A method and system for fusion and collaborative control of PIS and PA based on multi-source sensing

Through multi-source sensing data processing and fuzzy control optimization, the coordinated control of the PIS and PA systems was realized, solving the problems of inconsistent information dissemination and low resource utilization, and improving the intelligence and adaptability of the rail transit information dissemination system.

CN122386845BActive Publication Date: 2026-11-17BEIJING QINGYUAN ZHIXIN TECH CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202610575439.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-04-28
Publication Date
2026-11-17
Estimated Expiration
2046-04-28

AI Technical Summary

Technical Problem

Existing passenger information systems (PIS) and public address systems (PA) in rail transit suffer from inconsistent information dissemination, low system resource utilization, and high operational complexity. Furthermore, they lack the ability to comprehensively perceive and make intelligent decisions based on multi-source environmental information, making it difficult to achieve adaptive optimization and multi-system collaborative control.

Method used

By acquiring multi-source sensing data, performing spatiotemporal alignment and heterogeneous feature encoding, constructing scene feature vectors, and combining fuzzy control and improved Thompson sampling, multi-scene recognition and conflict resolution are achieved, the optimal collaborative control strategy is determined, and the linkage control and adaptive adjustment of the PIS and PA systems are carried out.

Benefits of technology

It has improved the intelligence level of the rail transit information release system, realized adaptive optimization of information release, improved decision-making accuracy and system stability, and enhanced real-time response capability and resource utilization efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122386845B_ABST
    Figure CN122386845B_ABST
Patent Text Reader

Abstract

This invention provides a method and system for fusion and collaborative control of PIS and PA based on multi-source sensing, relating to the field of traffic control technology. The method includes acquiring multi-source sensing data from the rail transit operating environment; performing spatiotemporal alignment and heterogeneous feature encoding on the multi-source sensing data to construct scene feature vectors; performing multi-scene identification and conflict resolution based on the scene feature vectors to determine the target scene; constructing a control strategy space for the target scene containing PIS and PA linkage parameters; determining the optimal collaborative control strategy through a strategy optimization mechanism combining fuzzy control and improved Thompson sampling; and performing linkage control of the PIS and PA systems according to the optimal collaborative control strategy, with adaptive adjustments based on execution feedback. This invention solves the problems of insufficient collaboration, reliance on fixed rules, and lack of adaptive capabilities in existing PIS and PA systems.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of traffic control technology, and more specifically, to a method and system for coordinated control of PIS and PA based on multi-source sensing. Background Technology

[0002] With the rapid development of urban rail transit systems, passenger information systems (PIS) and public address systems (PA) play a crucial role in station operation and information dissemination as key passenger-oriented systems. Traditional PIS systems are mainly used for displaying images, videos, and text information, while PA systems are used for voice broadcasts. The two are usually built independently and controlled separately, which leads to problems such as inconsistent information dissemination, low system resource utilization, and high operation and maintenance complexity.

[0003] With the development of digital and network technologies, PIS and PA systems are gradually evolving towards a fully digital architecture based on IP networks, providing the technological foundation for converged deployment. However, most existing converged solutions still remain at the level of simple integration at the device level or linkage control based on preset rules, lacking the ability to comprehensively perceive and make intelligent decisions based on multi-source environmental information.

[0004] In actual operation, the station environment is highly dynamic. Factors such as changes in passenger flow, train operation status, and environmental noise fluctuations can all affect the effectiveness of information dissemination. Existing technologies typically employ fixed strategies or manual intervention for adjustments, making it difficult to achieve adaptive optimization of information dissemination, resulting in poor broadcast effects or mismatched information display. Furthermore, when multiple operational scenarios occur simultaneously, existing systems lack effective conflict resolution and strategy optimization mechanisms, failing to achieve collaborative control and dynamic scheduling among multiple systems.

[0005] Therefore, there is an urgent need for a method that can perform scene recognition based on multi-source sensing information and realize collaborative control and adaptive adjustment of PIS and PA systems to improve the intelligence level and service quality of rail transit information release systems. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for coordinated control of PIS and PA based on multi-source sensing, so as to improve the above-mentioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0007] Firstly, this application provides a method for coordinated control of PIS and PA based on multi-source sensing, including:

[0008] Acquire multi-source sensing data in the rail transit operation environment, including train operation status, passenger flow density, environmental status, and manual trigger commands;

[0009] The multi-source sensing data is subjected to spatiotemporal alignment and heterogeneous feature encoding to construct a scene feature vector;

[0010] Based on the scene feature vectors, multi-scene identification and conflict resolution are performed to determine the target scene;

[0011] A control strategy space containing PIS and PA linkage parameters is constructed for the target scenario. The optimal cooperative control strategy is determined by a strategy optimization mechanism that combines fuzzy control and improved Thompson sampling.

[0012] The PIS system and PA system are controlled in conjunction with the optimal collaborative control strategy, and adaptive adjustments are made based on the execution feedback.

[0013] Secondly, this application also provides a PIS and PA fusion and cooperative control system based on multi-source sensing, including:

[0014] The acquisition module is used to acquire multi-source sensing data in the rail transit operation environment, including train operation status, passenger flow density, environmental status, and manual triggering commands.

[0015] The processing module is used to perform spatiotemporal alignment and heterogeneous feature encoding on the multi-source sensing data to construct scene feature vectors;

[0016] The identification module is used to perform multi-scene identification and conflict resolution based on the scene feature vector to determine the target scene;

[0017] The optimization module is used to construct a control strategy space containing PIS and PA linkage parameters for the target scenario, and determine the optimal cooperative control strategy through a strategy optimization mechanism that combines fuzzy control and improved Thompson sampling.

[0018] The control module is used to perform coordinated control of the PIS system and PA system according to the optimal cooperative control strategy, and to make adaptive adjustments based on the execution feedback.

[0019] The beneficial effects of this invention are as follows:

[0020] (1) This invention constructs scene feature vectors based on multi-source sensing data and combines them with multi-scene recognition and conflict resolution mechanisms to achieve refined modeling and accurate identification of complex operating environments in rail transit. Compared with traditional control methods based on a single data source or fixed rules, this invention can comprehensively consider multi-dimensional information such as train operating status, passenger flow distribution, environmental noise, and manual instructions. In the case of multiple concurrent scenarios, the target scenario is determined through a scenario matching degree and priority fusion mechanism, which effectively avoids the problems of information release conflicts and control strategy confusion, thereby significantly improving the decision-making accuracy and stability of the system in complex dynamic environments.

[0021] (2) Furthermore, this invention constructs a control strategy space containing PIS and PA linkage parameters, and integrates fuzzy control and improved Thompson sampling strategy optimization mechanisms to achieve adaptive learning and dynamic optimization of the control strategy. By using a multi-index comprehensive reward function to uniformly evaluate information coverage, broadcast clarity, and system load, and combining this with execution feedback to form a closed-loop adjustment, the system can continuously optimize the control strategy according to environmental changes. This ensures the consistency and effectiveness of information dissemination, while also improving the system's real-time response capability and resource utilization efficiency, overcoming the shortcomings of existing technologies such as reliance on manual intervention, lack of adaptive capability, and insufficient collaborative control.

[0022] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the PIS and PA fusion collaborative control method based on multi-source sensing described in this embodiment of the invention;

[0025] Figure 2 This is a schematic diagram of the structure of the PIS and PA fusion collaborative control device based on multi-source sensing described in an embodiment of the present invention.

[0026] The diagram is labeled as follows: 800, PIS and PA fusion collaborative control device based on multi-source sensing; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0028] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0029] Example 1:

[0030] This embodiment provides a method for coordinated control of PIS and PA based on multi-source sensing.

[0031] This invention is applicable to various complex operational scenarios in rail transit stations, including peak passenger flow management, train operation anomalies, sudden emergency events, high-noise environments, and concurrent conflicts in multiple scenarios. Through the fusion and analysis of multi-source sensing information, it achieves accurate identification of different operational scenarios and, based on a scenario-driven mechanism, performs coordinated control and adaptive adjustment of the PIS and PA systems, thereby improving the consistency, timeliness, and effectiveness of information dissemination. Furthermore, this invention can be extended to scenarios such as transfer guidance optimization, low-passenger-flow energy-saving control, and personalized services, demonstrating good versatility and scalability.

[0032] See Figure 1 The figure shows that the method includes steps S1, S2, S3, S4, and S5.

[0033] Step S1: Acquire multi-source sensing data in the rail transit operation environment, including train operation status, passenger flow density, environmental status, and manual triggering commands;

[0034] In this step, the train operation status is obtained through interface communication with the train control system or the train dispatching system. The train operation status includes the train location and section information, arrival and departure time and predicted arrival time, train operation status (normal, delayed, turnaround, suspended) and the opening / closing status of the doors.

[0035] Passenger flow density is obtained through station passenger flow monitoring equipment. Passenger flow density includes real-time passenger flow, passenger flow density distribution information, and passenger flow statistics in and out of the station hall, platform and passage areas.

[0036] Environmental conditions are acquired through monitoring equipment, including ambient noise intensity, temperature and humidity information, smoke concentration, air quality indicators, and equipment operating status information. Temperature, humidity, and air quality data are obtained through an environmental control system, while smoke concentration monitoring data is obtained through a fire prevention and firefighting system.

[0037] Manual triggering instructions are obtained through the station control terminal or dispatching system. These manual triggering instructions include emergency broadcast triggering instructions, temporary information release instructions, and emergency response instructions.

[0038] Step S2: Perform spatiotemporal alignment and heterogeneous feature encoding on the multi-source sensing data to construct a scene feature vector;

[0039] This step addresses the issues of asynchronicity, spatial dispersion, and inconsistency in units of multi-source sensing data. By using spatiotemporal alignment and structured encoding, a unified scene feature vector is constructed, enabling a computable representation of complex operating environments.

[0040] Step S2 includes:

[0041] Step S21: Synchronize the train operation status, passenger flow density, environmental status, and manual trigger commands according to a unified time base to obtain a time-aligned multi-source data sequence;

[0042] In this step, since the sampling frequency and timestamps of different data sources are different, such as train operation status data and passenger flow density data being updated at the second level, while environmental status data can adopt a custom continuous sampling mode, it is necessary to map various types of data to a unified time base through a time alignment mechanism to ensure that multi-source data have consistent semantics at the same time.

[0043] Step S22: Map the multi-source data sequence to different regions according to the station's spatial structure to obtain a subset of multi-source data organized by spatial regions;

[0044] In this step, based on the station's spatial structure (such as concourse level, platform level, passageway, etc.), the multi-source data sequence is mapped to the corresponding spatial area, so that different types of data have a unified spatial semantics, thereby reflecting the operational status of the local area.

[0045] Step S23: Perform dimensional unification processing on the multi-source data subset;

[0046] In this step, since different sensory data have different physical dimensions (such as passenger flow in number of people, noise in decibels, and time in seconds), they need to be normalized to transform them into dimensionless features, thereby preventing a certain type of feature from dominating the subsequent calculation.

[0047] Step S24: Perform feature structure encoding on a subset of multi-source data based on dimension-unified processing to obtain a multi-dimensional feature sub-vector corresponding to each spatial region;

[0048] In this step, after unifying the dimensions, the multi-source data is further structured and encoded, transforming the raw data into a multidimensional representation that includes state features, statistical features, and trend features. State features reflect the current operational state, statistical features characterize the distribution characteristics within a certain time window, and trend features describe the dynamic evolution of the environment, thereby enhancing the ability to represent changes in complex scenarios.

[0049] Step S25: Combine all multidimensional feature vectors to construct the scene feature vector.

[0050] Step S3: Based on the scene feature vector, perform multi-scene identification and conflict resolution to determine the target scene;

[0051] Understandably, when multiple scenarios such as peak passenger flow, train delays, high noise levels, and emergencies occur concurrently at a station, the system lacks an effective scenario identification and conflict resolution mechanism, easily leading to conflicting control objectives and chaotic information dissemination. Therefore, this step constructs a candidate scenario set and its reference feature representation, utilizes the matching degree between scenario feature vectors and reference features to achieve parallel identification of multiple scenarios, and introduces priority constraints and scenario conflict relationships to filter and optimize candidate scenarios, thereby determining the target scenario. This overcomes the limitations of traditional single-scenario identification, effectively handles multi-scenario concurrency and conflict issues, and improves the accuracy and stability of decision-making in complex environments.

[0052] Step S3 includes:

[0053] Step S31: Define multiple candidate scenarios based on the operational needs of rail transit;

[0054] In this step, based on the actual operational needs of rail transit, several typical candidate scenarios are predefined, and structured reference features are constructed for each candidate scenario as the basis for scenario identification. For example, candidate scenarios include high passenger flow congestion scenarios, low passenger flow idle scenarios, train delay scenarios, train dense entry scenarios, high noise interference scenarios, and emergency evacuation scenarios.

[0055] Step S32: Match the scene feature vector with the structured reference features of each candidate scene to obtain the scene matching degree;

[0056] In this step, the scene matching degree of each candidate scene is obtained by calculating the similarity between the current scene feature vector and the reference features of each candidate scene. This scene matching degree is used to reflect the degree of closeness between the current environmental state and each predefined scene.

[0057] The formula for calculating scene matching degree is:

[0058] ;

[0059] In the formula, Representation and candidate scenarios Scene matching degree Represents an exponential function. Represents the scene feature vector. Representing candidate scenarios The structured reference features, express transpose, This represents the feature weight matrix.

[0060] The value ranges from (0,1), with larger values ​​indicating a higher degree of matching. It is a positive definite symmetric feature weight matrix, with each dimension's weight element being greater than 0, used to assign differentiated importance to different feature dimensions.

[0061] Step S33: Select candidate scenes with a scene matching degree greater than a preset threshold as the candidate set;

[0062] In this step, multiple candidate scenes with a matching degree higher than a preset threshold are selected to form a candidate scene set, thereby allowing multiple scenes to be identified at the same time and enhancing the ability to express complex environments.

[0063] Step S34: Based on scene priority and scene matching degree, resolve conflicts in the candidate set to determine the target scene.

[0064] In this step, the rail transit environment is a multi-dimensional state superposition system, where different dimensions can simultaneously satisfy multiple scenario conditions. For example, during the morning rush hour, high passenger flow and train delays may coexist; when trains are densely entering stations, high noise levels may also occur; and in the event of an emergency, emergency scenarios may coexist with regular operation scenarios. Furthermore, manually triggered commands may coexist with scenarios automatically identified by the system, leading to conflicts in control objectives or resource competition. Therefore, when multiple candidate scenarios exist, conflict resolution based on scenario priority and conflict relationships is necessary to determine the final target scenario.

[0065] Specifically, based on the semantic attributes and control strategy characteristics of candidate scenarios, it is determined whether there are semantic conflicts, control strategy conflicts, or resource competition conflicts between scenarios, and a candidate set is constructed. Conflict matrix:

[0066] ;

[0067] In the formula, Representing candidate scenarios With candidate scenarios The conflict value, and Representing candidate scenarios With candidate scenarios .

[0068] Based on the existing scene matching degree, a priority system is introduced:

[0069] ;

[0070] In the formula, Representing candidate scenarios The overall evaluation value, and All represent weighting coefficients. Representation and candidate scenarios Scene matching degree Representing candidate scenarios Priority.

[0071] according to After sorting the candidate scenarios in descending order, a greedy selection mechanism combined with a conflict matrix is ​​used to filter and obtain the final target scenario. It should be noted that the target scenario can consist of a main scenario or a main scenario paired with a conflict-free auxiliary scenario. The main scenario is used to determine the control strategy framework, and the auxiliary scenario is used to fine-tune the control parameters. Specifically, the main scenario is the candidate scenario with the highest comprehensive evaluation value, and the auxiliary scenario is a candidate scenario that does not conflict with the main scenario and meets the matching degree condition. In the subsequent generation of the control strategy, a control strategy space is constructed based on the main scenario, and the parameters in the control strategy space are adjusted in conjunction with the auxiliary scenario, thereby achieving the fusion and utilization of multi-scenario information and improving the adaptability and accuracy of the control strategy.

[0072] Step S4: Construct a control strategy space for the target scenario that includes the linkage parameters of PIS and PA, and determine the optimal cooperative control strategy through a strategy optimization mechanism that combines fuzzy control and improved Thompson sampling.

[0073] Understandably, existing PIS and PA systems rely solely on simple preset rules or manual adjustments for control, lacking intelligent strategy optimization capabilities and struggling to adapt to dynamic uncertainties such as passenger flow and noise, resulting in poor information dissemination effectiveness. This step first defines the linkage control parameters of PIS and PA and constructs a control strategy space that conforms to scenario constraints. Then, it adapts to the fuzzy characteristics of the scene through fuzzy inference, combines improved Thompson sampling to consider both historical control effects and new strategy exploration, and utilizes a multi-index reward function to complete the optimal strategy selection. This effectively compensates for the shortcomings of traditional solutions, such as lack of intelligent decision-making and poor control performance, achieving precise collaborative adaptation between PIS and PA.

[0074] Step S4 includes:

[0075] Step S41: Define the PIS control parameter set and the PA control parameter set;

[0076] The PIS control parameter set includes at least the information display content, display priority, and switching frequency. The PA control parameter set includes at least the broadcast volume, number of trigger broadcasts, broadcast content, and repetition mechanism. The number of trigger broadcasts indicates the number of consecutive broadcasts when the trigger condition is met and the trigger event occurs once. The repetition mechanism is the anti-shake interval between adjacent trigger broadcasts.

[0077] Step S42: Combine the PIS control parameter set and the PA control parameter set, and determine the constraints through the set of constraints of the target scenario to construct a control strategy space that satisfies the scenario constraints;

[0078] In this step, the PIS control parameter set and the PA control parameter set are combined to form a candidate control strategy set. Then, based on the main scenario, constraints are constructed for the control strategies, and the candidate control strategies are screened to obtain an initial strategy set that meets the requirements of the main scenario. Furthermore, the control parameters in the initial strategy set are modulated and screened in conjunction with the constraints of the auxiliary scenario to obtain a control strategy space that simultaneously satisfies the constraints of both the main and auxiliary scenarios. This approach achieves the dominant constraint of the main scenario on the control strategy and the refined adjustment of the control parameters by the auxiliary scenario, thereby improving the adaptability of the control strategy to complex multi-scenario environments.

[0079] Step S43: Perform fuzzy reasoning based on scene feature vectors to obtain the membership degree of each control strategy in the control strategy space that satisfies scene constraints;

[0080] In this step, the scene feature vector is actually a multi-dimensional feature sub-vector corresponding to multiple spatial regions, and each multi-dimensional feature sub-vector includes multiple scene features. The scene feature vector is then fuzzified, mapping each scene feature to a corresponding fuzzy linguistic variable, and the membership degree of the scene features is calculated using a pre-constructed membership function.

[0081] Specifically, fuzzy linguistic variables are defined for each scene feature, such as low / medium / high passenger flow and weak / medium / strong noise. Membership function ,in, express right membership function, Indicates the first Each scene feature This represents a fuzzy set. For example, a fuzzy set can represent high passenger flow or low passenger flow.

[0082] Then, a fuzzy rule base is constructed, which describes the adaptation relationship between different scene feature combinations and control strategies.

[0083] Specifically, the fuzzy rule is represented as:

[0084] ;

[0085] In the formula, Indicates the first A fuzzy rule, and Representing the first scene feature and the second scene feature respectively. Each scene feature Indicates the first The fuzzy set corresponding to the first scene feature in the fuzzy rule. Indicates the first The first fuzzy rule A fuzzy set corresponding to each scene feature Indicates the first One control strategy.

[0086] The fuzzy rule base is actually a pre-built set of rules. Different scenarios participate in fuzzy inference calculations by activating different subsets of rules. Specifically, for a target scenario, a subset of rules related to that scenario is selected from the fuzzy rule base, the rule activation degree is calculated, and the control strategies are aggregated and calculated in combination with the rule weights to obtain the membership degree of each control strategy.

[0087] Specifically, the activation level is:

[0088] ;

[0089] ;

[0090] In the formula, Indicates the first Fuzzy rules in spatial regions activation level, Indicates the number of scene features. express right membership function, Representing a spatial region Next Each scene feature Indicates the first The first fuzzy rule A fuzzy set corresponding to each scene feature Indicates the first The activation degree of a fuzzy rule. Indicates the total number of spatial regions. Representing a spatial region Regional weights.

[0091] If the target scenario is a single scenario, then directly calculate the membership degree of each control policy:

[0092] ;

[0093] In the formula, Indicates the first Membership degree of each control strategy Indicates the number of fuzzy rules. Indicates the first The activation degree of a fuzzy rule. Indicates the first The fuzzy rule for the first Support for each control strategy.

[0094] in, It is set based on expert experience or historical operational experience to characterize the adaptability of different control strategies to the current scenario.

[0095] If the target scenario is a primary scenario and an auxiliary scenario, then the primary membership degree and auxiliary membership degree are calculated separately, and the membership degree is obtained through weighted fusion, thereby realizing the strategy adaptability evaluation under multiple scenario conditions.

[0096] Step S44: Based on the synergistic control effect of PIS and PA, construct a multi-indicator comprehensive reward function;

[0097] In this step, a set of evaluation indicators is constructed, including information coverage, broadcast clarity, information consistency, passenger response, and system load. Each indicator is normalized to obtain a multi-indicator comprehensive reward function.

[0098] ;

[0099] In the formula, Indicates the first individual control strategies The comprehensive reward value based on multiple indicators. Indicates the number of evaluation indicators. Indicates the first The weighting coefficients of each evaluation indicator, express Next The normalized value of each evaluation indicator.

[0100] Step S45: Select the optimal collaborative control strategy by using a multi-index comprehensive reward function and membership degree.

[0101] Step S45 includes:

[0102] Step S451: Generate decay weights for historical execution data using the time decay factor;

[0103] In this step, the attenuation weight is:

[0104] ;

[0105] In the formula, Representing historical moments The decay weight, Indicates the time decay factor. Indicates the current moment.

[0106] Step S452: The historical execution data is weighted and calculated using the decay weight and multi-index comprehensive reward function to obtain the historical reward value of each control strategy;

[0107] In this step, the historical reward value is:

[0108] ;

[0109] In the formula, Indicates the first A control strategy at a historical moment Historical reward value, Representing historical moments The decay weight, Indicates the first individual control strategies At a historical moment The comprehensive reward value based on multiple indicators.

[0110] Step S453: Construct the posterior reward distribution for each control policy using historical reward values;

[0111] In this step, the posterior reward distribution of the control strategy is constructed by combining the historical reward values ​​at historical execution times with a normal distribution.

[0112] Step S454: Randomly sample each control policy based on the posterior reward distribution to obtain the sampled reward value of the control policy;

[0113] Step S455: Construct an evaluation value for the control strategy based on the sampled reward value and the membership degree;

[0114] In this step, the evaluation value of the control strategy is:

[0115] ;

[0116] In the formula, Indicates the first The evaluation value of each control strategy Indicates the trade-off coefficient. Indicates the first The sampled reward value of each control strategy. Indicates the first Membership degree of each control strategy.

[0117] Step S456: Obtain the optimal cooperative control strategy from the control strategy space based on the evaluation value.

[0118] Understandably, standard Thompson sampling relies solely on random sampling based on the posterior reward distribution, failing to integrate prior knowledge of the scenario and fuzzy constraints. Furthermore, it cannot adapt to non-stationary environments and the demands of multi-indicator collaborative optimization, easily leading to policy conflicts, over- or under-exploration, and insufficient real-time performance in dynamic rail transit operation scenarios. The improved Thompson sampling designed in this step, by integrating policy membership degrees obtained through fuzzy inference, introducing historical rewards with time decay weights, and constructing a multi-indicator comprehensive reward function, achieves a synergistic improvement in scenario adaptability and policy optimization capabilities. It can stably output the optimal control strategy in highly dynamic, multi-constraint, and multi-objective PIS and PA fusion collaborative control scenarios, better meeting the requirements of practical engineering applications.

[0119] Step S5: Perform linkage control on the PIS system and PA system according to the optimal collaborative control strategy, and make adaptive adjustments based on the execution feedback.

[0120] Step S5 includes:

[0121] Step S51: Perform coordinated control of the PIS system and PA system according to the optimal cooperative control strategy;

[0122] Step S52: Update the scene feature vector by executing the feedback data, and redetermine the target scene based on the updated scene feature vector;

[0123] Step S53: If the redefined target scenario has not changed, then the control parameters of the optimal cooperative control strategy are dynamically adjusted;

[0124] Step S54: If the redefined target scenario changes, the control strategy space is reconstructed, and the next round of optimal collaborative control strategy generation is performed.

[0125] In this step, the optimal collaborative control strategy is parsed into specific control commands for the PIS and PA systems, and synchronously distributed via the network to achieve coordinated control of information display and voice broadcasting. Simultaneously, by collecting multi-dimensional feedback data such as system execution status, environmental changes, and passenger responses, the control effect is quantitatively evaluated, and the feedback information is fused to update the scene feature vector, thereby achieving a dynamic characterization of the current operating environment.

[0126] Based on this, the updated scene feature vector is re-identified to determine whether the target scene has changed: when the scene has not changed, the parameters of the current control strategy are adaptively fine-tuned to achieve continuous optimization and stable control; when the scene changes, the control strategy space is reconstructed and the global optimization mechanism is triggered to generate the optimal collaborative control strategy adapted to the new scene, forming a closed-loop adaptive control process, which effectively improves the system's response capability and control effect in complex dynamic environments.

[0127] Example 2:

[0128] This embodiment provides a PIS and PA fusion and collaborative control system based on multi-source sensing, the system comprising:

[0129] The acquisition module is used to acquire multi-source sensing data in the rail transit operation environment, including train operation status, passenger flow density, environmental status, and manual triggering commands.

[0130] The processing module is used to perform spatiotemporal alignment and heterogeneous feature encoding on the multi-source sensing data to construct scene feature vectors;

[0131] The identification module is used to perform multi-scene identification and conflict resolution based on the scene feature vector to determine the target scene;

[0132] The optimization module is used to construct a control strategy space containing PIS and PA linkage parameters for the target scenario, and determine the optimal cooperative control strategy through a strategy optimization mechanism that combines fuzzy control and improved Thompson sampling.

[0133] The control module is used to perform coordinated control of the PIS system and PA system according to the optimal cooperative control strategy, and to make adaptive adjustments based on the execution feedback.

[0134] The processing module includes:

[0135] The first processing unit is used to synchronously process the train operation status, passenger flow density, environmental status and manual triggering commands according to a unified time base to obtain a time-aligned multi-source data sequence.

[0136] The mapping unit is used to map multi-source data sequences according to the station's spatial structure to obtain multi-source data subsets organized by spatial regions.

[0137] The second processing unit is used to perform dimensional unification processing on multi-source data subsets.

[0138] The encoding unit is used to perform feature structure encoding on a subset of multi-source data based on dimensional uniform processing, to obtain a multi-dimensional feature sub-vector corresponding to each spatial region;

[0139] The combination unit is used to combine all multidimensional feature vectors to construct the scene feature vector.

[0140] The identification module includes:

[0141] The first definition unit is used to define multiple candidate scenarios based on the operational needs of rail transit;

[0142] The matching unit is used to match the scene feature vector with the structured reference features of each candidate scene to obtain the scene matching degree;

[0143] The filtering unit is used to select candidate scenes with a scene matching degree greater than a preset threshold as a candidate set.

[0144] The conflict resolution unit is used to resolve conflicts in the candidate set based on scene priority and scene matching degree to determine the target scene.

[0145] The optimization module includes:

[0146] Define the unit, used to define the PIS control parameter set and the PA control parameter set;

[0147] The first building unit is used to combine the PIS control parameter set and the PA control parameter set, and to determine the constraints through the set of constraints of the target scenario, thereby constructing a control strategy space that satisfies the scenario constraints.

[0148] The inference unit is used to perform fuzzy inference based on scene feature vectors to obtain the membership degree of each control strategy in the control strategy space that satisfies scene constraints.

[0149] The second building unit is used to construct a multi-indicator comprehensive reward function based on the synergistic control effect of PIS and PA;

[0150] The selection unit is used to select the optimal collaborative control strategy by combining a multi-index reward function and membership degree.

[0151] It should be noted that the specific methods by which each module performs operations in the system described in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0152] Example 3:

[0153] Corresponding to the above method embodiments, this embodiment also provides a PIS and PA fusion collaborative control device based on multi-source sensing. The PIS and PA fusion collaborative control device based on multi-source sensing described below and the PIS and PA fusion collaborative control method based on multi-source sensing described above can be referred to in correspondence.

[0154] Figure 2 This is a block diagram illustrating a PIS and PA fusion collaborative control device 800 based on multi-source sensing, according to an exemplary embodiment. Figure 2 As shown, the PIS and PA fusion collaborative control device 800 based on multi-source sensing may include: a processor 801 and a memory 802. The PIS and PA fusion collaborative control device 800 based on multi-source sensing may also include one or more of the following: a multimedia component 803, an I / O interface 804, and a communication component 805.

[0155] The processor 801 controls the overall operation of the PIS and PA fusion collaborative control device 800 based on multi-source sensing to complete all or part of the steps in the aforementioned PIS and PA fusion collaborative control method based on multi-source sensing. The memory 802 stores various types of data to support the operation of the PIS and PA fusion collaborative control device 800. This data may include, for example, instructions for any application or method operating on the PIS and PA fusion collaborative control device 800, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the multi-source sensing-based PIS and PA fusion collaborative control device 800 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or one or more combinations thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, and an NFC module.

[0156] In an exemplary embodiment, the PIS and PA fusion collaborative control device 800 based on multi-source sensing can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the above-described PIS and PA fusion collaborative control method based on multi-source sensing.

[0157] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, these program instructions implement the steps of the multi-source sensing-based PIS and PA fusion collaborative control method described above. For example, the computer-readable storage medium may be the memory 802 including the program instructions described above. These program instructions may be executed by the processor 801 of the multi-source sensing-based PIS and PA fusion collaborative control device 800 to complete the multi-source sensing-based PIS and PA fusion collaborative control method described above.

[0158] Example 4:

[0159] Corresponding to the above method embodiments, this embodiment also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the above-described method for PIS and PA fusion and collaborative control based on multi-source sensing.

[0160] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the multi-source sensing-based PIS and PA fusion collaborative control method described in the above method embodiments.

[0161] Specifically, the readable storage medium can be a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or any other readable storage medium capable of storing program code.

[0162] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0163] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for fusion and collaborative control of PIS and PA based on multi-source sensing, characterized in that, include: Acquire multi-source sensing data in the rail transit operation environment, including train operation status, passenger flow density, environmental status, and manual trigger commands; The multi-source sensing data is subjected to spatiotemporal alignment and heterogeneous feature encoding to construct a scene feature vector; Based on the scene feature vectors, multi-scene identification and conflict resolution are performed to determine the target scene; A control strategy space containing PIS and PA linkage parameters is constructed for the target scenario. The optimal cooperative control strategy is determined by a strategy optimization mechanism that combines fuzzy control and improved Thompson sampling. The PIS system and PA system are linked and controlled according to the optimal collaborative control strategy, and adaptive adjustments are made based on the execution feedback. The construction of a control strategy space for the target scene, including PIS and PA linkage parameters, and the determination of the optimal cooperative control strategy through a strategy optimization mechanism combining fuzzy control and improved Thompson sampling, includes: Define the PIS control parameter set and the PA control parameter set; By combining the PIS control parameter set and the PA control parameter set, and by using the constraint condition set of the target scenario to determine the constraints, a control strategy space that satisfies the scenario constraints is constructed. Fuzzy reasoning is performed based on scene feature vectors to obtain the membership degree of each control strategy in the control strategy space that satisfies scene constraints; Based on the synergistic control effect of PIS and PA, a multi-indicator comprehensive reward function is constructed. The optimal collaborative control strategy is obtained by using a multi-indicator comprehensive reward function and membership degree to select the strategy.

2. The PIS and PA fusion collaborative control method based on multi-source sensing according to claim 1, characterized in that, The process of performing spatiotemporal alignment and heterogeneous feature encoding on the multi-source sensing data to construct a scene feature vector includes: Train operation status, passenger flow density, environmental status and manual triggering commands are synchronized according to a unified time base to obtain a time-aligned multi-source data sequence; The multi-source data sequences are mapped to regions according to the station's spatial structure to obtain multi-source data subsets organized by spatial regions. Perform dimensional unification processing on subsets of multi-source data; Based on the multi-source data subset processed by dimension unification, feature structure encoding is performed to obtain multi-dimensional feature sub-vectors corresponding to each spatial region; By combining all the multidimensional feature vectors, the scene feature vector is constructed.

3. The PIS and PA fusion collaborative control method based on multi-source sensing according to claim 1, characterized in that, The step of determining the target scene by performing multi-scene identification and conflict resolution based on the scene feature vector includes: Based on the operational needs of rail transit, multiple candidate scenarios are defined; The scene matching degree is obtained by matching the scene feature vector with the structured reference features of each candidate scene; Candidate scenes with a scene matching degree greater than a preset threshold are selected as the candidate set; The candidate set is reconciled based on scene priority and scene matching degree to determine the target scene.

4. The PIS and PA fusion collaborative control method based on multi-source sensing according to claim 1, characterized in that, The process of selecting the optimal collaborative control strategy through a multi-index comprehensive reward function and membership degree includes: The decay weights of historical execution data are generated using a time decay factor. The historical reward value of each control strategy is obtained by weighting the historical execution data using a decay weight and a multi-index comprehensive reward function. Construct the posterior reward distribution for each control policy using historical reward values; For each control policy, a random sample is taken based on the posterior reward distribution to obtain the sampled reward value of the control policy; The evaluation value of the control strategy is constructed based on the sampled reward value and the membership degree; The optimal collaborative control strategy is obtained from the control strategy space based on the evaluation value.

5. The PIS and PA fusion collaborative control method based on multi-source sensing according to claim 1, characterized in that, The step of performing coordinated control of the PIS system and PA system according to the optimal cooperative control strategy, and making adaptive adjustments based on execution feedback, includes: The PIS system and PA system are controlled in conjunction according to the optimal cooperative control strategy described above. The scene feature vector is updated by executing feedback data, and the target scene is redefined based on the updated scene feature vector; If the redefined target scenario remains unchanged, the control parameters of the optimal cooperative control strategy are dynamically adjusted. If the redefined target scenario changes, the control strategy space is reconstructed, and the next round of optimal collaborative control strategy generation is carried out.

6. A PIS and PA fusion and collaborative control system based on multi-source sensing, characterized in that, include: The acquisition module is used to acquire multi-source sensing data in the rail transit operation environment, including train operation status, passenger flow density, environmental status, and manual triggering commands. The processing module is used to perform spatiotemporal alignment and heterogeneous feature encoding on the multi-source sensing data to construct scene feature vectors; The identification module is used to perform multi-scene identification and conflict resolution based on the scene feature vector to determine the target scene; The optimization module is used to construct a control strategy space containing PIS and PA linkage parameters for the target scenario, and determine the optimal cooperative control strategy through a strategy optimization mechanism that combines fuzzy control and improved Thompson sampling. The control module is used to perform coordinated control of the PIS system and PA system according to the optimal cooperative control strategy, and to make adaptive adjustments based on execution feedback. The optimization module includes: Define the unit, used to define the PIS control parameter set and the PA control parameter set; The first building unit is used to combine the PIS control parameter set and the PA control parameter set, and to determine the constraints through the set of constraints of the target scenario, thereby constructing a control strategy space that satisfies the scenario constraints. The inference unit is used to perform fuzzy inference based on scene feature vectors to obtain the membership degree of each control strategy in the control strategy space that satisfies scene constraints. The second building unit is used to construct a multi-indicator comprehensive reward function based on the synergistic control effect of PIS and PA; The selection unit is used to select the optimal collaborative control strategy by combining a multi-index reward function and membership degree.

7. The PIS and PA fusion and collaborative control system based on multi-source sensing according to claim 6, characterized in that, The processing module includes: The first processing unit is used to synchronously process the train operation status, passenger flow density, environmental status and manual triggering commands according to a unified time base to obtain a time-aligned multi-source data sequence. The mapping unit is used to map multi-source data sequences according to the station's spatial structure to obtain multi-source data subsets organized by spatial regions. The second processing unit is used to perform dimensional unification processing on multi-source data subsets. The encoding unit is used to perform feature structure encoding on a subset of multi-source data based on dimensional uniform processing, to obtain a multi-dimensional feature sub-vector corresponding to each spatial region; The combination unit is used to combine all multidimensional feature vectors to construct the scene feature vector.

8. The PIS and PA fusion and collaborative control system based on multi-source sensing according to claim 6, characterized in that, The identification module includes: The first definition unit is used to define multiple candidate scenarios based on the operational needs of rail transit; The matching unit is used to match the scene feature vector with the structured reference features of each candidate scene to obtain the scene matching degree; The filtering unit is used to select candidate scenes with a scene matching degree greater than a preset threshold as a candidate set. The conflict resolution unit is used to resolve conflicts in the candidate set based on scene priority and scene matching degree to determine the target scene.

Citation Information

Patent Citations

  • Dynamic beam forming control system and method based on multi-target reinforcement learning

    CN121308790A