Spatial data intelligent analysis system based on sojourn and old-age care residence

Abstract

The invention relates to the technical field of sojourn and old-age care management, and discloses an intelligent analysis system based on sojourn and old-age care residence space data. The system comprises a space multi-source sensing module, a space efficiency dynamic analysis module, a space resource configuration module and a space risk early warning module. The spatial multi-source sensing module collects physical environment, facility use state and resident behavior characteristic parameters in real time; after receiving the data, the space efficiency dynamic analysis module extracts space occupancy mode characteristics according to an evaluation period, and generates a space use frequency, a facility adaptation degree and a behavior safety evaluation value; the space resource configuration module calculates a space optimization demand index and outputs a space layout and facility configuration adjustment instruction; the space risk early warning module continuously monitors space structure stability, environment safety and abnormal behavior parameters and generates multi-level safety early warning signals. According to the system, intelligent management of the sojourn and old-age care residence space is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of travel and retirement management, in particular to a travel and retirement residential space data intelligent analysis system. BACKGROUND

[0002] With the acceleration of population aging, travel and retirement as a new model integrating retirement and tourism is gradually favored by the elderly. Travel and retirement housing, as the core carrier of this model, its spatial layout rationality, facility adaptation degree and safety level directly affect the living experience and quality of life of the elderly. However, there are still many deficiencies in the current travel and retirement housing in terms of space management and service optimization.

[0003] The existing residential space management relies on manual inspection and experience judgment, and it is difficult to grasp the physical environment changes in real time. Problems such as temperature and humidity fluctuations, insufficient light intensity, etc. are often ignored, which may cause the elderly to be unwell. In terms of facility use, there is a lack of effective monitoring of the use state of various facilities, such as the degree of wear and tear of fitness equipment, the condition of public rest area seats, etc., which cannot timely discover facility failures or aging problems, affecting the use experience and may also bring safety hazards.

[0004] The behavior characteristics of elderly residents are diverse and complex, and the traditional management method is difficult to accurately capture their activity rules, such as the gathering period in public space, the frequently used functional area, etc., resulting in unreasonable allocation of space resources, excessive congestion in some areas, and low utilization rate in some areas. In addition, the monitoring of space structure stability and environmental safety is mostly periodic inspection, which is difficult to discover sudden risks such as structure loosening and fire hazards in time, which poses a threat to the life safety of the elderly. SUMMARY

[0005] The purpose of the present application is to provide a travel and retirement residential space data intelligent analysis system to solve the problems raised in the background.

[0006] To achieve the above purpose, the present application provides a travel and retirement residential space data intelligent analysis system, which comprises: A space multi-source perception module for real-time collection of physical environment parameters, facility use state parameters and resident behavior characteristic parameters of travel and retirement housing; A space efficiency dynamic analysis module for receiving the collected data of the space multi-source perception module, setting an evaluation period, extracting features of space occupation mode in each evaluation period, and generating space use frequency evaluation value, facility adaptation degree evaluation value and behavior safety evaluation value; a space resource configuration module, configured to calculate a space optimization demand index in each evaluation period based on the space use frequency evaluation value, the facility adaptation evaluation value, and the behavior safety evaluation value, and output a space layout adjustment instruction and a facility configuration adjustment instruction; a space risk early warning module, configured to continuously monitor the space structure stability parameter, the environmental safety parameter, and the abnormal behavior parameter of the residential care home, and generate a multi-level space safety early warning signal.

[0007] Preferably, the physical environment parameters include illumination intensity distribution data, temperature and humidity distribution data, and air quality index data; the facility use state parameters include furniture use frequency data, electrical appliance operating state data, and barrier-free facility wear data; and the resident behavior characteristic parameters include activity trajectory data, stay duration data, and emergency call trigger data.

[0008] Preferably, the space efficiency dynamic analysis module extracts features of a space occupancy mode in each evaluation period, and specifically performs: dividing continuous time periods in the evaluation period, extracting a deviation amount of an actual occupancy duration of a space region in each continuous time period from a preset standard occupancy duration, and obtaining a space use frequency evaluation value; analyzing a matching degree of facility use frequency data in each continuous time period from a preset standard model for the elderly, and combining a degradation degree of barrier-free facility wear data, to obtain a facility adaptation evaluation value; counting an occurrence frequency of an abnormal activity trajectory and a distribution characteristic of emergency call trigger data in each continuous time period, to obtain a behavior safety evaluation value.

[0009] Preferably, when the space resource configuration module calculates the space optimization demand index, specifically performs: normalizing the space use frequency evaluation value, the facility adaptation evaluation value, and the behavior safety evaluation value; fusing the normalized space use frequency evaluation value, the facility adaptation evaluation value, and the behavior safety evaluation value by using a dynamic weighting algorithm, to output a space optimization demand index; when the space optimization demand index exceeds a preset demand threshold, activating the space layout adjustment instruction and the facility configuration adjustment instruction.

[0010] Preferably, a generation logic of the space layout adjustment instruction includes: identifying a region with a space use frequency evaluation value exceeding a first reference value, and marking the region as a high-frequency use region; identifying a region with a space use frequency evaluation value lower than a second reference value, and marking the region as an idle region; generating a furniture position resetting scheme as the space layout adjustment instruction according to a topological relationship between the high-frequency use region and the idle region.

[0011] Preferably, the generation logic of the facility configuration adjustment instruction comprises: When the facility adaptation evaluation value is lower than the preset adaptation threshold, the historical usage frequency data and the current wear data of the corresponding facility are called; Based on the facility degradation rate model, the remaining service life is predicted, and combined with the decline range of the behavior safety evaluation value, a facility replacement priority list is generated as the facility configuration adjustment instruction.

[0012] Preferably, the space resource configuration module further performs instruction feedback verification: After outputting the space layout adjustment instruction and the facility configuration adjustment instruction, the space multi-source perception data of the next evaluation period is re-collected; Based on the re-collected data, the space optimization demand index is updated, and it is verified whether the updated space optimization demand index is lower than the preset demand threshold; If it is not lower than the preset demand threshold, the parameter weight of the space layout adjustment instruction and the facility configuration adjustment instruction is corrected.

[0013] Preferably, when the space risk early warning module generates a multi-level space safety early warning signal, it specifically performs: Real-time comparison of the space structure stability parameter and the structure deformation safety threshold, if it exceeds, a first-level early warning signal is triggered; Continuously monitor the harmful gas concentration and fire risk factors in the environmental safety parameter, if it exceeds the corresponding threshold, a second-level early warning signal is triggered; Analyze the fall identification features and long-time stillness features in the abnormal behavior parameters, if it exceeds the behavior abnormality threshold, a third-level early warning signal is triggered.

[0014] Preferably, the space risk early warning module further comprises an early warning response mechanism: When the first-level early warning signal is triggered, the dangerous area is automatically locked and the structure reinforcing device is started; When the second-level early warning signal is triggered, the ventilation equipment and the fire-fighting equipment are linked to respond to the emergency; When the third-level early warning signal is triggered, the alarm information is pushed to the monitoring terminal in real time and the positioning tracking function is activated.

[0015] Preferably, the system further comprises a space data tracing module for storing the historical space optimization demand index, the space layout adjustment instruction execution record, the facility configuration adjustment instruction execution record and the multi-level space safety early warning signal trigger record, and generating a space use efficiency trend report.

[0016] Compared with the prior art, the present application has the following advantages: Through the collaborative work of multiple modules, it brings positive changes in many aspects for the management and service of the residential care for the elderly. The space multi-source perception module can collect physical environment parameters, facility usage state parameters and resident behavior characteristic parameters in real time, so that the management personnel can comprehensively and timely understand various situations in the residence, breaking the limitations of information lag and one-sidedness in traditional management.

[0017] The space performance dynamic analysis module sets an evaluation period and extracts space occupation mode characteristics based on the collected data, generates space usage frequency evaluation value, facility adaptation degree evaluation value and behavior safety evaluation value, providing quantitative reference basis for the management personnel, which helps to deeply understand the space usage law and facility matching degree, and potential safety problems in the behavior of the residents, so that the management decision is more targeted and scientific.

[0018] The space resource configuration module calculates the space optimization demand index according to the above evaluation values and outputs adjustment instructions, which can promote the space layout and facility configuration to be more reasonable. Through optimization, it can avoid the situation of resource waste in part of the area and resource shortage in part of the area, improve the space utilization rate and facility use efficiency, and enable the elderly residents to obtain more convenient and comfortable experience in the use process.

[0019] The space risk early warning module continuously monitors the stability of the space structure, the safety of the environment and the abnormal behavior and generates multi-level early warning signals, which can timely remind in the risk germination stage, so that the management personnel can quickly take measures to eliminate hidden dangers and reduce the possibility of safety accidents, providing more reliable protection for the life safety and physical health of the elderly residents. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The timing diagram of the intelligent analysis system based on space data of the residential care for the elderly described in the present application; Figure 2 The flowchart of space performance dynamic analysis; Figure 3 The flowchart of space optimization demand index calculation; Figure 4 The flowchart of facility configuration adjustment instruction generation; Figure 5 The flowchart of multi-level safety early warning signal generation. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0022] Referring to Figure 1 The application provides a smart analysis system based on residential space data for the elderly, which comprises: The space multi-source perception module collects physical environment parameters, facility usage state parameters and resident behavior characteristic parameters of the residential space for the elderly in real time. The space efficiency dynamic analysis module receives the collected data, sets an evaluation period, extracts features of space occupation modes in each evaluation period, and generates space usage frequency evaluation values, facility adaptation degree evaluation values and behavior safety evaluation values. The space resource configuration module calculates space optimization demand indexes of each evaluation period based on these evaluation values, and outputs space layout adjustment instructions and facility configuration adjustment instructions. The space risk early warning module continuously monitors space structure stability parameters, environmental safety parameters and abnormal behavior parameters, and generates multi-level space safety early warning signals. Each module works cooperatively, data is transmitted through an Internet of Things sensor network, distributed database storage is adopted, a data processing unit runs feature extraction algorithms and decision logic, and intelligent optimization and safety management of space are realized.

[0023] In Embodiment 1, the physical environment parameter collection of the space multi-source perception module is realized through a distributed sensor network. Light intensity distribution data is obtained by a light intensity sensor array, the sensors are installed at the junctions of the ceiling and the wall of the residence, the spacing is not more than 2 meters to form a coverage grid. The illuminance value is recorded every 5 minutes, the unit is lux (lx), and the data point contains a position coordinate and a time stamp. The temperature and humidity distribution data is obtained by a wall-embedded environment monitoring node, the node is deployed at the center points of the bedroom, living room, bathroom and corridor areas. The temperature and humidity readings are collected every 10 seconds, the temperature detection range is 0-50℃, the accuracy is ±0.5℃, the humidity detection range is 10-95% RH, and the accuracy is ±3% RH. The air quality index data is obtained by a multi-parameter gas sensor, which is fixed on the wall surface at a height of 1.5 meters from the ground. The PM2.5 (range 0-1000 μg / m 3 ), CO2 (range 400-5000 ppm) and TVOC (range 0-1187 ppb) concentration values are updated every 30 seconds, and the data packet is accompanied by a region number and a device ID code.

[0024] The furniture usage frequency data in the facility usage state parameter is collected by a pressure sensing system. A thin film pressure sensor is embedded in the sofa cushion, a piezoelectric array is arranged at the bottom of the mattress, and a strain gauge is installed on the chair leg. The pressure trigger threshold is set to 20 kg, and the start time, duration and pressure peak of each pressure are recorded. The appliance running state data is processed by an intelligent socket protocol analysis module, and the socket is provided with a current transformer and a voltage sampling circuit. The refrigerator running state records the compressor refrigeration cycle start interval, the air conditioner monitors the temperature deviation value, and the washing machine records the single running water consumption and vibration amplitude data. The contact type detection probe is installed on the inside of the stair handrail, the edge of the wheelchair ramp and the surface of the bathroom handrail. The probe scans at a distance of 15 minutes, measures the wear depth (accuracy 0.1 mm) and surface crack length, and generates a material degradation coefficient. All device state data is transmitted to the central gateway through the ZigBee protocol.

[0025] The resident behavior characteristic parameter collection relies on multi-modal perception technology. The activity trajectory data is realized by an ultra-wideband (UWB) positioning tag system, the tag is worn on the user's wrist, and the base station is deployed at the four corners of the space. The three-dimensional positioning engine outputs a set of coordinates (x, y, z) every second, with a positioning accuracy of ±15 cm. The stay duration data is processed by a trajectory point clustering algorithm, and when the spatial deviation of 20 consecutive trajectory points is less than 30 cm, it is determined as a stay state, and the time difference from the start to the end of the stay is calculated. The emergency call trigger data sources include two types: the infrared trigger signal of the one-key alarm button (response time <0.5 seconds) and the millimeter wave radar fall detection event. The radar is deployed in the center of the ceiling, and the fall posture is recognized by the Doppler feature change, and the body posture angle and impact acceleration are recorded when triggered.

[0026] The space data tracing module adopts a time series database architecture, and the database table structure is stored according to four types of data. The historical space optimization demand index record contains a timestamp (accurate to milliseconds), a region identifier and a normalized index value (floating point type). The space layout adjustment instruction execution record is stored in a structured log, and the fields include adjustment time, operation type (move / remove / add), furniture ID, original coordinates and target coordinates. The facility configuration adjustment instruction execution record contains device number, replacement priority score, operation suggestion (repair / replacement) and execution status flag. The multi-level space safety warning signal trigger record stores the event sequence, and each record includes a warning level code (1 / 2 / 3 level), a trigger position three-dimensional coordinate, a measured parameter overrun value and a standard threshold.

[0027] The storage mechanism of the data traceability module implements a hierarchical compression strategy: recent data (within 30 days) retains the original sampling frequency, medium-term data (30-180 days) is down-sampled to hourly mean storage, and long-term data (more than 180 days) is compressed into daily statistical feature values. The space performance trend report generation engine performs a four-step processing flow: monthly aggregation of historical space optimization demand index to form a line chart, with the horizontal axis representing the natural month sequence and the vertical axis representing the index fluctuation amplitude; spatial layout adjustment instruction record correlation area heat map visualizes furniture position change hot area; facility configuration record analysis into equipment replacement cycle prediction curve, marking the theoretical service life and actual replacement time node; warning signal record is converted into a three-dimensional space risk distribution model, with color depth identifying high-risk areas. The report is packaged in PDF format and automatically generated and written to an encrypted storage area every Monday at 00:00.

[0028] The sensor calibration mechanism ensures data reliability. The environmental monitoring node automatically performs zero-point calibration at 04:00 every day, and corrects drift error by referring to standard environmental parameters. The furniture pressure sensor performs a static load test once a month, with a 200kg weight load to verify linear error. The positioning base station performs time synchronization protocol (PTP protocol) every week, with clock difference controlled within nanoseconds. The traceability database performs off-site dual-active backup, with a 50km distance between master and slave nodes, and data consistency is maintained through blockchain hash verification mechanism. During report generation, all charts are labeled with data valid time range, and original sampling point data storage path is embedded as an appendix index in PDF document metadata.

[0029] Embodiment 2: refer to Figure 2 The space performance dynamic analysis module adopts a time window segmentation and multi-dimensional parameter fusion processing architecture when performing feature extraction during the evaluation period. The evaluation period is set to 24 hours, from 06:00 on the current day to 06:00 on the next day, divided into 6 consecutive time periods, each corresponding to a different typical activity pattern: morning activity period (06:00-10:00), morning active period (10:00-14:00), midday rest period (14:00-18:00), evening activity period (18:00-22:00), night rest period (22:00-02:00), and deep night silence period (02:00-06:00). Each time period is further subdivided into 15-minute analysis units, forming 96 basic time slices for micro-behavior modeling.

[0030] The calculation of the space usage frequency evaluation value is based on the dynamic comparison of the actual occupancy duration and the preset standard. The actual occupancy duration of each area (bedroom, living room, kitchen, etc.) is obtained by clustering the positioning system track points: when the user stays in the same area for more than 2 minutes, the timer starts, and the duration is recorded after leaving. The preset standard occupancy duration comes from the aging behavior model library, which contains the behavior pattern data of 12 typical old users. The deviation amount calculation uses the following formula:

[0031] wherein: represents the occupancy deviation rate of the i-th area in the j-th time slice, is the actual monitored occupancy minutes, is the model preset standard value. The deviation rates of all areas are weighted and averaged to convert the space usage frequency evaluation value to the range of 0-1, and the weight coefficient is dynamically adjusted according to the importance of the area function (bedroom 0.3, bathroom 0.25, living room 0.2, kitchen 0.15, corridor 0.1).

[0032] The generation of the facility adaptation degree evaluation value depends on the fusion analysis of dual data sources. The facility usage frequency data comes from the appliance operation log and furniture sensors. Taking the refrigerator as an example: record the number of door openings in a day (normal range 8-12 times) and the duration of each opening (standard value 15-30 seconds). The aging standard model establishes a facility usage feature matrix containing 32 threshold values. The matching degree calculation uses the sliding window correlation coefficient method: calculate the Pearson correlation coefficient between the actual usage curve and the model curve within a 7-day window, and the result is mapped to the 0-1 interval. The wear and tear data of barrier-free facilities are obtained by surface scanners, and the deterioration coefficient is generated by comparing the wear depth with the preset safety threshold (such as handrail wear ≤0.5mm):

[0033] wherein: is the deterioration coefficient, is the measured wear depth, is the safety threshold. The final facility adaptation degree evaluation value is the geometric mean of the usage matching degree and the deterioration coefficient, and when the coefficient is less than 0.4, a yellow warning state is triggered.

[0034] The analysis of the behavior safety assessment value focuses on abnormal pattern recognition. Abnormal activity trajectories are defined into three categories: wandering type (repeated path in the same area more than 5 times), staying type (staticity beyond time in non-rest area), and boundary crossing type (entry into restricted area). The frequency statistics use sliding counting method: record the number of abnormal events in each 15-minute time slice, and accumulate to form the daily frequency curve. The spatio-temporal distribution characteristics of emergency call trigger data are extracted by density clustering algorithm: project the alarm events into three-dimensional space-time coordinate system, identify the cluster by DBSCAN algorithm, and set the cluster radius to 2 meters / 30 minutes. The behavior safety assessment value is generated by the weighted harmonic mean of abnormal frequency and alarm density, and the alarm density is given a weight coefficient of 70% in the calculation formula.

[0035] The data processing flow implements a hierarchical parallel architecture. The raw sensor data enters the stream processing engine for time alignment and unit standardization, forming a uniform timestamp observation value sequence. The feature extraction layer deploys three parallel computing units: the spatial occupancy analysis unit runs the time series comparison algorithm, the facility assessment unit performs curve similarity calculation, and the behavior safety unit implements pattern recognition. The intermediate results are stored in distributed cache, and the aggregation engine triggers fusion calculation according to the evaluation period. The evaluation value output is packaged in JSON-LD format, including data quality identifiers (integrity, freshness, accuracy, and three check codes).

[0036] The dynamic calibration mechanism ensures the stability of the evaluation. The spatial usage frequency evaluation value is compared with the historical baseline every day, and when the fluctuation exceeds ±15% for 3 consecutive days, the model parameter review is started. The facility adaptation degree evaluation value performs sensor data cross-validation every week, and compares the consistency of the built-in log and external monitoring data. The behavior safety evaluation value updates the abnormal pattern definition library every month, and adds typical behavior samples to the training data set. All evaluation value generation processes record operation logs, including data traceability chain (original data ID→processing algorithm version→output timestamp), supporting full life cycle audit tracking.

[0037] The evaluation result visualization system presents a multi-dimensional dashboard. The spatial usage frequency heat map uses gradient color to mark the usage intensity of each area, and superimposes the time axis slider to show the periodical change. The facility adaptation degree radar chart shows five core indicators: usage compliance, wear state, energy efficiency, operation convenience, and safety coefficient. The behavior safety trend chart displays the seven-day moving average in the form of a line, and marks the abnormal event occurrence position marker. All charts support drilling query, and clicking the data point can drill down to the original sensor reading level. The visualization engine uses WebGL rendering, automatically updates the data source every 30 seconds, and maintains real-time information.

[0038] A hierarchical response mechanism is established within the exception handling process. When the space usage frequency assessment value consistently falls below 0.2, the space replanning recommendation generation process is triggered. When the facility suitability assessment value enters the red alert range (0-0.3), a maintenance work order is automatically created and sent to the property management terminal. If the behavioral safety assessment value drops by more than 20% in a single day, a real-time alert function is activated for supervisors, and a detailed behavioral analysis report is delivered. The execution status of all response actions is recorded, forming a closed-loop processing log for subsequent optimization of the decision-making rule base.

[0039] The system maintenance module implements preventative management. Compute node resource utilization is monitored at the minute level, and container instances are automatically scaled when memory usage exceeds 80%. Data pipelines have a breakpoint-resume mechanism, caching data packets for up to 72 hours in the event of a network outage. The evaluation model performs monthly version rolling updates, retaining the three most recent versions for rollback selection. The database executes scheduled defragmentation tasks, triggering compression and archiving when log files exceed 1GB.

[0040] Example 3: See Figure 3 The operating mechanism of the spatial resource allocation module is based on a dynamic optimization algorithm to achieve multi-parameter decision-making. The normalization process uses a piecewise linear transformation method to uniformly convert input parameters of different dimensions into a standard measurement range. The normalization function of the spatial use frequency assessment value defines the input value interval [0.1, 0.9] mapped to the interval [0, 1]. Values ​​below 0.1 are truncated to zero, and values ​​above 0.9 are saturated to one. The facility adaptability assessment value is processed using a logarithmic transformation to compensate for the right-skewed characteristics of the data distribution. The conversion formula is:

[0041] in: represents the normalized facility fitness evaluation value, = is the raw assessment value (range 0-1). The behavioral safety assessment value is directly linearly scaled to preserve the relative distribution characteristics of the original data. All normalization operations are performed in the floating-point unit, and the processed values ​​retain four decimal places of precision.

[0042] The dynamic weighting algorithm implements a three-tier decision logic. The basic weight allocation uses static preset values: spatial usage frequency weight =0.45, facility adaptability weight =0.35, behavioral safety weight =0.20. The weight adjustment factor is derived from environmental status monitoring data. When the air quality index (AQI) exceeds 100, the behavioral safety weight increases by 20%; when the user's mobility score decreases, the facility adaptability weight increases by 15%. The fusion calculation of the space optimization demand index uses a weighted power average model:

[0043] wherein: is the space optimization demand index, , , represent the normalized evaluation values of the three categories respectively, is the harmonic coefficient (default value 1.5). The index output range is 0-1, and the value is updated every 30 minutes. When the change amplitude exceeds 0.05, the visualization system refreshes.

[0044] High-frequency usage area identification uses a sliding window statistical method. The space usage frequency evaluation value is calculated as a moving average with a 6-hour window. When the value of three consecutive windows exceeds the first benchmark value (0.75), the area state is marked as "high-frequency usage". The idle area judgment combines time dimension analysis, requiring two conditions to be met simultaneously: the daily usage frequency evaluation value is lower than the second benchmark value (0.25), and the past seven-day average value has a significant downward trend (linear regression slope <-0.02). The area state marking information is stored in the space topology database, including area number, marking type, effective timestamp and confidence score.

[0045] The generation process of the furniture position reset scheme implements constraint optimization. The space topology relationship is modeled as a weighted undirected graph, with nodes representing functional areas and edge weights reflecting actual path distances. The optimization objective function minimizes the resource transfer cost between high-frequency areas and idle areas, while maximizing the space usage efficiency improvement amplitude. The solution is searched by genetic algorithm, with a population size of 50 and no more than 100 iterations. The output instructions include a list of moving items, target coordinate sequences and execution priority, with coordinate accuracy controlled within ±5 cm. Scheme verification uses discrete event simulation to pre-act the adjusted space usage in a virtual environment.

[0046] The instruction activation mechanism establishes multi-level trigger conditions. When the space optimization demand index exceeds the preset demand threshold (0.7) for more than 4 hours, the system enters the instruction preparation state. At this time, the stability of the current environmental parameters is checked: the temperature fluctuation is required to be less than ±2℃, and the intensity of human activity is within the normal range. After meeting the conditions, the space layout adjustment instructions are sent to the smart home control terminal in batches, with each instruction interval not less than 15 minutes to avoid user interference caused by centralized operation. The instruction execution state is fed back in real time through RFID tags, and when the deviation between the actual position of furniture and the target coordinates exceeds 10 cm, the position calibration process is triggered.

[0047] The anomaly handling subsystem monitors optimization process risk indicators. Instruction execution is suspended when the following conditions are detected: user health event alert, spatial structure safety warning signal activation, power supply instability state duration timeout. Interrupt event records detailed context data, including environmental sensor readings, user location trajectory, and system log snapshot. Resuming execution requires manual confirmation or waiting for the automatic diagnosis module to output a safety assessment report. All interrupt events generate an analysis summary for adaptive adjustment of optimization algorithm parameters.

[0048] The historical decision database stores complete optimization process records. Each spatial layout adjustment instruction is associated with the following metadata: trigger timestamp, demand index value, weight configuration version, region marker state snapshot, furniture movement trajectory record, and execution time consumption. Data is compressed and archived in columnar storage format, retaining complete operation logs for the last 90 days. A spatial optimization performance analysis report is generated monthly, with indicators such as instruction execution success rate, resource transfer efficiency improvement, and user adaptation period length.

[0049] The visual interactive interface provides three-dimensional space editing functions. Authorized users can manually adjust the system-generated scheme through touch devices, and the modified configuration is automatically saved as an alternative scheme. The system regularly compares the actual use effect differences between the automatic scheme and the manually adjusted scheme, and the difference data is used for optimization algorithm training. The interface also displays a real-time space state overview: the current high-frequency use area is highlighted in orange, idle areas are displayed in light gray, and the area corresponding to the adjustment instruction being executed presents a pulsing flashing effect.

[0050] The dynamic learning module continuously improves decision quality. A reinforcement learning framework is used to build a feedback loop, with the actual space use efficiency improvement after each optimization as the reward signal. The neural network model receives 126-dimensional feature input, including environmental parameters, user behavior patterns, and space configuration state. Training data is updated every 24 hours, and the model inference result is weighted and fused with the rule engine output, gradually reducing the preset rule weight proportion. Model version management implements an A / B testing mechanism, with the new version running in shadow mode for 72 hours before being put into actual decision-making.

[0051] Embodiment 4: see Figure 4, the generation process of facility configuration adjustment instructions is illustrated with the example of bathroom space renovation in a certain residential care home. When the system detects that the facility adaptation degree evaluation value of the area is below the preset adaptation threshold value of 0.5 for three consecutive days, the deep diagnosis process is started. The historical usage frequency data extracts the records of the last 30 days from the database, showing that the daily average usage number of the barrier-free handrail is 28 times (standard value range 20-35 times), but the single holding time is extended from the standard 3-5 seconds to 7-9 seconds. The wear detection data reflects that the handrail surface wear depth reaches 0.6mm, which exceeds the safety threshold of 0.5mm, and the right side fixed bolt appears a loosening displacement of 0.2mm.

[0052] Table 1: Analysis of facility status data retrieved by the system as follows.

[0053]

[0054] The facility degradation rate model analysis shows that the barrier-free handrail is expected to have a remaining service life of 14 days under the current usage intensity, which is combined with the following characteristics: the material fatigue curve shows that when the wear depth exceeds 0.65mm, the structural strength decreases by 40%; the bolt loosening degree increases by 0.1mm per week; the average grip strength detection value of the user is reduced by 8% compared with last month. The behavior safety evaluation value is analyzed synchronously, and it is found that the fall risk index of the area in the past seven days has increased by 22%, which has a significant correlation with the decrease in handrail usage stability.

[0055] The generation of the facility replacement priority list adopts a multi-criteria decision-making method. The scoring system includes five dimensions: safety risk weight 40%, usage frequency weight 25%, maintenance cost weight 15%, user dependence weight 15%, and alternative solution availability weight 5%. The final score of the barrier-free handrail is 83 points (full score 100), triggering the immediate replacement suggestion. The list details: the priority replacement handrail model is GH-2024 antibacterial, and the installation needs to be completed during the user's lunch break (13:00-15:00), and wall strength detection needs to be performed after the old parts are removed. The supporting suggestions include simultaneous replacement of the non-slip floor mat (score 71), while the electric toilet (score 58) is included in the next month's maintenance plan.

[0056] The instruction feedback verification phase implements a double-cycle evaluation mechanism. The first evaluation cycle (24 hours) focuses on monitoring the usage data of the new handrail: the usage number on the first day after installation returns to 32 times, the single holding time is shortened to 4.2 seconds, and the grip distribution uniformity is improved by 35%. The second evaluation cycle (7 days) comprehensively analyzes the overall safety indicators of the area: the fall risk index falls to the level before the renovation, and the number of false triggers of the emergency call button is reduced by 40%. The space optimization demand index decreases from the initial 0.72 to 0.51, verifying the effectiveness of the adjustment instructions.

[0057] When the verification result is not as expected (e.g., the index is still higher than 0.6), the system initiates the parameter weight correction process. Taking this renovation as an example, the initial weight setting is 70% for safety risk, 30% for usage frequency. After correction, it is adjusted to 60% for safety risk, 25% for usage frequency, and 15% for user habit adaptability. The correction is based on the newly added monitoring data: users have a 3-day adaptation period for the new handrail height, during which the usage pattern shows a clear learning curve feature.

[0058] The user interaction interface provides a confirmation function for the renovation scheme. After the handrail replacement instruction is generated, the system pushes a three-dimensional schematic diagram to the monitoring terminal: the new and old parts are displayed in contrast, showing that the anti-slip line density increases by 50% and the support area expands by 20%. The user side can make adjustment suggestions, such as modifying the installation height from the standard 85 cm to 82 cm to adapt to specific physical conditions. The system records all manual intervention items, and if the same modification is made more than three times, it will automatically update the size parameters in the elderly-friendly standard model.

[0059] The cost control system is embedded in the decision-making process. The budget analysis involved in this renovation shows that the priority replacement project accounts for 23% of the monthly maintenance funds, but is expected to reduce related maintenance expenses by 41% in the next 6 months. The system automatically generates a comparison report of alternative schemes, including key procurement indicators such as price-performance ratio, expected service life, and local service response time for different brands and models.

[0060] The abnormal situation handling establishes a multi-level response strategy. During this renovation process, when it is detected that users have allergic reactions to the new material (manifested as skin rash after use), the system immediately initiates a secondary response: temporarily restore the use of the old handrail, and at the same time trigger the biocompatibility detection process. The medical consultant terminal receives a detailed material composition report and gives a safety evaluation conclusion within 72 hours. Such events are recorded as special cases and updated to the contraindication database of the user's health record.

[0061] Long-term performance tracking adopts a monthly evaluation mode. Within 30 days after the renovation is completed, the system continuously collects the following indicators: handrail strain sensor data (daily average micro-deformation <0.01 mm), user satisfaction score (weekly voice evaluation sentiment analysis), and changes in the usage pattern of related facilities (e.g., toilet lift usage frequency increased by 12%). These data are used to generate a facility life cycle report, predict the next replacement cycle of similar components, and optimize the time node of preventive maintenance plans.

[0062] Example 5: see Figure 5The operational mechanism of the spatial risk warning module is built on a multi-channel processing architecture of real-time data streams. The spatial structural stability parameters are monitored by a distributed sensor array, in which strain measurement nodes are embedded inside load-bearing walls. A triaxial strain gauge is installed every 10 centimeters, with a sampling frequency set at 50 data points per second. Inclination sensors are installed at the beam-column junctions, with a two-axis dynamic detection accuracy of 0.01 degrees. The structural deformation safety threshold library contains critical parameters for 12 types of building components, such as the allowable strain value of 150 με for concrete walls and the deflection threshold of 1 / 400 of the span for wooden beams. When the real-time monitored wall strain value exceeds 140 με and the upward trend continues for more than 5 minutes, the first-level warning signal generation process is triggered. The signal activation includes automatically executed physical intervention measures: electromagnetic door locks deployed at the boundary of the danger zone are powered off within 200 milliseconds after receiving the signal and enter the locked state; at the same time, the structural reinforcement device in the basement starts the hydraulic support, which lifts the telescopic support rod to the beam bottom position at a speed of 15 centimeters per minute, forming a temporary support structure.

[0063] The monitoring network of environmental safety parameters covers the entire spatial three-dimensional grid. The harmful gas detection system combines electrochemical sensors with photoionization detectors, with a carbon monoxide sensor range of 0-1000 ppm and a methane detection sensitivity of 1 ppm. Gas concentration data are refreshed every two seconds, and sampling probes are arranged at different height layers from 0.3 meters to 1.8 meters above the ground. The fire risk factor monitoring uses multi-spectral flame detection technology combined with a point-type temperature detector array. When the CO concentration in the kitchen area exceeds 24 ppm and the temperature gradient rises by 4°C within 30 seconds, the system determines it as a precursor to gas leakage and combustion, triggering a second-level warning signal. The signal linkage response executes multiple device cooperative actions: the fresh air system starts the emergency mode to increase the air exchange volume to three times the normal value; the intelligent gas valve cuts off the pipeline supply within 500 milliseconds; the water mist sprinkler installed in the ceiling locates the fire source position based on the temperature distribution map and implements point cooling in a fan-shaped coverage manner.

[0064] The behavior anomaly monitoring system integrates non-contact sensing technology. The fall detection system uses a 60GHz millimeter wave radar array to extract human joint motion trajectories through micro-Doppler features. Among the pre-set 10 high-risk fall patterns, the focus is on detecting the sudden change in hip joint acceleration during lateral falls, with a trigger threshold of 3.5g instantaneous impact. Long-term stationary detection uses thermal imaging and UWB positioning composite verification technology. When the bedroom area detects a heat source that has been in the same position for 60 minutes with a position deviation of less than 20 centimeters, it is determined to be a potential loss of consciousness risk. The behavior anomaly threshold is dynamically adjusted based on the user's daily activity baseline, such as a 70% decrease in the movement frequency threshold at night compared to during the day. After triggering a three-level early warning signal, the alarm information is compiled into a standard communication protocol package, including event type code, spatial coordinate data, and biological feature summary, which is transmitted to the monitoring platform through a 5G private network. At the same time, the positioning system switches to a high refresh rate mode, and the update frequency of the wearable positioning tag is increased from 1 to 10 times per second, forming a continuous tracking trajectory.

[0065] The early warning response execution unit establishes multiple safety mechanisms. The closed area in the first level of early warning state sets up a pressure sensing boundary, and when the user is detected to be within 1 meter of the isolation area, the audible and visual warning device is started, and the pre-recorded safety prompt voice is played. During the second level of early warning, the device linkage follows a priority control logic, and the gas shut-off and air purification functions are executed first, and the fire fighting operation is delayed for 3 seconds to avoid injury. The information push triggered by the third level of early warning adopts multi-channel redundant transmission, and the monitoring terminal automatically starts the backup communication link after receiving the main alarm to ensure reliable delivery of the alarm information. All active intervention operations implement a forced confirmation mechanism, such as completing the structure stress simulation verification before lifting the support device, and reconfirming the personnel position before spraying the water mist.

[0066] The event data recording system captures the complete early warning life cycle. Each level of early warning signal automatically creates an event archive, recording the trigger time point to the nearest millisecond, and saving the original sensor readings at a frequency of 100Hz for a time window of 20 seconds. The response action execution process records the operation instruction sequence and timestamp, such as the door lock response time recorded as the time difference between "signal reception - execution completion". The subsequent disposal information integrates the artificial operation log, including the identity verification record of the on-site disposal personnel and the device reset operation time. The data archiving adopts a hierarchical storage strategy, with raw waveform data retained for 48 hours, feature extraction data retained for 90 days, and event summary information permanently stored in the disaster recovery center.

[0067] The early warning parameter dynamic optimization module implements a closed-loop learning mechanism. Every month, all early warning events are analyzed retrospectively, and the correspondence between the trigger threshold and the actual risk is compared. When a false alarm event occurs (such as smoke false triggering during the cooking peak period), the threshold coefficient in the corresponding scene is automatically adjusted. When a missed alarm occurs (such as a minor fall not being recognized), the system collects case data to update the pattern recognition feature library. The algorithm model version is updated quarterly, and the new model is only deployed to the production system after simulating 200 test cases in a sandbox environment. The user preference setting module allows guardians to customize response parameters, such as setting the notification delay time of the three-level early warning in the range of 0-30 seconds.

[0068] The system state monitoring system ensures operational reliability. The heartbeat packet of the early warning master control unit is sent every 500 milliseconds, and if no response is received for three consecutive times, it automatically switches to the backup controller. The health status of the sensor network is scanned every hour, and faulty nodes are automatically isolated and trigger maintenance work orders. The communication link implements a dual-ring redundant design, and immediately enables wireless Mesh networking when the main fiber channel is interrupted. The system performs full-link stress testing every month to simulate the resource load capacity under the three-level concurrent early warning scenario. All hardware devices are equipped with an electronic tag management system, and maintenance personnel can obtain device maintenance history and current state data by scanning the two-dimensional code.

[0069] The human-machine interaction interface realizes panoramic situation visualization. The status of each early warning area is displayed in real time in the three-dimensional building model, with the first-level early warning area displayed as red pulsing light, the second-level early warning area displayed as orange wave diffusion effect, and the third-level early warning position highlighted and flashing. When an alarm occurs, a multi-layer information card is automatically popped up: the first layer displays the event location and type identifier, the second layer expands the dynamic chart of environmental parameters, and the third layer provides response operation suggestion buttons. The historical early warning data analysis panel provides multi-dimensional filtering functions, and can generate statistical reports according to time distribution, spatial hotspots, and response effect indicators. Authorized users can mark the areas of interest in the three-dimensional space by gesture control system, and the system continuously pushes real-time monitoring data streams of the area.

[0070] It should be noted that, in the present text, relational terms such as first and second are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof are intended to cover non-exclusive inclusions, so that a process, method, article, or apparatus that includes a list of elements does not only include those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0071] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. An intelligent analysis system based on spatial data of elderly care housing, characterized in that: include: The spatial multi-source perception module is used to collect real-time physical environment parameters, facility usage status parameters, and resident behavior characteristic parameters of the elderly care residence; A spatial efficiency dynamic analysis module is configured to receive the data collected by the spatial multi-source perception module, set an evaluation period, extract features of the spatial occupancy pattern within each evaluation period, and generate a spatial usage frequency evaluation value, a facility adaptability evaluation value, and a behavioral safety evaluation value; A space resource configuration module is configured to calculate the space optimization demand index within each evaluation period based on the space usage frequency evaluation value, the facility adaptability evaluation value, and the behavior safety evaluation value, and output space layout adjustment instructions and facility configuration adjustment instructions; The spatial risk warning module is used to continuously monitor the spatial structural stability parameters, environmental safety parameters and abnormal behavior parameters of the retirement residence and generate multi-level spatial safety warning signals.

2. The intelligent analysis system based on spatial data of elderly care housing according to claim 1 is characterized in that: The physical environment parameters include light intensity distribution data, temperature and humidity distribution data, and air quality index data; the facility usage status parameters include furniture usage frequency data, electrical appliance operation status data, and barrier-free facility wear data; the resident behavior characteristic parameters include activity trajectory data, stay duration data, and emergency call trigger data.

3. The intelligent analysis system based on spatial data of elderly care housing according to claim 2 is characterized in that: The space efficiency dynamic analysis module extracts features of space occupancy patterns within each evaluation period, specifically performing the following steps: Dividing the evaluation period into consecutive time periods, extracting the deviation between the actual occupancy time of the space area and the preset standard occupancy time in each consecutive time period, and obtaining a space usage frequency evaluation value; Analyze the matching degree between the frequency of facility use data in each continuous time period and the preset standard model for elderly-friendly facilities, and combine the deterioration degree of barrier-free facility wear data to obtain the facility adaptability assessment value; The frequency of abnormal activity trajectories and the distribution characteristics of emergency call triggering data in each continuous time period are counted to obtain the behavioral safety assessment value.

4. The intelligent analysis system based on spatial data of elderly care housing according to claim 3 is characterized in that: When the space resource allocation module calculates the space optimization demand index, it specifically performs the following steps: Normalizing the space usage frequency assessment value, facility adaptability assessment value, and behavior safety assessment value; A dynamic weighting algorithm is used to integrate the normalized space usage frequency assessment value, facility adaptability assessment value, and behavioral safety assessment value to output the space optimization demand index. When the space optimization demand index exceeds the preset demand threshold, the space layout adjustment instruction and the facility configuration adjustment instruction are activated.

5. The intelligent analysis system based on spatial data of elderly care housing according to claim 4 is characterized in that: The generation logic of the spatial layout adjustment instruction includes: Identify areas where the space usage frequency assessment value exceeds a first benchmark value and mark them as high-frequency usage areas; Identify areas where the space usage frequency assessment value is lower than a second benchmark value and mark them as idle areas; Based on the topological relationship between high-frequency use areas and idle areas, a furniture repositioning plan is generated as a spatial layout adjustment instruction.

6. The intelligent analysis system based on spatial data of elderly care housing according to claim 5 is characterized in that: The generation logic of the facility configuration adjustment instruction includes: When the facility suitability evaluation value is lower than the preset suitability threshold, the historical usage frequency data and current wear data of the corresponding facility are retrieved; Based on the remaining service life prediction of the facility degradation rate model and the decline of the behavioral safety assessment value, a facility replacement priority list is generated as an instruction for facility configuration adjustment.

7. The intelligent analysis system based on spatial data of elderly care residences according to claim 6 is characterized in that: The space resource configuration module also performs instruction feedback verification: After outputting the spatial layout adjustment instructions and facility configuration adjustment instructions, re-collect the spatial multi-source perception data for the next evaluation cycle; Update the space optimization demand index based on the re-collected data, and verify whether the updated space optimization demand index is lower than the preset demand threshold; If it is not lower than the preset demand threshold, the parameter weights of the spatial layout adjustment instructions and the facility configuration adjustment instructions are corrected.

8. The intelligent analysis system based on spatial data of elderly care residences according to claim 1 is characterized in that: When the space risk warning module generates a multi-level space safety warning signal, it specifically performs the following operations: Compare the spatial structure stability parameters with the structural deformation safety threshold in real time, and trigger a first-level warning signal if it exceeds the threshold; Continuously monitor the concentration of harmful gases and fire risk factors in environmental safety parameters, and trigger a secondary warning signal if the corresponding threshold is exceeded; Analyze the fall recognition features and long-term inactivity features in the abnormal behavior parameters. If the abnormal behavior threshold is exceeded, a third-level warning signal is triggered.

9. The intelligent analysis system based on spatial data of elderly care residences according to claim 8 is characterized in that: The space risk warning module also includes a warning response mechanism: When a level one warning signal is triggered, the dangerous area is automatically locked and the structural reinforcement device is activated; When the second-level warning signal is triggered, the ventilation equipment and fire-fighting equipment will be linked to respond to the emergency; When the third-level warning signal is triggered, the alarm information is pushed to the monitoring terminal in real time and the positioning tracking function is activated.

10. The intelligent analysis system based on spatial data of elderly care residences according to claim 1 is characterized in that: It also includes a spatial data tracing module, which is used to store historical space optimization demand indexes, space layout adjustment instruction execution records, facility configuration adjustment instruction execution records, and multi-level space safety warning signal trigger records, and generate space utilization efficiency trend reports.

Images