Barrier-free public facility layout optimization method

Abstract

The invention relates to the field of barrier-free public facilities, in particular to a barrier-free public facility layout optimization method. Constructing a demand prediction model by using a multi-source data fusion technology; predicting future barrier-free facility demand distribution by adopting a machine learning algorithm of a regression model; developing a site selection optimization algorithm in combination with a geographic information system (GIS) and a demand prediction model; a sensor and a camera are deployed in a barrier-free facility, and the use condition, congestion degree and damage information data of the facility are collected in real time; equipment state monitoring is carried out in real time; based on real-time monitoring data, by analyzing the use frequency and congestion degree of the facilities, generating an adjustment scheme which temporarily comprises path guidance adjustment and temporary mobile facility layout increase; generating a personalized barrier-free path according to specific requirements, physical conditions and travel preferences of the user; and updating and optimizing path recommendation in real time based on user data and environmental perception data including gradient and road conditions.

Description

Technical Field

[0001] The present invention relates to the field of barrier-free public facilities, and particularly to a method for optimizing the layout of barrier-free public facilities. Background Art

[0002] In the field of optimizing the layout of barrier-free public facilities, various solutions have been proposed in the prior art. For example, the method for optimizing the layout of barrier-free public facilities described in Chinese Invention Patent No. 2023113779592. This solution optimizes the layout of facilities through simulation operations by obtaining the location information of barrier-free public facilities, mobile phone signaling data, as well as the user's staying location and time, confirming the residence of the target user in combination with the characteristics of the place of residence, and calculating the service hinterland of the facilities. However, there are still some deficiencies and drawbacks in the practical application of this prior art, mainly reflected in the limitations of data collection and analysis, insufficient dynamic adaptation ability, inadequate consideration of user needs personalization, limited real-time optimization effect, and insufficient utilization of environmental perception data. First of all, the prior art relies relatively heavily on mobile phone signaling data as the main information source for obtaining user location and behavior, and this data has certain limitations. Although mobile phone signaling data can provide the user's staying location and staying time, its accuracy and real-time performance are insufficient. Especially in complex urban environments, the coverage and accuracy of mobile phone signals may be interfered by factors such as buildings and underground facilities. For example, in enclosed places such as subway stations and shopping malls, the signal may attenuate, resulting in the system's difficulty in accurately capturing the user's precise location, thereby affecting the calculation of the service scope of the facilities. In addition, mobile phone signaling data cannot effectively distinguish the types of user behaviors. For example, users who stay in an area for a long time may just stop temporarily, rather than being target users who need barrier-free facilities in the long term. Since these data are difficult to accurately distinguish the actual needs of users, the system may make incorrect optimization decisions on the layout of barrier-free facilities, resulting in waste of resources or insufficient coverage. Secondly, the inadequate consideration of user needs personalization is a major defect of the existing solutions. Although the existing patent infers the user's place of residence and travel patterns by analyzing the user's staying time and frequency, this method is still based on static behavior pattern analysis and lacks in-depth exploration of individual user needs. For example, groups such as wheelchair users, visually impaired people, and the elderly have quite different needs for barrier-free facilities, and it is difficult for the prior art to effectively capture these differences and conduct targeted layout optimization. Different types of users have different priorities and specific requirements when using barrier-free facilities. For example, wheelchair users have higher requirements for slopes, while visually impaired people may need voice navigation or blind path guidance more. The prior art relies more on the characteristics of the user's place of residence to speculate on the overall demand distribution, while ignoring the diversity and dynamics of individual user needs, which results in the lack of flexibility and adaptability of its optimization results in practical applications and cannot provide truly personalized services for special groups. There are also significant deficiencies in the utilization of environmental perception data in the prior art. Although the existing solution can optimize the service scope of barrier-free facilities by combining mobile phone signaling data and the characteristics of the place of residence, its consideration of actual environmental factors is relatively limited. In the actual environment, road conditions, slopes, weather conditions, etc. will all have an important impact on the use of barrier-free facilities.For example, the needs of wheelchair users for ramps are different on rainy days. Ramps may become slippery on rainy days, posing a safety hazard. If the system only relies on location data and user behavior history and cannot effectively identify these dynamic environmental changes, it may not be able to adjust path recommendations and facility layouts in a timely manner. Similarly, in extreme situations such as earthquakes or bad weather, the system lacks the ability to perceive and give early warnings, which may lead to the failure of barrier-free facilities to serve users in a timely manner, reducing the safety and practicality of facility layouts. Therefore, the lack of real-time environmental perception data (such as weather, road surface quality, etc.) will cause existing solutions to react slowly when dealing with environmental changes, affecting the effective use of facilities.

[0003] In summary, the main deficiencies of the existing optimization methods for the layout of barrier-free public facilities include: relying on mobile phone signaling data resulting in insufficient accuracy and inability to distinguish the actual needs of individual users; insufficient consideration of the personalization of user needs and lack of in-depth exploration of the differentiated needs of special groups; lack of dynamic adaptation ability and difficulty in making real-time responses during peak periods and emergencies; insufficient utilization of environmental perception data and inability to effectively handle actual influencing factors such as weather and road surfaces; limited real-time optimization effect and difficulty in quickly adjusting facility layouts; lack of a flexible deployment mechanism for mobile facilities and difficulty in coping with dynamic changes during facility use. Summary of the Invention

[0004] The purpose of the present invention is to provide an optimization method for the layout of barrier-free public facilities, so as to solve some of the drawbacks and deficiencies pointed out in the background technology.

[0005] The technical solutions adopted by the present invention to solve its above-mentioned technical problems are as follows: an optimization method for the layout of barrier-free public facilities, including: S1. Prediction of barrier-free facility needs and optimization of site selection:

[0006] S1.1. Using multi-source data fusion technology, integrating different data types including static geographical data and dynamic traffic and pedestrian flow data to construct a demand prediction model;

[0007] S1.2. Adopting a machine learning algorithm of a regression model to analyze historical usage data and population behavior to predict the future demand distribution of barrier-free facilities;

[0008] S1.3. Combining the geographical information system GIS and the demand prediction model to develop a site selection optimization algorithm; the site selection optimization algorithm includes calculating: terrain, traffic convenience, distance and coverage of existing facilities; through a simulated annealing algorithm, generating a facility layout plan;

[0009] S2. Dynamic monitoring and intelligent layout:

[0010] S2.1. Deploy sensors and cameras in barrier-free facilities including wheelchair ramps, elevators, and blind paths to collect real-time data on the usage, congestion level, and damage information of the facilities; transmit the data to the central control system through Internet of Things technology for real-time monitoring of the device status;

[0011] S2.2. Based on the real-time monitoring data, generate an adjustment plan that temporarily includes adjusting path guidance and temporarily increasing the layout of mobile facilities by analyzing the usage frequency and congestion level of the facilities;

[0012] S2.3. Use the monitoring data for fault prediction of the facilities, adopt predictive maintenance technology, and predict potential faults of the facilities by analyzing historical data and real-time usage data;

[0013] S3. Personalized barrier-free path planning and adaptive navigation:

[0014] S3.1. Generate a personalized barrier-free path according to the specific needs, physical condition, and travel preferences of the user; then, based on the user data and environmental perception data including slope and road conditions, update and optimize the path recommendation in real time;

[0015] S3.2. Provide real-time navigation through a mobile application or wearable device, and adaptively adjust the user's travel path according to dynamic data including road construction and facility congestion.

[0016] Furthermore, the strategy for generating the temporary adjustment plan includes: using a machine learning model to analyze the historical usage data of the facilities, and combining real-time data to predict the future usage situation of the facilities; the usage frequency f u (t) and the congestion level f c (t) are recorded and calculated, and it is determined whether to make an adjustment by judging whether they exceed the normal range through a threshold:

[0017]

[0018] where Φ(t) represents the change in facility usage pressure, which is used to measure the change in the usage and congestion status of the facility within time t; f u (t) is the facility usage frequency function, which describes the usage density of the facility at each moment and represents the frequency of use of the facility; f c (t) is the facility congestion level function, which represents the congestion situation of the facility per unit time and reflects whether the facility is in an overused state; and are the change rates of the usage frequency and congestion level, which represent the change speed of the facility usage frequency and congestion level within time τ; Θ(x) is the step function, which is used to judge whether the usage frequency or congestion level of the facility exceeds a certain threshold; when x > 0, the value of the step function is 1, otherwise it is 0;

[0019] u th and c th are the usage frequency threshold and the congestion level threshold respectively. When the usage frequency or congestion level exceeds the threshold, a dynamic adjustment plan is triggered; by calculating Φ(t), based on real-time monitoring data, it is determined whether the usage pressure of the facility is too high to decide whether adjustment is needed.

[0020] Furthermore, the strategy for generating a temporary adjustment plan includes: when it is detected that the usage frequency or congestion level of the facility exceeds the threshold, a temporary path adjustment plan is generated; by real-time monitoring the usage of the facility and combining with the path planning function P(x, y, t), an alternative path is provided for the user; and the path planning is optimized by combining path distance, time cost, and facility usage to generate a dynamic path guidance:

[0021] P(x, y, t) = min(∫ A (f dist (x, y) + λ 1 ·f time (x, y, t) + λ 2 ·f usage (x, y, t))dx dy)

[0022] where P(x, y, t) is the path weight function, describing the total weight of choosing a path at time t starting from location (x, y); the path is recommended by minimizing the value; f dist (x, y) is the path distance function, representing the physical distance from the current location (x, y) to the target location; f time (x, y, t) is the time cost function, representing the time consumption of using the path at time t; f usage (x, y, t) is the facility usage frequency function, representing the real-time usage of the facilities passed by the path; λ 1 and λ 2 are adjustment coefficients, respectively controlling the influence of time cost and facility usage frequency on path selection.

[0023] Furthermore, the strategy for generating a temporary adjustment plan includes: when the barrier-free facilities in a certain area are continuously congested or in a high-frequency usage state and the temporary path adjustment cannot effectively solve the problem, it is recommended to temporarily add mobile facilities including temporary ramps, handrails, and signs through a resource allocation algorithm; based on real-time monitoring data, analyze the current demand of the facilities, and through the resource scheduling function R(z, t), dynamically determine whether it is necessary to add mobile facilities and optimize the deployment location of the facilities:

[0024]

[0025] Among them, R(z, t) is the resource allocation priority function, which is used to measure the priority of adding mobile facilities at location z and time t; f need (z, t) represents the regional demand function, which describes the demand intensity of the region for mobile facilities at location z and time t; Γ(f mob (z, t)) is the mobile facility availability function, which reflects the number of schedulable mobile facilities at location z at time t; T is the time period of resource scheduling, indicating that the regional demand and facility availability are evaluated within the period T.

[0026] Furthermore, the method for generating personalized barrier-free paths includes: generating personalized paths that meet the needs according to the types of users including wheelchair users, visually impaired people, and the elderly; constructing a preliminary personalized path weight function through the following formula:

[0027]

[0028] Among them, P u (x, y) represents the preliminary path weight value of user u at location x, y, and is a path recommended for the user according to the user's personalized needs and travel preferences; Ω represents the regional scope of the entire path, and the optimal point of the path is searched within the scope; S u (x, y) is the slope function, which measures the slope suitability of user u at location x, y; a low slope is suitable for wheelchair users, and 1 - S u (x, y) is used to reflect the impact of the slope on the user. The smaller the slope, the higher the value, and the better the path.

[0029] D u (x, y) is the distance function, which represents the physical distance from the user's current location to the target location; is used in the formula to reduce the priority of long paths; U u (x, y) is the device usage preference function of user u, which reflects the degree of dependence of the user on barrier-free facilities during passage; α 1 , α 2 , α 3 are weight coefficients, corresponding to slope, distance, and user preference respectively, and are used to balance the importance of factors in path recommendation.

[0030] Furthermore, the method for generating personalized barrier-free paths includes: after generating the preliminary path, real-time obtaining environmental perception data through deployed Internet of Things devices and including slope sensors, road surface quality monitoring devices, and weather monitoring, monitoring the environmental status of the path, and dynamically optimizing the path based on the real-time data; when the environmental conditions change, adjusting the path; dynamically adjusting the path through the formula:

[0031] P e (x, y, t) = Pu (x, y) + ∫ Ω (β 1 ·W(x, y, t) + β 2 ·R(x, y, t)) dx dy

[0032] where P e (x, y, t) represents the path weight after environmental optimization, which is the path weight dynamically adjusted based on the preliminary weight P u (x, y) and environmental perception data; W(x, y, t) is the weather condition function, which describes the weather conditions at position x, y at time t; including rain, snow, and slippery roads increasing the difficulty of passing and increasing the path weight value P e (x, y, t) to prompt the selection of a safe alternative path;

[0033] R(x, y, t) is the road surface quality function, which is used to reflect the road surface conditions of the path segment at time t; poor path quality will cause the value to rise, increasing the weight of the path to avoid users choosing poor road segments; β 1 , β 2 are the weight coefficients of environmental data, respectively adjusting the influence of weather conditions and road surface quality on path recommendation; weather and road surface conditions affect path selection to varying degrees.

[0034] The barrier-free public facility layout optimization method of the present invention has multiple beneficial effects. First, by introducing technologies such as multi-source data fusion, machine learning, and geographic information system (GIS), this method can accurately predict the future demand for barrier-free facilities and optimize the location layout of facilities accordingly, thereby maximizing the utilization rate of facilities and avoiding resource waste. Second, the present invention dynamically generates a path adjustment plan by real-time monitoring the usage situation, congestion level, and fault information of barrier-free facilities, ensuring that users can quickly obtain alternative barrier-free access plans in case of congestion or facility failures, improving the travel experience and safety of users. In addition, based on the dynamic optimization mechanism of environmental perception data (such as slope, weather, road surface quality, etc.), the travel path of users can be adjusted in real time to ensure that users can obtain the safest and most convenient travel path in case of sudden environmental changes, which is especially suitable for special groups such as wheelchair users, visually impaired people, and the elderly. Finally, the present invention also flexibly adds mobile facilities (such as temporary ramps, handrails, signs, etc.) through intelligent resource scheduling algorithms, effectively relieving the facility pressure during high-frequency use and ensuring the efficient operation of the barrier-free facility layout under dynamic demand. Generally speaking, this method significantly improves the management efficiency of barrier-free public facilities, the user experience, and the reasonable allocation of facility resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flowchart of the barrier-free public facility layout optimization method of the present invention.

[0036] Figure 2 A strategy flowchart for generating a temporary adjustment plan for the present invention.

[0037] Figure 3 A method flowchart for generating a personalized barrier-free path for the present invention. Detailed implementation manners

[0038] The following makes a detailed description of the detailed implementation manners of the present invention with reference to the accompanying drawings.

[0039] Combined with the attached Figure 1 , the barrier-free public facility layout optimization method includes the following steps: First, conduct barrier-free facility demand prediction and site selection optimization. Using multi-source data fusion technology, different data types are integrated, including static geographical data and dynamic traffic data and pedestrian flow data. By this integration, a demand prediction model is constructed to better understand the usage demand and distribution of barrier-free facilities. Then, a machine learning algorithm of a regression model is adopted to analyze historical usage data and population behavior patterns. Through learning the historical data, the machine learning model can identify the rules of users' use of barrier-free facilities and accurately predict the future facility demand, generating an estimate of the future barrier-free facility demand distribution. Subsequently, combined with the geographical information system GIS and the demand prediction model, a site selection optimization algorithm is developed. This algorithm not only calculates and considers factors such as terrain features and traffic convenience, but also combines the distribution, distance, and coverage range of existing barrier-free facilities, so as to ensure that the optimized layout plan can cover high-demand areas and has good traffic convenience. In this process, by introducing a simulated annealing algorithm, which gradually optimizes through random search and temperature parameter control and can effectively jump out of the local optimal solution, finally a globally optimal barrier-free facility layout plan is generated.

[0040] The second step involves dynamic monitoring and intelligent layout. Sensors and cameras are deployed at barrier-free facilities such as wheelchair ramps, elevators, and blind paths to collect data on the usage situation, congestion degree, and damage information of the facilities in real time. Through the Internet of Things technology, these data are transmitted to the central control system, thereby realizing real-time monitoring of the device status and ensuring that the usage situation and potential problems of the facilities can be obtained in a timely manner. Then, based on the data monitored in real time, the system analyzes the usage frequency and congestion situation of the facilities and generates temporary adjustment plans, including path guidance adjustment and layout adjustment plans for temporarily increasing mobile facilities, so as to ensure that the barrier-free facilities can still operate smoothly during high demand. Finally, the system uses the monitored data for fault prediction of the facilities, adopts predictive maintenance technology, and by analyzing historical data and real-time usage data, identifies potential faults occurring in the facilities and conducts maintenance in advance to reduce the failure rate and downtime of the devices and ensure the long-term stable operation of the barrier-free facilities.

[0041] The third step involves personalized barrier-free path planning and adaptive navigation. First, the system generates a personalized barrier-free path according to the user's specific needs, physical condition, and travel preferences. This personalized path generation process takes into account the user's special needs. For example, wheelchair users require roads with a smaller slope, visually impaired users need voice navigation support, and the elderly prefer shorter walking paths. After generating the path, the system updates and optimizes it in real time based on the user's personalized data and environmental perception data (such as slope, road surface conditions, weather, etc.) to ensure that the recommended path always meets the best passing conditions for the user. Then, the system provides real-time navigation services through a mobile application or a device worn by the user (such as a smartwatch, etc.) to ensure that the user can pass smoothly and safely. At the same time, the system adaptively adjusts the path according to dynamic data (such as road construction, facility congestion, etc.) to guide the user to avoid congested or impassable sections, thus ensuring that the user's travel experience remains efficient, personalized, and barrier-free.

[0042] Example 1:

[0043] Refer to Figure 2 , in a busy subway station in the center of a large city, the municipal government launched a plan to optimize the layout of barrier-free public facilities, aiming to improve the use experience of barrier-free facilities, especially to provide better travel convenience for special groups such as wheelchair users, visually impaired people, and the elderly. The municipal government selected the wheelchair ramp, elevator, and blind path of this station as pilots, installed sensors and cameras, and used machine learning models to predict and optimize the use of facilities.

[0044] First, in the system, the usage frequency f u (t) and congestion level f c (t) of the facilities are monitored in real time. The sensor records the number of wheelchairs passing through per minute, the usage times of the elevator, and the usage frequency of the blind path. Based on historical data, the system knows that the usage frequency of the elevator during the morning rush hour (7:30 - 9:30) on each working day is usually 5 times per minute, the congestion level f c (t) is relatively high, and the average waiting time is about 3 minutes; the usage frequency of the wheelchair ramp is relatively low, with 2 - 3 users passing through per minute.

[0045] To determine whether dynamic adjustment of the facility usage is needed, the system sets a threshold u th for the usage frequency and a threshold c th for the congestion level. According to historical analysis, the usage frequency threshold u th of the elevator is set to 8 times per minute, and the congestion level threshold c th is set to an average waiting time of 5 minutes. According to the actual data, the system detects the congestion and frequent usage during the morning rush hour, which triggers the adjustment mechanism.

[0046] According to the formula in the barrier-free facility layout optimization method:

[0047]

[0048] It is determined whether the facility needs to be adjusted through calculation. Suppose at time t = 8:30, the usage frequency f u (t) of the elevator is 10 times per minute, and the average waiting time of the congestion level f c (t) is 6 minutes. According to the formula, first calculate the change rate of the usage frequency per minute Suppose this change rate is 2 times / minute 2 . At the same time, the change rate of the congestion level is 1 minute / minute 2 . Since f u (t) exceeds the threshold u th and f c (t) exceeds the threshold c th , the values of the step functions Θ(f u (t) - u th ) and Θ(f c (t) - c th ) are both 1. Therefore:

[0049]

[0050] Suppose the morning peak period is 2 hours (120 minutes), then Φ(t) = 3×120 = 360, indicating that the usage pressure of the facility increases significantly during the peak period, and the system determines that the facility needs to be adjusted during the peak period.

[0051] Through the above calculation, the system confirms that the usage pressure of the current facility exceeds the normal range, so the dynamic adjustment plan is triggered. The adjustment strategies include temporarily adjusting the path guidance, adding a spare wheelchair passage and a mobile ramp. The system will guide users to a smoother barrier-free path according to the real-time situation and send a notice to users through the mobile application to prompt the use of the spare facilities.

[0052] Suppose the blind path is congested due to sudden water accumulation outside the subway station from 9:00 to 10:00 in the morning. The congestion level f c (t) of the blind path suddenly rises and exceeds the set c th threshold. The sensor detects that the number of people passing through the blind path per minute increases from the original 1 person to 5 people per minute. At the same time, due to the water accumulation, the passing speed becomes slower, resulting in an increase in the waiting time to 7 minutes. At this time, the system calculates Φ(t) again, triggers the dynamic adjustment mechanism, and recommends another relatively dry blind path for users to pass through to avoid the safety risks brought by the water accumulation area.

[0053] Regarding the coefficient value range, the usage frequency threshold u th can be set to 5 - 10 times per minute according to the regular usage of the facility type, and the congestion degree threshold c th can be set to an average waiting time of 3 - 7 minutes according to the design load capacity of the facility. Through this reasonable coefficient value, the system can flexibly respond to the optimization requirements of barrier-free facility layouts in different scenarios and make precise dynamic adjustments based on real-time data.

[0054] The usage of barrier-free facilities at the subway station during the morning rush hour triggered the system's dynamic adjustment strategy. Next, it will be explained how to generate a temporary path adjustment plan through path planning and real-time monitoring.

[0055] When the system detects that the facility usage frequency or congestion degree exceeds the threshold, the dynamic adjustment mechanism will trigger the path planning function to generate an alternative path suitable for the user. Suppose the system detects that the usage frequency of the wheelchair access path rises from 2 times per minute to 5 times between 8:30 and 9:00, exceeding the set usage frequency threshold u th = 4 times / minute, and at the same time the congestion degree f c (t) also increases, and the average waiting time reaches 6 minutes, exceeding the set congestion threshold c th = 5 minutes. Therefore, the system generates a dynamic path adjustment plan through real-time monitoring data and combines it with the path planning formula to provide a smoother alternative path for the user.

[0056] The formula for path planning is as follows:

[0057] P(x, y, t) = min(∫ A (f dist (x, y) + λ 1 ·f time (x, y, t) + λ 2 ·f usage (x, y, t))dx dy)

[0058] Where:

[0059] P(x, y, t) is the path weight function, which is used to calculate the total weight of choosing a path at time t starting from the current location (x, y). The system recommends the optimal path for the user by minimizing this weight value.

[0060] f dist (x, y) is the path distance function, which represents the physical distance from the current position to the target position. In this example, the user hopes to go from the entrance / exit of the subway station to the nearby barrier-free elevator, so the distance becomes one of the important factors for path optimization.

[0061] f time(x, y, t) is the time cost function, representing the time consumption of the user on the path. During peak hours, due to the increased congestion, the time cost is higher, so the system will try to select a path with a lower time cost.

[0062] f usage (x, y, t) represents the usage frequency of the facilities passed by this path. During the morning peak period, due to the high usage frequency of elevators and wheelchair ramps, the system will provide alternative facilities or paths with a low usage frequency for the user.

[0063] λ 1 and λ 2 are adjustment coefficients, respectively controlling the weights of the time cost and the facility usage frequency in path selection.

[0064] Now, assign specific values and calculate the path recommended by the system for the user. Assume that the user is currently located at the entrance of the subway station (x 0 , y 0 ), and needs to reach the location of the accessible elevator (x 1 , y 1 ), and the physical distance is 200 meters. Therefore, the value of f dist (x 0 , y 0 ) is 200. Assume that due to congestion, the expected time cost function f time (x 0 , y 0 , t) on the path is 10 minutes, and assume that the usage frequency of the elevator on this path f usage (x 0 , y 0 , t) is 7 times per minute, which has exceeded the usage frequency threshold u th = 5 times / minute.

[0065] The system substitutes these data into the path planning formula:

[0066] P(x 0 , y 0 , t) = min(200 + λ 1 ·10 + λ 2 ·7)

[0067] Assume that the adjustment coefficient is assigned λ 1 = 1.5 and λ 2 = 2.0, then the path weight value is:

[0068] P(x 0 , y 0 , t) = min(200 + 1.5·10 + 2.0·7) = min(200 + 15 + 14) = 229

[0069] At this time, the system found that the path weight value was relatively high, mainly because the usage frequency of the facilities on this path was too high. So the system calculated an alternative path for the user, setting that this path was through a newly built barrier-free passage beside it, and the physical distance increased to 250 meters. However, due to less usage, the time cost function was 7 minutes and the usage frequency was 3 times per minute. Substituting back into the formula:

[0070] P′(x 0 ,y 0 ,t)=min(250 + 1.5·7 + 2.0·3)=min(250 + 10.5 + 6)=266.5

[0071] Although the physical distance of the alternative path is longer, due to the lower usage frequency, the actual travel time of the user is shortened. Finally, the system provides this alternative path to the user as a temporary adjustment plan.

[0072] This path optimization method not only considers the physical distance, but also provides the most suitable travel plan for the user by comprehensively evaluating the time cost on the path and the usage frequency of the facilities. The value ranges of the adjustment coefficients λ 1 and λ 2 can be flexibly adjusted according to the actual situation. The value range of λ 1 is from 1.0 to 2.0, which is used to balance the time cost, and the value range of λ 2 is from 1.0 to 3.0 to control the influence of the usage frequency of the facilities in path selection.

[0073] During the morning rush hour, there was continuous congestion in the wheelchair access and elevators at the subway station, and the previous path adjustment strategy failed to effectively alleviate this problem. In this case, the system decided to further adopt a strategy, that is, to dynamically increase temporary mobile facilities, such as temporary ramps, handrails or signs, through a resource allocation algorithm to relieve the congestion pressure.

[0074] First of all, based on the real-time monitoring data, the system analyzes the demand situation of the current facilities and the availability of mobile facilities, and uses the resource scheduling function R(z, t) for resource allocation to decide whether to add mobile facilities. The formula is as follows:

[0075]

[0076] Among them:

[0077] R(z, t) is the resource allocation priority function, which represents the priority of adding mobile facilities at position z and time t. The larger the value, the higher the urgency of adding facilities at this position and time.

[0078] f need(z, t) is the regional demand function, representing the demand intensity of the region for mobile facilities at location z and time t. This demand function is comprehensively derived based on factors such as facility usage frequency, users' congestion perception, and usage feedback. The higher the demand, the larger the value.

[0079] Γ(f mob (z, t)) is the mobile facility availability function, reflecting the number of mobile facilities available for scheduling at location z at the current moment. If the facility resources are tight, the value of this function is small.

[0080] T is the time period for resource scheduling, indicating that the system evaluates the regional demand and facility availability within this period.

[0081] Set the current time period to 8:30 to 9:30 in the morning, and the facility demand during the peak period continues to increase within this period. Through real-time monitoring, the system finds that in a certain key barrier-free facility area (such as the main wheelchair passage) of the subway station, the demand function f need (z, t) reaches a relatively high value, approximately f need (z, t) = 8. At the same time, through data retrieval, the system knows that the number of temporary ramps and handrails in this area is limited, and the mobile facility availability function Γ(f mob (z, t)) is 0.5, indicating that there are fewer schedulable resources. Therefore, the system substitutes into the formula:

[0082]

[0083] Through calculation, the resource allocation priority R(z, t) = 4 shows that the demand in this area is relatively urgent, and the system judges that mobile facilities need to be added immediately.

[0084] Next, the system suggests temporarily adding two ramps and a sign in this area to guide wheelchair users to use the alternative path. At the same time, combined with the mobile facility availability, the system allocates the temporary ramps from other locations of the subway station to this area and ensures that the sign guides users to avoid the congested wheelchair passage and use the newly set temporary ramps.

[0085] To further test the effectiveness of the scheduling strategy, the system continuously monitors the usage of the temporary ramps. It is set that the usage frequency of the newly added ramps reaches 3 times per minute within the next 30 minutes, and the congestion situation has been significantly alleviated, and the average waiting time has been reduced to 2 minutes. At this time, through the real-time updated data, the system compares and finds that the congestion degree f c (t) has decreased significantly, and finally judges that this temporary adjustment plan has successfully solved the congestion problem.

[0086] In this case, the value of the adjustment coefficient is crucial. The regional demand function f needThe value range of (z, t) can be set from 0 to 10 to reflect the urgency of demand; while the mobile facility availability function Γ(f mob (z, t)) can have a value range between 0.1 and 1. When the value is close to 1, it indicates that there are sufficient mobile facilities available for scheduling in this area. When Γ(f mob (z, t)) is low, the system gives priority to resource allocation to ensure that the areas in greatest need of facilities can obtain additional temporary facilities in a timely manner.

[0087] Example 2:

[0088] Combined with Figure 3 , in this example, in a large shopping mall, three main types of users - wheelchair users, visually impaired people, and the elderly - need to travel from the entrance of the shopping mall to a specific area (such as the dining area) on a certain floor.

[0089] First, use the formula to generate the user's personalized preliminary path weight function. The formula is as follows:

[0090]

[0091] Where:

[0092] P u (x, y) represents the preliminary path weight value of user u at location x, y. The lower the weight value, the more suitable the path is for the user;

[0093] Ω is the area range of the entire path. The system searches for the optimal point of the path within this area range;

[0094] S u (x, y) is the slope function, which measures the slope suitability of user u at location x, y. A lower slope is more suitable for wheelchair users. Therefore, 1 - S u (x, y) is used to reflect the attractiveness of a low-slope path to the user. The smaller the slope, the higher this value;

[0095] D u (x, y) is the distance function, which represents the physical distance from the user's current location to the target location. Use to reduce the priority of long-distance paths;

[0096] U u (x, y) is the equipment usage preference function, which reflects the user's dependence on barrier-free facilities;

[0097] α 1 , α 2 , α 3 are weight coefficients, corresponding to slope, distance, and user preference respectively, and are used to balance the relative importance of these factors in path recommendation.

[0098] 1. Path generation for wheelchair users:

[0099] In this shopping mall, wheelchair users need to travel from the entrance to the dining area on the third floor. The user hopes to choose a barrier-free path with a smaller slope and smooth passage. There are two paths from the entrance to the dining area: The first is through the elevator and ramp, with a total path distance of 200 meters and a slope of 1 degree; the second is through the elevator and ordinary stairs, with a total path distance of 150 meters, but the slope is 5 degrees. Since wheelchair users are particularly sensitive to slopes, the system takes the slope as the main optimization parameter.

[0100] Substitute into the formula:

[0101] For the first path (elevator + ramp): S u (x, y) = 0.01 (1-degree slope), D u (x, y) = 200 meters. The user has a high dependence on the ramp, and set U u (x, y) = 0.9. The weight coefficients are α 1 = 0.7, α 2 = 0.2, α 3 = 0.1.

[0102] For the second path (elevator + stairs): S u (x, y) = 0.05 (5-degree slope), D u (x, y) = 150 meters. The user cannot use the stairs, so U u (x, y) = 0.1.

[0103] Calculate the weights of the two paths respectively:

[0104] The weight of the first path:

[0105]

[0106] The weight of the second path:

[0107]

[0108] Although the second path is shorter, since wheelchair users cannot use the stairs, the system excludes this path according to the user's device dependence (U u (x, y)) and finally recommends the first path (elevator + ramp).

[0109] 2. Path generation for visually impaired people:

[0110] When visually impaired people move within the same shopping mall, the system pays more attention to voice navigation and clear path support. Suppose a visually impaired person needs to move from the entrance to the same dining area. Since they rely on the voice navigation system and guiding signs, the system needs to consider the dependence on barrier-free facilities as an important factor.

[0111] Substitute into the formula:

[0112] The first path (elevator + ramp): S u (x, y) = 0.01, D u (x, y) = 200 meters, with a relatively high degree of voice navigation support. Set U u (x, y) = 0.8. The weight coefficient is α 1 = 0.4, α 2 = 0.2, α 3 = 0.4.

[0113] The second path (elevator + stairs): S u (x, y) = 0.05, D u (x, y) = 150 meters, with relatively low voice navigation support. Set U u (x, y) = 0.2.

[0114] Calculate the weights of the two paths respectively:

[0115] The weight of the first path:

[0116]

[0117] The weight of the second path:

[0118]

[0119] In this case, since visually impaired users are more dependent on voice navigation facilities, the system recommends the second path (elevator + stairs) as the optimal path.

[0120] 3. Path generation for the elderly:

[0121] For the elderly, the system needs to balance the length and convenience of the path and avoid choosing paths that are too long or too steep. Suppose the elderly hope to minimize the walking distance, so the distance becomes the primary consideration factor.

[0122] Substitute into the formula:

[0123] The first path (elevator + ramp): S u (x, y) = 0.01, D u (x, y) = 200 meters, with a relatively low demand for ramps from the elderly. Set U u (x, y) = 0.3. The weight coefficient is α 1= 0.3, α 2 = 0.5, α 3 = 0.2.

[0124] The second path (elevator + stairs): S u (x, y) = 0.05, D u (x, y) = 150 meters. The elderly cannot use the stairs. Set U u (x, y) = 0.1.

[0125] Calculate the weights of the two paths respectively:

[0126] The weight of the first path:

[0127]

[0128] The weight of the second path:

[0129]

[0130] Due to the difficulties the elderly have in using the stairs, the system finally selects the first path (elevator + ramp) as the recommended path.

[0131] In these examples, personalized path planning generates path recommendations that are most suitable for various types of users by considering the specific needs of different types of users (slope, distance, equipment dependence). By setting different weight coefficients α 1 , α 2 , α 3 and evaluating key factors (such as slope, distance, and facility dependence), the system can achieve dynamic optimization to ensure the safety, convenience, and personalization of barrier-free paths. The value range of the weight coefficient can be set as α 1 takes 0.1 - 0.7, α 2 takes 0.2 - 0.6, α 3 takes 0.1 - 0.4. The specific value depends on the user's preferences and needs for each factor.

[0132] Further introduce environmental perception data to optimize the generated personalized barrier-free path in real time. When the user (whether a wheelchair user, a visually impaired person, or an elderly person) starts to move along the recommended path, the system obtains environmental data through deployed Internet of Things devices (such as slope sensors, road surface quality monitoring devices, and weather monitoring systems). When the environmental conditions change, such as when the road surface is slippery or there is water accumulation, the system will make dynamic adjustments based on real-time data and re-recommend the optimal path.

[0133] Use the following formula to make dynamic adjustments to the path:

[0134] P e (x, y, t) = P u(x, y) + ∫ Ω (β 1 ·W(x, y, t) + β 2 ·R(x, y, t)) dx dy

[0135] Where:

[0136] P e (x, y, t) is the path weight optimized for the environment, which is dynamically adjusted based on the initial weight P u (x, y) and environmental perception data (such as weather and road surface quality).

[0137] W(x, y, t) is the weather condition function that describes the weather of the path segment at time t. Weather conditions such as rainfall, snow, or haze will increase the path weight, prompting the system to select a safer alternative path for the user.

[0138] R(x, y, t) is the road surface quality function that reflects the road surface condition of the path segment at time t, such as water accumulation, damage, etc. The worse the road surface quality, the higher the weight, prompting the system to avoid these poor road segments for the user.

[0139] β 1 and β 2 are the weight coefficients of environmental data, which are used to adjust the influence of weather and road surface quality on path selection respectively.

[0140] 1. When a wheelchair user encounters sudden weather changes:

[0141] In a shopping mall, a wheelchair user is moving along the first recommended path (elevator + ramp), and the initial path weight P u (x, y) is 0.784. Suppose the weather suddenly changes, there is heavy rain in the external area, and the ramp is slippery due to water accumulation. The system detects the deterioration of weather and road surface conditions through sensors and calculates the impact of environmental factors on the path in real time.

[0142] For the first path, suppose the value of the weather condition function W(x, y, t) is 0.3 (rainfall), and the road surface quality function R(x, y, t) is 0.5 (slippery and water accumulation). Suppose the weight coefficients of environmental data are β 1 = 0.5 and β 2 = 0.4.

[0143] Substitute into the formula:

[0144] P e (x, y, t) = 0.784 + ∫ Ω (0.5·0.3 + 0.4·0.5) dx dy = 0.784 + 0.15 + 0.2 = 1.134

[0145] Due to the influence of weather and road surface conditions, the weight of the first path has increased to 1.134, indicating that this path is no longer suitable for users to continue passing through in the current environment.

[0146] By comparing the environmental conditions of other available paths, the system finds that the area where the second path (elevator + stairs) is located is not affected by rainfall. The weather condition function W(x, y, t) is 0.1, and the road surface quality function R(x, y, t) is 0.2 (the road surface is dry and in good condition). Therefore, the system dynamically adjusts the path weight:

[0147] P e (x, y, t) = 0.6763 + ∫ Ω (0.5·0.1 + 0.4·0.2)dx dy = 0.6763 + 0.05 + 0.08 = 0.8063

[0148] At this time, the weight of the second path is 0.8063. Although the initial weight is relatively high, because it provides a safer passing environment in bad weather, the system finally recommends that the user switch to this path.

[0149] 2. An elderly person encounters damaged road surface:

[0150] Another elderly person is passing through the shopping mall and selects the recommended first path (elevator + ramp). The initial path weight is 0.3595. However, the road surface on a certain floor of the shopping mall is damaged due to construction, affecting the passing safety. The sensor monitors that the road surface quality of this section has decreased, and the system inputs this information into the dynamic path optimization formula for adjustment.

[0151] Set the road surface quality function R(x, y, t) of the section affected by the construction area. Its value is 0.6. At the same time, the weather has no change, and the weather condition function W(x, y, t) remains at 0.1. The weight coefficients β 1 = 0.3 and β 2 = 0.6 correspond to the influence of weather and road surface quality.

[0152] Calculate the new path weight:

[0153] P e (x, y, t) = 0.3595 + ∫ Ω (0.3·0.1 + 0.6·0.6)dx dy = 0.3595 + 0.03 + 0.36 = 0.7495

[0154] The path weight has increased significantly, and the system determines that the damaged road surface poses a threat to the passing safety of users.

[0155] In this case, the system recommends another path (elevator + stairs) for the elderly. The road surface quality of this path is good, and the road surface quality function R(x, y, t) is 0.2, while the weather condition function W(x, y, t) remains 0.1. After dynamic optimization:

[0156] P e (x, y, t) = 0.3083 + ∫ Ω (03·0.1 + 0.6·0.2)dx dy = 0.3083 + 0.03 + 0.12 = 0.4583

[0157] At this time, the system determines that this path is more suitable for the elderly to pass through, so it adjusts the navigation guidance.

[0158] 3. Visually impaired people face sudden changes in both weather and road surface quality:

[0159] Visually impaired people rely on the barrier-free paths recommended by the system and highly rely on voice navigation and clear guiding signs. Suppose that the road surface suddenly becomes damaged on the section where the visually impaired people are passing, and the weather also deteriorates, resulting in a slippery road surface. The system obtains the latest weather and road surface quality data through sensors and updates the path recommendation in real time.

[0160] For the first path (elevator + ramp), the weather condition function W(x, y, t) is 0.4 (light rain), and the road surface quality function R(x, y, t) is 0.5 (slippery and damaged), and the weight coefficients β 1 = 0.4 and β 2 = 0.5.

[0161] The path weight is adjusted as follows:

[0162] P e (x, y, t) = 0.717 + ∫ Ω (0.4·0.4 + 0.5·0.5)dx dy = 0.717 + 0.16 + 0.25 = 1.127

[0163] At this time, the path weight increases significantly, indicating that the current passing conditions are poor and no longer suitable for visually impaired people to pass through.

[0164] After dynamic optimization by the system, it is recommended that visually impaired people switch to another path with better weather and road surface quality. Suppose the weather condition function of this path is W(x, y, t) = 0.2, and the road surface quality function is R(x, y, t) = 0.3. The final weight is:

[0165] P e (x, y, t) = 0.4613 + ∫ Ω (0.4·0.2 + 0.5·0.3)dx dy = 0.4613 + 0.08 + 0.15 = 0.6913

[0166] Since the second path is safer, the system immediately switches the navigation and provides clear path guidance to the visually impaired.

[0167] By introducing environmental perception data, the system can dynamically adjust the barrier-free path according to real-time weather and road surface quality changes, ensuring the safety and smooth passage of users. The weight coefficients β 1 and β 2 range from 0.1 to 0.5 respectively and are adjusted according to the influence degree of environmental conditions on path recommendation.

[0168] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing the layout of barrier-free public facilities, characterized in that The following steps are involved: S1. Demand forecast and site selection optimization for barrier-free facilities: S1.

1. Use multi-source data fusion technology to integrate different data types including static geographic data and dynamic traffic and passenger flow data to build a demand forecasting model; S1.

2. Use a machine learning algorithm based on a regression model to analyze historical usage data and crowd behavior to predict the future demand distribution of barrier-free facilities; S1.

3. Develop a site selection optimization algorithm by combining the Geographic Information System (GIS) and the demand forecasting model; the site selection optimization algorithm includes calculations of terrain, transportation convenience, distance and coverage of existing facilities; and generates a facility layout plan through a simulated annealing algorithm; S2. Dynamic monitoring and intelligent layout: S2.

1. Deploy sensors and cameras in barrier-free facilities including wheelchair access, elevators and blind paths to collect real-time data on the use, congestion and damage of facilities; transmit data to the central control system through the Internet of Things technology to monitor the status of equipment in real time; S2.

2. Based on real-time monitoring data, by analyzing the frequency of use and congestion level of facilities, generate temporary adjustment plans including adjusting route guidance and temporarily increasing the layout of mobile facilities; S2.

3. Use monitoring data to predict facility failures and adopt predictive maintenance technology to predict potential facility failures by analyzing historical data and real-time usage data; S3, Personalized barrier-free path planning and adaptive navigation: S3.

1. Generate personalized barrier-free paths based on the user's specific needs, physical condition and travel preferences; then update and optimize path recommendations in real time based on user data and environmental perception data including slope and road conditions; S3.

2. Provide real-time navigation through mobile applications or wearable devices, and adaptively adjust the user's travel path based on dynamic data including road construction and facility congestion.

2. The method for optimizing the layout of barrier-free public facilities according to claim 1, characterized in that The strategy for generating a temporary adjustment plan includes: using a machine learning model to analyze the historical usage data of the facility and combining it with real-time data to predict the future usage of the facility; the usage frequency of the facility f u (t) and congestion level f c (t) is recorded and calculated, and the threshold is used to determine whether it exceeds the normal range to determine whether adjustment is needed: Among them, Φ(t) represents the change in facility usage pressure, which is used to measure the change in facility usage and congestion status within time t; f u (t) is the facility usage frequency function, which describes the usage density of the facility at each moment and indicates how frequently the facility is used; f c (t) is the facility congestion function, which indicates the congestion of the facility per unit time and reflects whether the facility is in an overused state; and is the rate of change of the frequency of use and the degree of congestion, indicating the speed of change of the frequency of use and the degree of congestion of the facility within the time τ; Θ(x) is a step function, which is used to determine whether the frequency of use or the degree of congestion of the facility exceeds a certain threshold; when x>0, the value of the step function is 1, otherwise it is 0; u th and c th They are the usage frequency threshold and congestion level threshold respectively. When the usage frequency or congestion level exceeds the threshold, the dynamic adjustment plan is triggered. By calculating Φ(t), based on real-time monitoring data, it is determined whether the usage pressure of the facility is too high and whether adjustment is needed.

3. The method for optimizing the layout of barrier-free public facilities according to claim 2, characterized in that The strategy for generating a temporary adjustment plan includes: generating a temporary path adjustment plan when it is detected that the frequency of facility use or the degree of congestion exceeds a threshold; providing users with alternative paths by real-time monitoring of facility use and combining a path planning algorithm; and optimizing path planning by combining path distance, time cost and facility use.

4. The method for optimizing the layout of barrier-free public facilities according to claim 3, characterized in that The strategy for generating a temporary adjustment plan includes: when the barrier-free facilities in a certain area are continuously congested or frequently used, and the temporary path adjustment cannot effectively solve the problem, it is recommended to temporarily add mobile facilities including temporary ramps, handrails, and signs through the resource allocation algorithm; according to the real-time monitoring data, the current demand for the facilities is analyzed, and through the resource scheduling function R(z,t), it is dynamically determined whether additional mobile facilities are needed, and the deployment location of the facilities is optimized: Where R(z,t) is the resource allocation priority function, which is used to measure the priority of adding mobile facilities at location z and time t; f need (z,t) represents the regional demand function, which describes the intensity of regional demand for mobile facilities at location z and time t; Γ(f mob (z,t)) is the mobile facility availability function, which reflects the number of mobile facilities that can be dispatched at location z at time t; T is the time period of resource scheduling, which means that the regional demand and facility availability are evaluated within the period T.

5. The method for optimizing the layout of barrier-free public facilities according to claim 1, characterized in that The method for generating a personalized barrier-free path includes: generating a personalized path that meets the needs according to the types of users including wheelchair users, visually impaired people, and elderly people; and constructing a preliminary personalized path weight function through the following formula: Among them, P u (x, y) represents the initial path weight value of user u at position x, y, and the path recommended for the user based on the user's personalized needs and travel preferences; Ω represents the area range of the entire path, and the optimal point of the path is found within the range; S u (x, y) is the slope function, which measures the suitability of the slope for user u at position x, y; a low slope is suitable for wheelchair users, 1-S u (x,y) is used to reflect the impact of slope on users. The smaller the slope, the higher the value and the better the path. D u (x, y) is the distance function, which represents the physical distance from the user's current location to the target location; the formula uses To reduce the priority of long paths; U u (x, y) is the device usage preference function of user u, which reflects the user's reliance on barrier-free facilities during travel; α1, α2, and α3 are weight coefficients, corresponding to slope, distance, and user preference, respectively, which are used to balance the importance of factors in path recommendation.

6. The method for optimizing the layout of barrier-free public facilities according to claim 5, characterized in that The method for generating a personalized barrier-free path includes: after generating a preliminary path, acquiring environmental perception data in real time through deployed Internet of Things devices and including slope sensors, road quality monitoring equipment, and weather monitoring, monitoring the environmental status of the path, and dynamically optimizing the path based on real-time data; when environmental conditions change, adjusting the path.

Images