Indoor wireless signal rapid coverage method and system

Abstract

This invention belongs to the field of mobile communication technology and discloses a method and system for rapid indoor wireless signal coverage. The method includes: calculating the power deviation index of M mobile radio frequency units per unit time based on the collected reference signal received power distribution; obtaining the spatial characteristics of the combination of each mobile radio frequency unit and power level based on the area function and power level; performing feature engineering operations on the monitoring data and time attribute set collected by each mobile radio frequency unit per unit time, as well as the power level, power deviation index and spatial characteristics, to obtain the historical feature data of each mobile radio frequency unit per unit time; and defining a preset window length, setting training data, and training the prediction model of each mobile radio frequency unit based on the historical feature data of each mobile radio frequency unit per unit time. This invention significantly improves the accuracy of predicting the transmit power level of mobile radio frequency units and achieves the best balance between service quality and energy consumption.

Description

Technical Field

[0001] This invention relates to the field of mobile communication technology, and more specifically, to a method and system for rapid indoor wireless signal coverage. Background Technology

[0002] Semi-enclosed spaces, such as underground parking garages with reinforced concrete and metal structures, severely attenuate and repeatedly reflect radio frequency signals, creating coverage blind spots. Macro base station signals can hardly penetrate these areas, necessitating the deployment of distributed antenna systems (DAS) to improve indoor coverage. However, traditional threshold-triggered energy-saving strategies (such as low-load power reduction and sleep mode) struggle to balance energy efficiency and service quality due to static configuration: they consume significant standby power during low traffic periods, while high traffic periods suffer from interruptions and reduced throughput due to wake-up delays. While digital DAS offers the advantage of flexible programmability, it lacks accurate prediction and intelligent resource management, hindering its full potential.

[0003] With the advancement of 5G / 6G technologies, network energy efficiency is just as important as spectrum efficiency. In wireless access networks, the power consumption of base stations or mobile radio frequency units (RF units) is dominant. In underground parking lot scenarios, most mobile RF units are in a low-load or idle state for extended periods. To implement green communications and reduce operating costs, a strategy for dynamically adjusting the state of mobile RF units is urgently needed.

[0004] In view of this, the present invention proposes a method and system for rapid indoor wireless signal coverage to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and achieve the above objectives, the present invention provides the following technical solution: a method for rapid indoor wireless signal coverage, comprising:

[0006] Based on the power distribution of the acquired reference signal, calculate the power deviation index of M mobile radio frequency units per unit time.

[0007] Based on the area function and power level, the spatial characteristics of the combination of each mobile radio frequency unit and power level are obtained;

[0008] For each mobile radio frequency unit, the monitoring data and time attribute set collected per unit time, as well as the power level, power deviation index and spatial characteristics, are subjected to feature engineering operations to obtain the historical feature data of each mobile radio frequency unit per unit time.

[0009] Based on the historical feature data of each mobile radio frequency unit per unit time, a preset window length is defined, training data is set, and a prediction model for each mobile radio frequency unit is trained.

[0010] The prediction results are generated based on the prediction model of each mobile radio frequency unit and the historical characteristic data of each unit time. Based on the prediction results, the power level of each mobile radio frequency unit is adjusted in advance for the next unit time.

[0011] Furthermore, methods for constructing the area function include:

[0012] For each combination of mobile RF unit and power level, n measurement points are generated throughout the underground parking lot using a pseudo-random uniform sampling algorithm. The received power of the reference signal at these n measurement points is measured using a handheld drive test instrument or portable spectrum analyzer, and outliers are removed through statistical filtering to obtain the received power of the reference signal at n′ preprocessed locations. The signal strength level is calculated based on the received power of the reference signal to obtain the signal strength level at the n′ preprocessed locations. The effective strength levels of the preprocessed locations are retained to obtain the effective strength levels of the n” effective signal locations. The functional type of the spatial area where the n” effective signal locations are located is obtained. For each combination of effective strength level and functional type, the number of effective signal locations is counted to obtain the equivalent strength and similarity quantity for each combination of mobile RF unit, power level, effective strength level, and functional type. The total area of ​​the underground parking lot is calculated, and the average area of ​​each measurement point is obtained by dividing the total area by n. The equivalent strength and similarity quantity is multiplied by the average area to obtain the equivalent strength and similarity area for each combination of mobile RF unit, power level, effective strength level, and functional type. The area function is obtained by using the mobile RF unit, power level, effective strength level, and functional type as independent variables and the equivalent strength and similarity area as the dependent variable.

[0013] Furthermore, the method for constructing power levels includes: dividing the acquired signal transmission power into intervals to obtain power levels;

[0014] The method for constructing signal strength levels includes: dividing the received power of the acquired reference signal into intervals to obtain power levels;

[0015] The effective strength level is the signal strength level that is higher than the preset strength threshold.

[0016] Furthermore, methods for dividing spatial regions include:

[0017] Create a digital map of the underground parking lot, dividing it into different spatial areas according to function type, and marking the function type of each spatial area; the function types include passage area, charging area, and parking area.

[0018] Furthermore, methods for obtaining spatial features include:

[0019] For each combination of mobile radio frequency unit, power level, and effective strength level, the area of ​​the same type with equal strength is obtained by the area function; the sum of the areas of the same type with equal strength corresponding to all function types is calculated to obtain the total area of ​​equal strength; the total area of ​​equal strength and the areas of the same type with equal strength corresponding to all function types are written into column vectors according to a preset order to obtain the vector of equal strength corresponding to each combination of mobile radio frequency unit, power level, and effective strength level; for each combination of mobile radio frequency unit and power level, the vector of equal strength corresponding to each effective strength level is written into matrix form according to a preset order to obtain the spatial characteristics corresponding to each combination of mobile radio frequency unit and power level.

[0020] Furthermore, the monitoring data includes: uplink throughput, downlink throughput, physical resource block utilization, and user capacity.

[0021] Furthermore, the methods for acquiring signal transmission power, monitoring data, and reference signal received power distribution include:

[0022] Through the network management system of the digital distributed antenna system, at the beginning of each unit of time, the real-time data of the signal transmission power, monitoring data, and reference signal reception power distribution of each mobile radio frequency unit are read and the timestamps of each data are recorded.

[0023] Furthermore, the calculation method for the power deviation index includes:

[0024] For each mobile radio frequency unit (RF unit), the reference signal received power value at a preset percentile is calculated based on the reference signal received power distribution per unit time to obtain the cell signal strength per unit time. When the cell signal strength is greater than or equal to the upper limit of the preset quality level, the power deviation index of the mobile RF unit per unit time is -1; when the cell signal strength is less than the lower limit of the preset quality level, the power deviation index of the mobile RF unit per unit time is 1; when the real-time signal strength is within the range of the preset quality level, the power deviation index of the mobile RF unit per unit time is 0.

[0025] Furthermore, methods for adjusting power levels in advance include:

[0026] For each mobile radio frequency unit (RF unit), a preset number of time units, including the current time unit, are used as a preset window length. Historical feature data within this preset window length are extracted and input into the prediction model to obtain the predicted power level and power deviation index for each RF unit in the next time unit. These are labeled as the predicted power level and the prediction deviation index, respectively. When the prediction deviation index is -1, the actual power level configured for the corresponding RF unit in the next time unit is equal to the power level one level lower than the predicted power level. When the prediction deviation index is 1, the actual power level configured for the corresponding RF unit in the next time unit is equal to the power level one level higher than the predicted power level. When the prediction deviation index is 0, the actual power level configured for the corresponding RF unit in the next time unit is equal to the predicted power level.

[0027] An indoor wireless signal rapid coverage system is provided to implement a method for rapid indoor wireless signal coverage. The system includes:

[0028] The first processing module calculates the power deviation index of M mobile radio frequency units per unit time based on the power distribution of the collected reference signal.

[0029] The second processing module obtains the spatial characteristics of the combination of each mobile radio frequency unit and power level based on the area function;

[0030] The feature engineering module performs feature engineering operations on the monitoring data and time attribute set collected per unit time for each mobile radio frequency unit, as well as the power level, power deviation index and spatial characteristics, to obtain the historical feature data of each mobile radio frequency unit per unit time.

[0031] The training module, based on the historical feature data of each mobile radio frequency unit per unit time, defines a preset window length, sets training data, and trains the prediction model for each mobile radio frequency unit.

[0032] The control module predicts the results based on the prediction model of each mobile radio frequency unit and the historical characteristic data of each unit time, and adjusts the power level of each mobile radio frequency unit in advance for the next unit time based on the prediction results.

[0033] The beneficial effects of the indoor wireless signal rapid coverage method and system of the present invention are as follows:

[0034] By refining the spatial division of underground parking lots into functional areas such as passageways, charging areas, and parking areas, and combining this with quantitative analysis of the signal coverage area of ​​mobile radio frequency units at different power levels, the accuracy of predicting the transmission power level of mobile radio frequency units has been significantly improved.

[0035] Pseudo-random sampling and Z-score anomaly filtering are used to rigorously screen drive test data, ensuring the reliability of the coverage area function and eliminating error accumulation and outlier interference.

[0036] By using historical feature data to train a prediction model, the system can accurately predict future loads and intelligently adjust the transmission power of mobile radio frequency units in advance, achieving the best balance between service quality and energy consumption, and avoiding the energy waste caused by traditional static threshold strategies.

[0037] Taking into account spatial characteristics (coverage area of ​​each functional area), time attributes (weekdays / holidays, peak and off-peak periods) and real-time signal indicators (transmit power, throughput, physical resource block utilization, user capacity, and reference signal received power distribution), the power configuration is dynamically evaluated and adjusted in real time. This not only effectively fills blind spots but also flexibly responds to sudden traffic fluctuations, thereby significantly improving user experience and system robustness. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of a module of an indoor wireless signal rapid coverage system according to the present invention;

[0039] Figure 2 This is a flowchart of a method for rapid indoor wireless signal coverage according to the present invention;

[0040] Figure 3 This is a schematic diagram of a scenario for an indoor wireless signal rapid coverage system according to the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] Example 1

[0043] See Figure 1 The indoor wireless signal rapid coverage system described in this embodiment includes: a first processing module, a second processing module, a feature engineering module, a training module, and a control module.

[0044] A digital distributed antenna system (DAS) consists of two main parts: a baseband processing unit (BBU) and a remote radio unit (RRU) (see [reference]). Figure 3The Remote Radio Unit (RRU), also known as the Mobile Radio Unit (MRU), is responsible for digitizing analog radio frequency signals from the operator's core network (including demodulation, channel coding / decoding, and multi-carrier multiplexing) to generate a unified digital baseband data stream. This digital baseband data stream is then distributed to multiple mobile radio units. Simultaneously, it receives the uplink digital baseband data streams from each mobile radio unit, processes them in reverse, and sends them back to the operator's core network. Mobile radio units are deployed in the service area (e.g., the various zones of the underground parking lot in this system). Their main function is to convert between analog radio frequency signals and digital baseband data streams. Specifically, the mobile radio unit receives the digital baseband data stream from the BBU and converts it back to analog radio frequency signals using a DAC (digital-to-analog converter). Simultaneously, the mobile radio unit receives uplink analog radio frequency signals transmitted by the end user equipment (UE), amplifies them with low noise, performs an ADC (analog-to-digital converter), and packages them into a digital baseband data stream before sending it back to the BBU. This system ignores the transmission delay between the BBU and the mobile radio units (they are connected via a high-speed, low-latency method, such as fiber optic cable). Digital distributed antenna systems are existing technology and will not be discussed further here.

[0045] Signal transmission power directly affects the coverage and penetration capability of mobile radio frequency unit signals. Uplink throughput measures a user's ability to initiate upload services (such as cloud storage synchronization, live video streaming, or large file uploads), reflecting the uplink capacity of the network. For video conferencing or real-time monitoring, insufficient uplink throughput can lead to video stuttering and intermittent audio. Downlink throughput directly reflects the user's experience quality for downlink services (such as webpage loading, video-on-demand, and file downloads). High downlink throughput supports smooth 4K / 8K video playback and fast large file downloads. Insufficient downlink throughput can cause video buffering and slow webpage response. Physical resource block (PRB) utilization measures the intensity of cell wireless resource usage. High PRB utilization indicates resource scarcity, which may lead to queuing or transmission delays. When PRB utilization consistently exceeds 85%, the average throughput for users will significantly decrease, impairing the service experience. By monitoring PRB utilization, scheduling algorithms can be dynamically adjusted or capacity can be expanded to ensure service quality during peak periods. User capacity directly reflects the cell load level and is a key indicator for determining whether capacity expansion or resource optimization is needed. More users mean that the total bandwidth is distributed among more terminals, naturally reducing the throughput available to a single user. Reference Received Power (RSRP) is defined as the average power of a specific reference signal received by a user equipment from the serving cell. The higher the RSRP value, the stronger the signal received by the user equipment. The RSRP distribution is a cumulative distribution function of the RSRP values ​​reported by user equipment within the coverage area of ​​a mobile radio unit. The RSRP distribution can be used to calculate the average, median (50th percentile), standard deviation, minimum, maximum, and various percentiles (such as the 5th, 10th, and 90th percentiles).

[0046] The monitoring data includes: uplink throughput, downlink throughput, physical resource block utilization, and user capacity.

[0047] The methods for acquiring signal transmit power, monitoring data, and reference signal receive power distribution include: using the network management system (NMS) of a digital distributed antenna system (such as Huawei iMasterNCE-Campus), at the beginning of each unit of time (e.g., every 60 seconds), reading the real-time data of signal transmit power, monitoring data, and reference signal receive power distribution of each mobile radio frequency unit and recording the timestamp of each data item.

[0048] The first processing module calculates the power deviation index of the M mobile radio frequency units per unit time based on the power distribution of the collected reference signal.

[0049] The calculation method for the power deviation index includes: for each mobile radio frequency unit, the reference signal received power value at a preset percentile is calculated based on the reference signal received power distribution per unit time to obtain the cell signal strength per unit time. When the cell signal strength is greater than or equal to the upper limit of the preset quality level, the power deviation index of the mobile radio frequency unit per unit time is -1; when the cell signal strength is less than the lower limit of the preset quality level, the power deviation index of the mobile radio frequency unit per unit time is 1; when the real-time signal strength is within the range of the preset quality level, the power deviation index of the mobile radio frequency unit per unit time is 0.

[0050] For example, when the preset percentile is the fifth percentile (ensuring that the signal quality of most users is not lower than the preset quality level), the preset quality level is signal availability, and the reference signal received power value of the fifth percentile of the reference signal received power distribution of a certain mobile radio frequency unit in a certain unit of time is -105dBm, -90dBm is the upper limit of the preset quality level, -100dBm is the lower limit of the preset quality level, the cell signal strength of a certain mobile radio frequency unit in a certain unit of time is -105dBm (5% of user equipment have a reference signal received power value less than -105dBm), the power deviation index of a certain mobile radio frequency unit in a certain unit of time is 1 (indicating that the power level of a certain mobile radio frequency unit in a certain unit of time needs to be improved).

[0051] The power deviation index is a key data indicator for achieving intelligent control and energy efficiency optimization. It not only accurately quantifies the degree to which the current cell signal strength conforms to the preset quality level, providing an indispensable objective basis for the subsequent feature engineering module to construct high-quality historical feature data, but it is also a crucial component of the prediction model's learning and prediction target (i.e., the prediction deviation index). Through the precise calculation and application of the power deviation index and the prediction deviation index, the control module can transcend traditional threshold judgments and achieve refined and proactive adjustments to the power level of the mobile radio frequency unit—moderately reducing power when the prediction deviation index is -1, increasing power when it is 1, and maintaining the optimized state when it is 0. This closed-loop feedback and predictive control mechanism based on the power deviation index is a fundamental solution to the problem of insufficient dynamic adjustment capabilities and the inability to balance service quality and energy consumption in existing technologies. It directly contributes to the achievement of the key beneficial effect of "significantly improving the accuracy of mobile radio frequency unit transmit power level prediction and achieving the optimal balance between service quality and energy consumption."

[0052] The second processing module obtains the spatial characteristics of the combination of each mobile radio frequency unit and power level based on the area function.

[0053] The method for constructing power levels includes dividing the acquired signal transmission power into intervals to obtain power levels.

[0054] Power levels include: Deep Sleep, Light Sleep, Active Low Power, and Active Full Power. Deep Sleep consumes the lowest power (in the milliwatt range) and is used for scenarios where service is not needed for extended periods. Light Sleep consumes less power (in the tens to hundreds of milliwatts range) than Active Low Power but more power than Deep Sleep; it is used for scenarios where there is short periods of no use but a fast response is required. Active Low Power consumes a medium power level (30% to 50% of full power) and is used for light-load periods with only a few users. Active Full Power consumes the highest power (tens to hundreds of watts) and is used to ensure maximum coverage and throughput during peak traffic periods. In practical applications, signal transmission power levels can be divided in 3dBm intervals to achieve finer power adjustment (signal transmission power is not adjusted continuously but discretely according to power levels).

[0055] The method for constructing signal strength levels includes dividing the received power of the acquired reference signal into intervals to obtain power levels.

[0056] Signal strength levels include: excellent signal, good signal, usable signal, poor signal, and unusable signal. In this embodiment, RSRP ≥ -80dBm is defined as "excellent signal", -90 ≤ RSRP < -80dBm is defined as "good signal", -100 ≤ RSRP < -90dBm is defined as "usable signal", -110 ≤ RSRP < -100dBm is defined as "poor signal", and RSSI < -110dBm is defined as "unusable signal".

[0057] The effective strength level is the signal strength level that is higher than a preset strength threshold. For example, when the preset strength threshold is "signal unavailable," all preprocessed locations with a signal strength level of "signal unavailable" will be excluded. The preset strength threshold is used to distinguish between effective coverage locations and insufficient coverage locations. The effective strength level excludes insufficient coverage locations, which can significantly reduce the range of the area function's domain and reduce storage space.

[0058] The functional types include: passageway area, charging area, and parking area. It is worth noting that in actual applications, the functional types in this embodiment can be further subdivided; for example, the passageway area can be subdivided into entrance passageway area, exit passageway area, etc., and the parking area can be subdivided into first-floor parking area, second-floor parking area, etc.

[0059] The method for dividing the space includes: creating a digital map of the underground parking lot, dividing the underground parking lot into different spatial areas according to the function type, and marking the function type of each spatial area.

[0060] The system meticulously divides service environments such as underground parking lots into spatial areas based on functional types (e.g., "passage area," "charging area," "parking area"), and labels the functional type of each spatial area. This constitutes a key link in achieving accurate prediction and intelligent resource management. This functional type-based spatial area division allows the system to transcend the limitations of traditional methods that rely solely on network management system data, gaining a deeper understanding of the unique user behavior and data demand patterns within different functional areas. The resulting area function, which includes the coverage of each functional type, and further extracted spatial features, provide the feature engineering module with refined spatial context information reflecting the differences in real-world business scenarios. Integrating these spatial features rich in regional functional attributes into historical feature data and using them to train the prediction model is one of the core elements that enables the system to significantly improve the accuracy of mobile radio frequency unit transmit power level prediction. It is this meticulous consideration and utilization of spatial areas and their functional types that allows power regulation decisions to fully take into account the differentiated needs of different areas, thereby effectively supporting the ultimate achievement of the beneficial effect of "optimal balance between service quality and energy consumption," ensuring intelligent and efficient wireless signal coverage in complex indoor environments.

[0061] The construction method of the area function includes: for each combination of a mobile radio unit and a power level, n measurement point positions are generated throughout the underground parking lot through a pseudo-random uniform sampling algorithm (ensuring the spatial representativeness of the measurement positions), and a hand-held road test instrument (such as Huawei's GENEX Probe) or a portable spectrum analyzer (such as Keysight's FieldFox series) is used to measure the reference signal received power at the n measurement point positions, and outliers are removed through statistical filtering (such as using the Z-score method and setting that when the absolute value of the Z-score is greater than 3, it is an outlier), obtaining the reference signal received power at n' (n' < n) preprocessed positions; calculating the signal strength level based on the reference signal received power, obtaining the signal strength levels at the n' preprocessed positions; retaining the preprocessed positions with valid strength levels, obtaining the valid strength levels at n'' (n'' < n') valid signal positions; obtaining the functional type of the spatial region where the n'' valid signal positions are located; for each combination of a valid strength level and a functional type, counting the number of valid signal positions, obtaining the equal-strength and same-type quantity for each combination of a mobile radio unit, a power level, a valid strength level, and a functional type (for example, for a combination of a certain mobile radio unit and low-power activation, if the number of valid signal positions belonging to the charging area with excellent signal is 15, then the corresponding equal-strength and same-type quantity is equal to 15); calculating the total area of the underground parking lot, dividing the total area by n to obtain the average area of each measurement point position; multiplying the equal-strength and same-type quantity by the average area to obtain the equal-strength and same-type area for each combination of a mobile radio unit, a power level, a valid strength level, and a functional type; using the mobile radio unit, the power level, the valid strength level, and the functional type as independent variables and the equal-strength and same-type area as the dependent variable to obtain the area function.

[0062] The area function plays a crucial role in accurately quantifying and correlating the power level of a mobile radio frequency unit (RF unit) with its actual signal coverage performance. Through systematic pseudo-random uniform sampling measurement and statistical filtering of the reference signal received power, and combined with the functional type of the spatial area, the area function establishes a mapping relationship between the mobile RF unit, power level, effective strength level, and functional type to areas of equal strength and type. This refined, multi-dimensional modeling of coverage provides a solid and reliable foundation for subsequently extracting spatial features that truly reflect the coverage characteristics of areas with different functional types. It is precisely because of these accurate spatial features derived from reliable area functions that the prediction model of this system can more accurately learn and predict the signal coverage requirements and corresponding power level configurations under specific functional types, thus directly supporting a fundamental improvement over existing technologies that suffer from inaccurate predictions and inability to effectively adapt to the dynamic needs of complex indoor environments. Therefore, the construction and application of the area function not only ensures the quality of spatial features and eliminates error accumulation and outlier interference, but also significantly improves the accuracy of predicting the transmit power level of the mobile RF unit, providing a key technical guarantee for ultimately achieving the beneficial effect of "optimal balance between service quality and energy consumption."

[0063] The method for obtaining spatial features includes: for each combination of mobile radio frequency unit, power level, and effective strength level, obtaining the area of ​​the same type with equal strength corresponding to each functional type using an area function; calculating the sum of the areas of the same type with equal strength corresponding to all functional types to obtain the total area of ​​equal strength; writing the total area of ​​equal strength and the areas of the same type with equal strength corresponding to all functional types into column vectors according to a preset order to obtain the equal strength vector corresponding to each combination of mobile radio frequency unit, power level, and effective strength level (for example, using the total area of ​​equal strength as the first element of the equal strength vector, the area of ​​the same type with equal strength in the channel area as the second element, the area of ​​the same type with equal strength in the charging area as the third element, and the area of ​​the same type with equal strength in the parking area as the fourth element); for each combination of mobile radio frequency unit and power level, writing the equal strength vector corresponding to each effective strength level into a matrix according to a preset order to obtain the spatial features corresponding to each combination of mobile radio frequency unit and power level.

[0064] For example, if the preset strength threshold is that the signal is unavailable, and the optimal signal strength vector corresponding to a combination of a mobile radio frequency unit and a power level is a column vector (4.0, 1.4, 1.0, 1.6), then... T The equal-strength vector with good signal strength is the column vector (6.2, 1.5, 2.4, 2.3). T The available equal-intensity vector for the signal is the column vector (5.5, 2.2, 3.3, 0). T The vector with the weaker signal strength is the column vector (8.9, 2.1, 2.8, 4.0). TThen the spatial characteristics are matrices.

[0065]

[0066] Spatial features are a key data structure carrying detailed information on the coverage performance of mobile radio frequency units (RF units) across different functional areas at specific power levels. Originating from a deep analysis of area functions, spatial features systematically organize information such as the equivalent area and total equivalent area for each functional type into a matrix-like form for different effective power levels. This provides the feature engineering module with quantified and structured, refined spatial context information. These spatial features, which precisely describe the specific relationship between coverage and functional type, enable the prediction model to transcend single signal or load indicators during the learning process, deeply understanding the actual impact of different power level configurations on the coverage of various functional areas (such as "channel areas," "charging areas," and "parking areas"). Therefore, the introduction and application of spatial features play a decisive role in solving the bottleneck problem of inaccurate transmission power level prediction and difficulty in dynamically adapting to the differentiated needs of various functional areas in existing technologies under complex indoor environments. It is an indispensable technical support for achieving a significant improvement in the accuracy of mobile RF unit transmission power level prediction and ultimately achieving the core beneficial effect of "optimal balance between service quality and energy consumption."

[0067] The feature engineering module performs feature engineering operations on the monitoring data and time attribute set collected per unit time for each mobile radio frequency unit, as well as the power level, power deviation index and spatial characteristics, to obtain the historical feature data of each mobile radio frequency unit per unit time.

[0068] The methods for obtaining the time attribute set include: for each unit of time, obtaining all time attributes to form the time attribute set for each unit of time.

[0069] Time attributes include: work attributes and peak / valley attributes.

[0070] The methods for obtaining work attributes include: obtaining the unit time to be processed, extracting the corresponding date information, and determining whether the extracted date belongs to a working day or a non-working day (weekend / statutory holiday) by calling the system calendar service or the built-in holiday table; if it is a working day, the work attribute is 1; if it is a non-working day, the work attribute is 0.

[0071] The method for obtaining peak and valley attributes includes: obtaining the unit time to be processed, extracting the corresponding hour and minute, and comparing it with preset time period thresholds for morning peak (e.g., 07:00–09:00), midday peak (e.g., 11:30–13:30), evening peak (e.g., 17:00–19:00), and nighttime trough (e.g., 23:00–05:00 the next day); if it falls within the morning peak time period threshold, the peak and valley attribute is 0; if it falls within the midday peak time period threshold, the peak and valley attribute is 1; if it falls within the evening peak time period threshold, the peak and valley attribute is 2; if it falls within the nighttime trough time period threshold, the peak and valley attribute is 3.

[0072] Feature engineering operations include: cleaning, encoding, standardization, and unifying the timestamp format.

[0073] The training module, based on the historical feature data of each mobile radio frequency unit per unit time, defines a preset window length, sets training data, and trains the prediction model for each mobile radio frequency unit.

[0074] Use a preset number of unit times, including the current unit time, as the preset window length.

[0075] The method for setting up training data includes: setting a preset window length as y, taking the historical feature data of y units from the kth to the k+y-1th unit as input samples, taking the power level and power deviation index of the k+yth day as output samples, and combining the input samples and output samples to form a set of training data.

[0076] The training method for the prediction model includes: for each mobile radio frequency unit, dividing the historical feature data per unit time into a training set and a test set (the ratio can be set to 7:3), using the training set to train the selected machine learning model, and using the test set to test the selected machine learning model. Convergence occurs when the prediction accuracy reaches a preset prediction accuracy. Finally, a prediction model capable of predicting power levels and power deviation exponents is obtained. The machine learning model can be a Long Short-Term Memory (LSTM) network model.

[0077] The control module predicts the results based on the prediction model of each mobile radio frequency unit and the historical characteristic data of each unit time, and adjusts the power level of each mobile radio frequency unit in advance for the next unit time based on the prediction results.

[0078] The method for advance adjustment of power levels includes: for each mobile radio frequency unit, extracting historical feature data of a preset window length and inputting it into the prediction model to obtain the predicted values ​​of the power level and power deviation index for each mobile radio frequency unit in the next unit time, which are marked as the predicted power level and the prediction deviation index, respectively; when the prediction deviation index is -1, the actual value of the power level configured for the corresponding mobile radio frequency unit in the next unit time is equal to the power level one level lower than the predicted power level; when the prediction deviation index is 1, the actual value of the power level configured for the corresponding mobile radio frequency unit in the next unit time is equal to the power level one level higher than the predicted power level; when the prediction deviation index is 0, the actual value of the power level configured for the corresponding mobile radio frequency unit in the next unit time is equal to the predicted power level.

[0079] This embodiment proposes an indoor wireless signal rapid coverage system, which effectively solves the problem in the prior art that it is impossible to predict and adjust the transmission power level of each mobile radio frequency unit in real time.

[0080] Specifically, this system is implemented through the following methods:

[0081] The system collects real-time data on the signal transmission power, monitoring data, and reference signal received power distribution of each mobile radio frequency unit; it divides the spatial areas of the underground parking lot according to functional types (channel area, charging area, parking area); it constructs an area function using pseudo-random sampling and drive test data to quantify the signal coverage area of ​​each functional type of spatial area under different power levels; it calculates the power deviation index based on the reference signal received power distribution to determine whether the transmission power needs to be adjusted; and it combines time attributes (weekdays / holidays, peak and off-peak periods) and spatial characteristics (a functional area coverage matrix) to train a prediction model using historical data through a sliding window method, thereby achieving dynamic adaptive adjustment of the transmission power level of the mobile radio frequency unit.

[0082] Compared with existing technologies, this system has significant advantages:

[0083] The underground parking lot is spatially divided into functional areas such as passageways, charging areas, and parking areas. The prediction accuracy is improved by combining the coverage range of mobile radio frequency unit signals at each power level across these functional areas (quantized using an area function). Through pseudo-random sampling and Z-score anomaly filtering, the measured values ​​of the drive-test signals are rigorously screened to ensure the accuracy of the area function and effectively avoid error accumulation and outlier interference.

[0084] Based on historical feature data, a prediction model is trained through a sliding window to achieve accurate prediction of future load and adjust the transmit power level of the mobile radio frequency unit in advance to balance service quality and energy consumption, thereby reducing energy waste caused by static threshold strategies.

[0085] By combining the coverage area (spatial characteristics), time attributes (weekdays / holidays, peak and off-peak periods) and real-time signal indicators (transmit power, throughput, physical resource block utilization, user capacity, and reference signal received power distribution), the power configuration is evaluated and dynamically adjusted in real time. This not only effectively fills coverage blind spots but also flexibly responds to sudden load peaks or off-peak fluctuations, significantly improving user experience and system robustness.

[0086] Example 2

[0087] See Figure 2 As shown, this embodiment provides a method for rapid indoor wireless signal coverage, including:

[0088] Based on the power distribution of the acquired reference signal, calculate the power deviation index of M mobile radio frequency units per unit time.

[0089] Based on the area function and power level, the spatial characteristics of the combination of each mobile radio frequency unit and power level are obtained;

[0090] For each mobile radio frequency unit, the monitoring data and time attribute set collected per unit time, as well as the power level, power deviation index and spatial characteristics, are subjected to feature engineering operations to obtain the historical feature data of each mobile radio frequency unit per unit time.

[0091] Based on the historical feature data of each mobile radio frequency unit per unit time, a preset window length is defined, training data is set, and a prediction model for each mobile radio frequency unit is trained.

[0092] The prediction results are generated based on the prediction model of each mobile radio frequency unit and the historical characteristic data of each unit time. Based on the prediction results, the power level of each mobile radio frequency unit is adjusted in advance for the next unit time.

[0093] For any parts not mentioned in this application, existing technologies may be used or referenced.

[0094] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0095] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for rapid indoor wireless signal coverage, characterized in that, include: Based on the power distribution of the acquired reference signal, calculate the power deviation index of M mobile radio frequency units per unit time. Based on the area function and power level, the spatial characteristics of the combination of each mobile radio frequency unit and power level are obtained; For each mobile radio frequency unit, the monitoring data and time attribute set collected per unit time, as well as the power level, power deviation index and spatial characteristics, are subjected to feature engineering operations to obtain the historical feature data of each mobile radio frequency unit per unit time. Based on the historical feature data of each mobile radio frequency unit per unit time, a preset window length is defined, training data is set, and a prediction model for each mobile radio frequency unit is trained. The prediction results are generated based on the prediction model of each mobile radio frequency unit and the historical characteristic data of each unit time. Based on the prediction results, the power level of each mobile radio frequency unit is adjusted in advance for the next unit time.

2. The method for rapid indoor wireless signal coverage according to claim 1, characterized in that, Methods for constructing area functions include: For each combination of mobile RF unit and power level, n measurement points are generated throughout the underground parking lot using a pseudo-random uniform sampling algorithm. The received power of the reference signal at these n measurement points is measured using a handheld drive test instrument or portable spectrum analyzer, and outliers are removed through statistical filtering to obtain the received power of the reference signal at n′ preprocessed locations. The signal strength level is calculated based on the received power of the reference signal to obtain the signal strength level at the n′ preprocessed locations. The effective strength levels of the preprocessed locations are retained to obtain the effective strength levels of the n” effective signal locations. The functional type of the spatial area where the n” effective signal locations are located is obtained. For each combination of effective strength level and functional type, the number of effective signal locations is counted to obtain the equivalent strength and similarity quantity for each combination of mobile RF unit, power level, effective strength level, and functional type. The total area of ​​the underground parking lot is calculated, and the average area of ​​each measurement point is obtained by dividing the total area by n. The equivalent strength and similarity quantity is multiplied by the average area to obtain the equivalent strength and similarity area for each combination of mobile RF unit, power level, effective strength level, and functional type. The area function is obtained by using the mobile RF unit, power level, effective strength level, and functional type as independent variables and the equivalent strength and similarity area as the dependent variable.

3. The method for rapid indoor wireless signal coverage according to claim 2, characterized in that, The method for constructing power levels includes: dividing the acquired signal transmission power into intervals to obtain power levels; The method for constructing signal strength levels includes: dividing the received power of the acquired reference signal into intervals to obtain power levels; The effective strength level is the signal strength level that is higher than the preset strength threshold.

4. The method for rapid indoor wireless signal coverage according to claim 2, characterized in that, Methods for dividing spatial regions include: Create a digital map of the underground parking lot, dividing it into different spatial areas according to function type, and marking the function type of each spatial area; the function types include passage area, charging area, and parking area.

5. The method for rapid indoor wireless signal coverage according to claim 1, characterized in that, Methods for obtaining spatial features include: For each combination of mobile radio frequency unit, power level, and effective strength level, the area of ​​the same type with equal strength is obtained by the area function; the sum of the areas of the same type with equal strength corresponding to all function types is calculated to obtain the total area of ​​equal strength; the total area of ​​equal strength and the areas of the same type with equal strength corresponding to all function types are written into column vectors according to a preset order to obtain the vector of equal strength corresponding to each combination of mobile radio frequency unit, power level, and effective strength level; for each combination of mobile radio frequency unit and power level, the vector of equal strength corresponding to each effective strength level is written into matrix form according to a preset order to obtain the spatial characteristics corresponding to each combination of mobile radio frequency unit and power level.

6. The method for rapid indoor wireless signal coverage according to claim 1, characterized in that, The monitoring data includes: uplink throughput, downlink throughput, physical resource block utilization, and user capacity.

7. The method for rapid indoor wireless signal coverage according to claim 1, characterized in that, The methods for acquiring signal transmission power, monitoring data, and reference signal received power distribution include: Through the network management system of the digital distributed antenna system, at the beginning of each unit of time, the real-time data of the signal transmission power, monitoring data, and reference signal reception power distribution of each mobile radio frequency unit are read and the timestamps of each data are recorded.

8. The method for rapid indoor wireless signal coverage according to claim 1, characterized in that, The calculation methods for the power deviation index include: For each mobile radio frequency unit, the reference signal received power value at a preset percentile is calculated based on the reference signal received power distribution per unit time to obtain the cell signal strength per unit time. When the cell signal strength is greater than or equal to the upper limit of the preset quality level, the power deviation index of the mobile radio frequency unit per unit time is -1. When the cell signal strength is less than the lower limit of the preset quality level, the power deviation index of the mobile radio frequency unit per unit time is 1. When the real-time signal strength is within the range of the preset quality level, the power deviation index of the mobile radio frequency unit per unit time is 0.

9. A method for rapid indoor wireless signal coverage according to claim 1, characterized in that, Methods for adjusting power levels in advance include: For each mobile radio frequency unit (RF unit), a preset number of time units, including the current time unit, are used as a preset window length. Historical feature data within this preset window length are extracted and input into the prediction model to obtain the predicted power level and power deviation index for each RF unit in the next time unit. These are labeled as the predicted power level and the prediction deviation index, respectively. When the prediction deviation index is -1, the actual power level configured for the corresponding RF unit in the next time unit is equal to the power level one level lower than the predicted power level. When the prediction deviation index is 1, the actual power level configured for the corresponding RF unit in the next time unit is equal to the power level one level higher than the predicted power level. When the prediction deviation index is 0, the actual power level configured for the corresponding RF unit in the next time unit is equal to the predicted power level.

10. An indoor wireless signal rapid coverage system, characterized in that, The system for implementing the indoor wireless signal rapid coverage method according to claim 1 includes: The first processing module calculates the power deviation index of M mobile radio frequency units per unit time based on the power distribution of the collected reference signal. The second processing module obtains the spatial characteristics of the combination of each mobile radio frequency unit and power level based on the area function; The feature engineering module performs feature engineering operations on the monitoring data and time attribute set collected per unit time for each mobile radio frequency unit, as well as the power level, power deviation index and spatial characteristics, to obtain the historical feature data of each mobile radio frequency unit per unit time. The training module, based on the historical feature data of each mobile radio frequency unit per unit time, defines a preset window length, sets training data, and trains the prediction model for each mobile radio frequency unit. The control module predicts the results based on the prediction model of each mobile radio frequency unit and the historical characteristic data of each unit time, and adjusts the power level of each mobile radio frequency unit in advance for the next unit time based on the prediction results.

Images