall may increase by 10 millimeters for every 100 meters of height increase. Using clustering analysis method, the data is divided into highland heavy rainfall, lowland weak rainfall, etc. groups to form a sixth geographical data set. This grouping can clearly reveal the influence of terrain on weather. When judging the matching degree of obstacle data and meteorological information.

[0080] Preferably, it can be checked whether the building dense area coincides with the low wind speed area. If the matching degree is lower than the preset threshold, such as the correlation coefficient is less than 0.6, then it is adjusted by data fusion method, for example, the wind speed data is weighted and fused with the building height to form a seventh geographical data set. This adjustment can improve the coordination between data and provide more accurate basis for subsequent application.

[0081] For example, in a mountainous scene at the edge of a city, satellite data shows that the height of a certain grid is 300 meters, the number of buildings is 5, and the rainfall is 20 millimeters per hour. After grid division, it is found that the height of the adjacent grid is only 50 meters, which is smoothed to a gradual height after interpolation adjustment. After mean filtering processing of the rainfall time series, the fluctuation is reduced from ±5 millimeters to ±2 millimeters, and Kalman filtering further optimizes the wind speed data, making it more consistent with the terrain distribution. After clustering analysis, the highland group is associated with high heavy rainfall, and after fusion adjustment, the wind speed data of the building dense area is more consistent with the actual situation. This multi-step processing can provide more reliable data support for signal analysis in complex terrain areas.

[0082] It should be noted that the application of linear interpolation and Kalman filtering can smooth data anomalies and improve prediction accuracy, while clustering analysis and data fusion enhance the integration capability of multi-source data. These methods support each other in a single scene to ensure data consistency and practicality.

[0083] S104, according to the first geographic data set, using fast Fourier transform algorithm to analyze the terrain complexity characteristics, combined with real-time requirements, using parallel computing framework to speed up the processing, get the second geographic data set containing terrain features.

[0084] The terrain complexity information is obtained from the first geographic data set, and the fast Fourier transform algorithm is used for processing to obtain the second geographic data set. The terrain features in the second geographic data set are processed by a parallel computing framework to obtain a third geographic data set. For the terrain features in the third geographic data set, a clustering analysis method is used to group the data to obtain a fourth geographic data set. According to the grouping result of the fourth geographic data set, the related features of terrain complexity and real-time requirements are extracted to obtain a fifth geographic data set. If the distribution of terrain features in the fifth geographic data set exceeds a preset threshold, a mean filter is used for smoothing to obtain a sixth geographic data set. The matching features of terrain complexity and computing framework are obtained from the sixth geographic data set, and a data fusion method is used for adjustment to obtain a seventh geographic data set. Through the seventh geographic data set, the correlation degree of terrain features and real-time requirements is determined to obtain the final processing result.

[0085] Specifically, when obtaining terrain complexity information from the first geographic data set.

[0086] It can be understood that the terrain complexity generally reflects the diversity and frequency of the surface relief.

[0087] For example, in a mountainous area, the terrain complexity may be characterized by frequent and large amplitude changes in elevation, while in a plain area it is relatively flat. When using the fast Fourier transform algorithm, the core is to convert the terrain data from the spatial domain to the frequency domain to extract high and low frequency features.

[0088] For example, in a 1000m x 1000m area, assuming that the terrain elevation data is sampled every 10m, the fast Fourier transform can identify sharp peaks with high frequency and gentle slope areas with low frequency, thereby generating a second geographic data set. This method can quickly separate the main fluctuation mode of the terrain, facilitating subsequent analysis. When the second geographic data set is processed by a parallel computing framework.

[0089] It should be noted that parallel computing can significantly improve the processing efficiency of large-scale data.

[0090] In particular, on a terrain dataset containing 100,000 sampling points, traditional serial computation can take several minutes, while with the aid of parallel framework, such as GPU-based computation, the processing time can be reduced to several seconds. The resulting third geographic dataset preserves the details of the terrain features while improving the real-time performance, laying the foundation for subsequent grouping. For the terrain features in the third geographic dataset, when using clustering analysis method for grouping, one possible implementation is to divide the terrain features into several categories according to complexity through K-means algorithm.

[0091] For example, in an area with large terrain undulations, it can be divided into three groups of high complexity, medium complexity and low complexity. Assuming that the standard deviation of elevation change greater than 50 meters is classified as high complexity, 20 to 50 meters is classified as medium complexity, and less than 20 meters is classified as low complexity. The resulting fourth geographic dataset can clearly reflect the spatial distribution characteristics of the terrain features. According to the grouping results of the fourth geographic dataset, when extracting the features related to terrain complexity and real-time requirements.

[0092] In one embodiment, the demand of different complexity areas on computing resources can be analyzed.

[0093] Preferably, high complexity areas can require more intensive sampling and faster processing speed.

[0094] For example, the real-time requirement of a high complexity mountainous area can be to complete feature extraction within 1 second, while a low complexity plain can be relaxed to 5 seconds. The fifth geographic dataset can thus reflect the priority differences of different areas. If the distribution of terrain features in the fifth geographic dataset exceeds the preset threshold, such as a standard deviation of more than 60 meters, mean filtering is used for smoothing.

[0095] Illustratively, for an area containing steep cliffs, the original data can show elevation mutations, and through mean filtering, the abnormal peaks can be smoothed into more continuous slope changes to generate the sixth geographic dataset. This processing helps to reduce noise interference and improve data usability. When obtaining the matching features of terrain complexity and computing framework from the sixth geographic dataset, data fusion methods can be used for adjustment.

[0096] For example, in a high complexity area, if the processing capacity of the computing framework is insufficient, it can cause delay. At this time, additional elevation smoothing data can be fused to optimize resource allocation, and finally the seventh geographic dataset is obtained. This adjustment can better match the actual computing demand. When judging the correlation between terrain features and real-time requirements through the seventh geographic dataset.

[0097] In one embodiment, it can be determined whether the processing delay of the high complexity area is less than 2 seconds. If the requirement is met, it indicates that the terrain features are highly matched with the real-time performance.

[0098] For example, a real-time navigation system can quickly plan a path to avoid computational bottlenecks caused by complex terrain. Such analysis can provide key evidence for system optimization.

[0099] S105, extracting obstacle data from the second geographic data set, generating a preliminary three-dimensional model by voxel grid method, superimposing volume interpolation algorithm for the spatial distribution characteristics of the obstacle data, obtaining the first optimized three-dimensional model.

[0100] The obstacle data is extracted from the second geographic data set, and the initial three-dimensional model is generated by using the voxel grid method to obtain the first three-dimensional model. For the spatial distribution characteristics in the first three-dimensional model, the volume interpolation algorithm is superimposed for processing to obtain the second three-dimensional model. The distribution characteristic data is obtained from the second three-dimensional model, and the obstacle distribution information is separated by the data extraction method to obtain the first distribution data set. If the distribution characteristics in the first distribution data set exceed the preset threshold, the mean filter is used for smoothing processing to obtain the second distribution data set. According to the obstacle distribution information in the second distribution data set, the data is grouped by using the clustering analysis method to obtain the first grouping data set. The matching characteristics of the spatial distribution and the obstacle data are obtained through the grouping results in the first grouping data set to obtain the first matching data set. The matching characteristics in the first matching data set are adjusted by using the data fusion method to obtain the second matching data set.

[0101] Specifically, when the obstacle data is extracted from the second geographic data set.

[0102] It can be understood that the obstacles can include terrain elements such as vegetation, buildings or rocks.

[0103] For example, in a mountainous scene, the second geographic data set can contain elevation information and land cover data. By analyzing the elevation mutation points and cover types, the rock distribution area can be identified.

[0104] In one embodiment, if the elevation change rate of a certain area exceeds 20%, and the cover type shows bare rock, it is marked as an obstacle, and the related coordinates and range are extracted to form the initial obstacle data. When the first three-dimensional model is generated by using the voxel grid method.

[0105] It should be noted that the voxel grid divides the space into regular cubic units, which is convenient for representing complex terrain.

[0106] Specifically, the space can be divided according to the resolution of 1 meter x 1 meter x 1 meter.

[0107] Exemplarily, if a mountain slope region contains 1000 voxels, and 300 of which are marked as obstacles, an initial three-dimensional model can be constructed by connecting these voxels, intuitively reflecting the obstacle distribution. A second three-dimensional model is generated by superimposing the first three-dimensional model with a stereovision interpolation algorithm.

[0108] It can be understood that stereovision interpolation infers the characteristics of unknown regions through existing data points.

[0109] For example, in a sparse vegetation area, the first three-dimensional model can only mark the positions of trees, and after interpolation, a continuous vegetation density distribution map can be generated.

[0110] Preferably, if the voxel spacing in a certain region is large, interpolation can fill the gaps, making the second three-dimensional model closer to the real terrain. When obtaining distribution characteristic data from the second three-dimensional model and separating obstacle information.

[0111] In one possible implementation, it can be completed by threshold screening. It is assumed that the region with a vegetation density greater than 0.8 is considered as a dense obstacle area, and after extraction, a first distribution data set is formed.

[0112] For example, in a certain forest area, voxels with a density higher than 0.8 account for 40%, and after separation, the obstacle and non-obstacle regions can be clearly divided. If the first distribution data set exceeds the preset threshold, smoothing processing is required.

[0113] Specifically, mean filtering can reduce the influence of noise.

[0114] Exemplarily, if the density of a certain obstacle edge region changes dramatically, exceeding the threshold 0.9, the mean value of the surrounding 5x5 voxels is taken to generate a second distribution data set, making the distribution smoother and more coherent, which is helpful for subsequent analysis. When performing clustering analysis according to the second distribution data set.

[0115] In one embodiment, the K-means method can be used for grouping.

[0116] For example, the obstacle distribution is divided into high, medium and low three categories according to the density to obtain a first grouping data set. It is assumed that the proportion of high-density obstacles in a certain region is 30%, and the proportion of medium-density obstacles is 50%, and after grouping, the characteristics of different terrain regions can be reflected. When obtaining matching features through the first grouping data set.

[0117] In one possible implementation, the correlation between obstacles and spatial distribution can be analyzed.

[0118] For example, in a certain valley, low-density obstacles are concentrated in flat areas, and high-density obstacles are located in steep slopes, generating a first matching data set to reveal the distribution rule. When generating a second matching data set according to the first matching data set, the data fusion method can integrate multi-source information.

[0119] Preferably, if the elevation data and the density data are inconsistent, the adjustment can be made by weighted fusion.

[0120] For example, the elevation weight is set to 0.6 and the density weight is set to 0.4. The second matching data set after fusion can more accurately reflect the matching relationship between the terrain and the obstacles, which helps to improve the accuracy of spatial analysis.

[0121] S106, obtain real-time meteorological information flow, calculate the influence of meteorology on signal transmission through fluid dynamics simulation algorithm, and obtain the second three-dimensional model matched with the current environment by using time series prediction method for dynamic changes of meteorological information flow.

[0122] Obtain real-time meteorological information flow, collect dynamic change data through sensor network, obtain first meteorological data set. Process the first meteorological data set through fluid dynamics simulation algorithm, calculate the signal transmission parameter, obtain the first influence data set. For the dynamic change characteristics in the first influence data set, use time series prediction method for analysis, obtain the first prediction data set. According to the prediction result in the first prediction data set, match the current environment parameter, obtain the second three-dimensional model. If the signal transmission parameter in the second three-dimensional model exceeds the preset threshold, process the first influence data set through mean filtering to obtain the second influence data set. Use clustering analysis method to group the data in the second influence data set to obtain the first grouped data set. Adjust the spatial distribution characteristics of the second three-dimensional model through the grouping result in the first grouped data set to obtain the third three-dimensional model.

[0123] Specifically, when obtaining real-time meteorological information flow, dynamic change data can be collected through sensor network.

[0124] For example, a plurality of meteorological sensors arranged in the upper air of a city record temperature, humidity, wind speed and other parameters every second to form a first meteorological data set.

[0125] For example, these data may show that the temperature in a certain area fluctuates between 25 and 30 degrees, and the wind speed is 5 meters per second.

[0126] In one possible implementation, the sensor network covers an area of 10 square kilometers, and the collection frequency is 1 time per minute, ensuring the real-time nature of the data. When processing the first meteorological data set through fluid dynamics simulation algorithm, the signal transmission parameter can be calculated.

[0127] Specifically, assuming that wind speed and humidity affect signal attenuation, the simulation result may show that the signal attenuation rate in a certain area is 2 decibels per kilometer, obtaining the first influence data set.

[0128] It should be noted that the fluid dynamics simulation takes into account the air flow characteristics, such as the deflection effect of wind direction on the signal path. This method can help analyze the impact of weather conditions on transmission. For the dynamic change characteristics in the first influence data set, a time series prediction method is used for analysis.

[0129] For example, based on the wind speed data of the past 1 hour, it is predicted that the wind speed may rise to 7 meters per second in the next 30 minutes, generating a first prediction data set.

[0130] In an embodiment, time series prediction can predict weather changes in advance through historical trend and periodic fluctuation analysis. This prediction helps to optimize subsequent model adjustment. According to the prediction results in the first prediction data set, match the current environmental parameters to generate a second three-dimensional model.

[0131] Preferably, the environmental parameters include terrain height and vegetation density, and after combining the predicted wind speed, the model may show that the signal coverage range of a certain area is reduced by 10%.

[0132] It can be understood that this matching can reflect the interaction between weather and space, providing dynamic basis for the model. If the signal transmission parameters in the second three-dimensional model exceed the preset threshold, such as the attenuation rate exceeding 3 decibels per kilometer, the first influence data set is processed through mean filtering.

[0133] For example, the attenuation rate of a certain point in the original data is 4 decibels per kilometer, and after filtering, it is reduced to 2.5 decibels per kilometer, generating a second influence data set.

[0134] Specifically, mean filtering smooths out outliers, improving data stability and helping the accuracy of subsequent analysis. When grouping data in the second influence data set using clustering analysis, it can be divided into 3 groups according to the attenuation rate.

[0135] For example, the low attenuation group is 0-1 decibel per kilometer, the medium attenuation group is 1-2 decibels per kilometer, and the high attenuation group is 2 decibels per kilometer or more, obtaining a first grouped data set.

[0136] In a possible implementation, the clustering result reflects the difference in signal characteristics of different regions, facilitating targeted optimization. Through the grouping results in the first grouped data set, the spatial distribution characteristics of the second three-dimensional model are adjusted to generate a third three-dimensional model.

[0137] For example, the area corresponding to the high attenuation group may be marked as a signal blind area, and the model will adjust the coverage range to reduce the blind area by 15%.

[0138] Exemplarily, this adjustment can improve the adaptability of the model to complex environments and ensure more reasonable signal distribution. This method optimizes spatial characteristics through data-driven, which has high practical value.

[0139] S107, according to the second three-dimensional model and the second signal set, a particle filter algorithm is used to realize signal and geographic information fusion processing, model parameters are adjusted to optimize signal coverage range for low-altitude navigation requirements, and a third signal set is obtained.

[0140] The feature data in the signal set is extracted by the second model to obtain a first feature set. According to the first feature set and the geographic information, a particle filter algorithm is used for fusion processing to obtain a first fusion set. According to the data distribution characteristics in the first fusion set, the parameters related to low-altitude navigation are adjusted to obtain a first optimization set. The spatial distribution information of the enhanced signal is obtained from the first optimization set, and the boundary value of the coverage range is determined. If the boundary value exceeds the preset threshold, the first optimization set is processed by mean filtering to obtain a second optimization set. According to the data characteristics in the second optimization set, a clustering analysis method is used for grouping to obtain a first grouping set. The spatial distribution of the enhanced signal is adjusted by the first grouping set to obtain a third signal set.

[0141] Specifically, the feature data in the signal set is extracted by the second model to obtain a first feature set, which aims to extract key information from complex signal data.

[0142] For example, in a low-altitude navigation scenario, the signal set may contain raw data such as electromagnetic wave intensity, delay time and multipath effect. After feature extraction, the first feature set may include the peak intensity of the signal and the direction angle of the main propagation path.

[0143] Specifically, the waveform of the signal can be decomposed into multiple feature components by time domain analysis method, such as extracting the maximum amplitude value of every 10 milliseconds, such as 5 volts, as one of the features. According to the first feature set and the geographic information, a particle filter algorithm is used for fusion processing to obtain a first fusion set, which is the key to combining signal features and environmental elements. Particle filtering optimizes data fusion results by simulating multiple possible state distributions.

[0144] For example, in a mountainous environment, the geographic information may include a terrain fluctuation of 500 meters in height, and the signal direction angle in the first feature set is 30 degrees. Particle filtering can generate a set of distributed particles based on this information to estimate the real propagation path of the signal in this area, and finally fuse the first fusion set, which may show that the optimal estimated intensity of the signal at a certain point is 4.8 volts. According to the data distribution characteristics in the first fusion set, the parameters related to low-altitude navigation are adjusted to obtain a first optimization set, which is adjusted to adapt to the actual application requirements.

[0145] In a possible implementation, the first fusion set can show that the signal strength is low in some areas, such as below 3 volts, at which point the parameters can be optimized by increasing the transmission power or adjusting the antenna angle to form a first optimization set, to ensure more uniform signal coverage and effectively improve the navigation accuracy. The spatial distribution information of the enhanced signal is obtained from the first optimization set, and the boundary value of the coverage range is determined, and this process focuses on the effective range of the signal.

[0146] It can be understood that the optimized signal can maintain a strength higher than 4 volts within a horizontal distance of 10 kilometers, and the boundary value can be set to 3.5 volts. If the threshold is exceeded, it indicates that the coverage is insufficient.

[0147] For example, in the urban fringe area, the signal strength can be reduced to below 3 volts due to building shielding, exceeding the threshold. If the boundary value exceeds the preset threshold, the first optimization set is processed by mean filtering to obtain a second optimization set, and this method is used to smooth the data fluctuations.

[0148] Preferably, if the signal strength in a certain area frequently jumps between 3 and 4 volts, the mean filtering can take the average of the previous and next 5 sampling points, such as smoothing 3.2, 3.8, and 3.5 volts to 3.5 volts, to form a more stable second optimization set and improve signal consistency. According to the data characteristics in the second optimization set, a clustering analysis method is used for grouping to obtain a first grouping set, which helps to identify the pattern of signal distribution.

[0149] In an embodiment, the second optimization set can include multiple signal strength intervals, such as 3.5 to 4 volts and 4 to 4.5 volts, which can be divided into two groups by clustering analysis. The first grouping set can show that the low-intensity group is concentrated in the urban fringe, and the high-intensity group is close to the base station, which is helpful for subsequent optimization layout. The spatial distribution of the enhanced signal is adjusted by the first grouping set to obtain a third signal set, which further improves the signal coverage.

[0150] For example, for the low-intensity group, a relay station can be added to increase the signal strength from 3.5 volts to more than 4 volts, and finally the spatial distribution of the third signal set is more balanced, ensuring the stable operation of the low-altitude navigation equipment. This way significantly improves the reliability and coverage range of the signal.

[0151] Embodiment 2:

[0152] The application also provides a multi-channel-based low-altitude helicopter communication and navigation signal enhancement system, comprising:

[0153] The first processing module is configured to obtain a first signal set with synchronization and low interference from raw signal data collected by the multi-channel technology.

[0154] The second processing module is configured to extract feature parameters from the first signal set, perform preliminary enhancement processing on the signal by using a Kalman filtering algorithm, superimpose a preset attenuation compensation model for multipath effects caused by terrain complexity, and obtain an enhanced second signal set.

[0155] The third processing module is configured to obtain a unified first geographic data set according to terrain data, obstacle data and meteorological information flow in a geographic information source.

[0156] The fourth processing module is configured to obtain a second geographic data set containing terrain features according to the first geographic data set.

[0157] The fifth processing module is configured to extract obstacle data from the second geographic data set, generate a preliminary three-dimensional model by using a voxel grid method, superimpose a volume interpolation algorithm for spatial distribution characteristics of the obstacle data, and obtain an optimized first three-dimensional model.

[0158] The sixth processing module is configured to obtain a real-time meteorological information flow and obtain a second three-dimensional model matched with a current environment.

[0159] The seventh processing module is configured to perform fusion processing of signals and geographic information by using a particle filtering algorithm according to the second three-dimensional model and the second signal set, adjust model parameters to optimize signal coverage range for low-altitude navigation requirements, and obtain a fusion-enhanced third signal set.

[0160] As an embodiment of the present application, the first processing module obtains original signal data collected by a multi-channel technology, performs time alignment on channel data by using a preset time domain synchronization algorithm, suppresses cross interference between channels by using an adaptive filter for signal synchronization requirements, and obtains a first signal set that is synchronized and has low interference.

[0161] As an embodiment of the present application, the third processing module obtains terrain data, obstacle data and meteorological information flow in a geographic information source, converts heterogeneous data into a unified format by using a data preprocessing module, and obtains a unified first geographic data set by using a standardized mapping method for data heterogeneity problems.

[0162] As an embodiment of the present application, the fourth processing module analyzes terrain complexity characteristics by using a fast Fourier transform algorithm according to the first geographic data set, combines real-time requirements, and uses a parallel computing framework to speed up processing, and obtains a second geographic data set containing terrain features.

[0163] As an embodiment of the present application, the sixth processing module obtains a real-time meteorological information flow, calculates the influence of meteorology on signal transmission by using a fluid dynamics simulation algorithm, and obtains a second three-dimensional model matched with a current environment by using a time series prediction method for dynamic changes of the meteorological information flow.

[0164] The above description is merely that of a plurality of preferred embodiments of the present disclosure, and is not intended to limit the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall fall within the scope of protection of the present disclosure.

Claims

1. A method for enhancing communication and navigation signals of low altitude helicopter based on multi-channel, characterized in that, The method comprises the following steps: According to the original signal data collected by the multi-channel technology, a first signal set with synchronization and low interference is obtained; From the first signal set, feature parameters are extracted, and a Kalman filter algorithm is used for preliminary enhancement processing of the signal. In view of the multipath effect caused by the complexity of the terrain, a preset attenuation compensation model is superimposed to obtain a second signal set after enhancement; According to the terrain data, obstacle data and meteorological information flow in the geographic information source, a first geographic data set is obtained; According to the first geographic data set, a second geographic data set containing terrain features is obtained; From the second geographic data set, obstacle data is extracted, and a preliminary three-dimensional model is generated by the voxel grid method. In view of the spatial distribution characteristics of the obstacle data, a volume interpolation algorithm is superimposed to obtain an optimized first three-dimensional model; Real-time meteorological information flow is obtained to obtain a second three-dimensional model matched with the current environment; According to the second three-dimensional model and the second signal set, a particle filter algorithm is used to realize the fusion processing of the signal and the geographic information. In view of the low-altitude navigation requirement, the model parameters are adjusted to optimize the signal coverage range, and a third signal set after fusion enhancement is obtained; The original signal data collected by the multi-channel technology is obtained, and the time alignment of each channel data is performed by using a preset time domain synchronization algorithm. In view of the signal synchronization requirement, an adaptive filter is used to suppress the cross interference between channels, and a first signal set with synchronization and low interference is obtained; The terrain data, obstacle data and meteorological information flow in the geographic information source are obtained, and the heterogeneous data is converted into a unified format by using a data preprocessing module. In view of the data heterogeneity problem, a standardized mapping method is used to obtain a first geographic data set. According to the first geographic data set, the terrain complexity characteristics are analyzed by using a fast Fourier transform algorithm, and a parallel computing framework is used for acceleration processing in combination with the real-time requirement, so as to obtain a second geographic data set containing terrain features; Real-time meteorological information flow is obtained, and the influence of meteorology on signal transmission is calculated by using a fluid dynamics simulation algorithm. In view of the dynamic change of the meteorological information flow, a time series prediction method is used to obtain a second three-dimensional model matched with the current environment.

2. A multi-channel based low altitude helicopter communication navigation signal enhancement system characterized in that, The method comprises the following steps: The first processing module is used for obtaining a first signal set with synchronization and low interference according to the original signal data collected by the multi-channel technology; The second processing module is used for extracting feature parameters from the first signal set, and a Kalman filter algorithm is used for preliminary enhancement processing of the signal. In view of the multipath effect caused by the complexity of the terrain, a preset attenuation compensation model is superimposed to obtain a second signal set after enhancement; The third processing module is used for obtaining a first geographic data set according to the terrain data, obstacle data and meteorological information flow in the geographic information source; The fourth processing module is used for obtaining a second geographic data set containing terrain features according to the first geographic data set; The fifth processing module is used for extracting obstacle data from the second geographic data set, and a preliminary three-dimensional model is generated by the voxel grid method. In view of the spatial distribution characteristics of the obstacle data, a volume interpolation algorithm is superimposed to obtain an optimized first three-dimensional model; The sixth processing module is used for obtaining a second three-dimensional model matched with the current environment by using real-time meteorological information flow; The seventh processing module is used for realizing the fusion processing of the signal and the geographic information according to the second three-dimensional model and the second signal set. In view of the low-altitude navigation requirement, the model parameters are adjusted to optimize the signal coverage range, and a third signal set after fusion enhancement is obtained. The seventh processing module is configured to implement signal and geographic information fusion processing by using a particle filter algorithm according to the second three-dimensional model and the second signal set, adjust model parameters to optimize signal coverage range for low-altitude navigation requirements, and obtain a third signal set enhanced by fusion; The first processing module acquires original signal data collected by multi-channel technology, performs time alignment on channel data by using a preset time domain synchronization algorithm, suppresses cross interference between channels by using an adaptive filter for signal synchronization requirements, and obtains a first signal set that is synchronized and has low interference; The third processing module acquires terrain data, obstacle data and meteorological information flow in a geographic information source, converts heterogeneous data into a unified format by using a data preprocessing module, uses a standardization mapping method for data heterogeneity problems, and obtains a first geographic data set that is consistent; The fourth processing module analyzes terrain complexity characteristics by using a fast Fourier transform algorithm according to the first geographic data set, uses a parallel computing framework to speed up processing for real-time requirements, and obtains a second geographic data set containing terrain characteristics; The sixth processing module acquires real-time meteorological information flow, calculates the influence of meteorology on signal transmission by using a fluid dynamics simulation algorithm, uses a time series prediction method for dynamic changes of the meteorological information flow, and obtains a second three-dimensional model matched with the current environment.

Images