Multi-source coupled personnel positioning system and method for cruise ship

Abstract

The application discloses a kind of cruise ship multi-source coupling personnel positioning system and method, including the positioning system of bracelet terminal, shipborne infrastructure and server, wherein bracelet terminal is worn on the wrist of passenger or crew on cruise ship, for collecting wearer relevant information and carrying out wireless communication with shipborne infrastructure;Shipborne infrastructure includes several LoRa base stations, WiFi access points, WiFi anchor points, UHF RFID readers and ship monitoring equipment distributed in different positions of ship body;Server is used to receive and process various data, vote screening, weighted fusion, Kalman filtering processing are carried out to the multi-source positioning data received, and the final positioning result of personnel is output.Effective inhibition of the error and failure of single data source, significantly improve the robustness and reliability of positioning system, reduce the mispositioning risk caused by personnel and terminal mismatch, with good expansibility and application value.

Description

Technical Field

[0001] This invention belongs to the field of cruise ship personnel positioning, specifically relating to a cruise ship multi-source coupled personnel positioning system and method. Background Technology

[0002] With the rapid development of the cruise tourism industry, the number of passengers and the size of cruise ships are constantly increasing, and the internal space of ships is becoming increasingly complex, typically including multiple decks, large public halls, catering and entertainment areas, and a large number of cabins. In the current cruise ship safety management process, the location information of passengers and crew mainly relies on manual roll call, access control card records, and surveillance video playback, which makes it difficult to grasp the spatial distribution of people on board in a timely and continuous manner.

[0003] To improve personnel management, some ships have attempted to introduce indoor positioning systems based on single technologies, such as WiFi-based RSSI positioning and RFID-based access control positioning. However, due to the significant signal obstruction and reflection effects of metal compartments, enclosed cabins, and multiple decks in cruise ship structures, single-technology positioning systems are susceptible to multipath interference, resulting in poor positioning accuracy and stability, and ineffective coverage in certain areas. Furthermore, in emergencies such as fires or flooding that cause local power or communication system failures, positioning systems relying on a single infrastructure often become completely paralyzed, failing to provide reliable data support for subsequent emergency evacuation and search and rescue operations.

[0004] On the other hand, cruise ships have a large number of key passenger groups, such as children and the elderly. Existing systems can only assist in location tracking through room card swipe records or simple video searches, which cannot provide continuous monitoring of key groups or prevent children from getting lost. Furthermore, existing technologies lack an effective mechanism to determine whether wearable devices are actually being worn by the person, which can easily lead to mislocation problems caused by a mismatch between the person and the device.

[0005] Therefore, there is an urgent need for a personnel positioning system that can integrate multiple positioning technologies in the complex cruise ship environment, has fault tolerance and redundancy capabilities, and can still provide basic positioning services in emergency situations, in order to improve the level of safety management during cruise ship operations. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a multi-source coupled personnel positioning system and method for cruise ships, which solves the problem of mispositioning caused by mismatch between personnel and terminals on cruise ships in the prior art.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] A cruise ship multi-source coupled personnel positioning system includes a positioning system comprising a wristband terminal, shipboard infrastructure, and a server. The wristband terminal is equipped with a LoRa communication module, a WiFi communication module, a UHF RFID tag, and a heart rate sensor. It is worn on the wrist of passengers or crew members on the cruise ship to collect relevant information about the wearer and to communicate wirelessly with the shipboard infrastructure.

[0009] The shipboard infrastructure includes several LoRa base stations, WiFi access points, WiFi anchor points, UHF RFID readers and ship monitoring equipment distributed in different locations on the hull, which collect video images of public areas and upload them to the server for computer vision processing, and card swiping devices installed at cabin doors, boarding gates and entrances to certain service facilities.

[0010] The server is located inside the ship's safety control center or computer room to receive and process various types of data. It performs voting filtering, weighted fusion, and Kalman filtering on the received multi-source positioning data to output the final positioning result of the personnel.

[0011] The wristband terminal periodically sends wireless signals containing the terminal's identification to the LoRa base station and WiFi access point through its built-in LoRa communication module, WiFi communication module, and UHF RFID tag, and passively responds to read the data when passing through the coverage area of ​​UHF RFID readers placed in passageways or stairwells; it also collects the wearer's heart rate data in real time through its built-in heart rate sensor and uploads it to the server.

[0012] The server obtains RSSI positioning data based on LoRa 4MHz from the LoRa base station, fingerprint positioning, TDOA positioning, or RSSI ranging positioning data based on WiFi from the WiFi access point, presence detection positioning data from the UHF RFID reader, and personnel positioning data based on human detection and facial recognition from the ship monitoring equipment. It also generates event record positioning data based on the card swiping events uploaded by the card swiping device, forming a multi-source positioning data set.

[0013] The process by which the server performs a voting filter on multi-source location data is as follows:

[0014] First, the server selects three positioning points from the multi-source positioning data at the current moment, calculates the geometric mean position of the three positioning points, and uses the geometric mean position as the candidate position.

[0015] Then, the distance between the remaining positioning points and the candidate position is calculated, and positioning points whose distance is less than a preset distance threshold are regarded as voting support for the candidate position; when the number of positioning points voting support accounts for more than half of the total number of positioning points participating in the vote, the server determines the candidate position as the winning position for this positioning, and marks the positioning data corresponding to the data source that is consistent with or similar to the winning position as valid positioning data.

[0016] Finally, for location data sources that significantly deviate from the candidate positions in multiple voting rounds, the server marks them as abnormal data sources and reduces their weight or temporarily removes their location data.

[0017] The server performs weighted fusion of valid location data as follows:

[0018] The server assigns basic weights to various positioning data sources based on their hardware accuracy and environmental adaptability. The weights are arranged from highest to lowest as follows: human positioning data based on computer vision, positioning data based on WiFi fingerprint or TDOA, anchored cooperative positioning data based on WiFi RSSI, and positioning data based on LoRa RSSI and UHF RFID presence detection have relatively lower weights.

[0019] For location data obtained based on UHF RFID presence detection and card swipe events, the server also dynamically adjusts the weight of the event over time using an exponential decay function based on the difference between the event time and the current time, so that the influence of earlier event location data on the result gradually weakens during the fusion process.

[0020] The process by which the server obtains the final location result is as follows:

[0021] First, the server inputs the weighted fusion measurement positions as observations into the Kalman filter model to smooth the personnel position trajectories;

[0022] Then, the server constructs a state vector containing position and velocity components, predicts the position of the person at the next moment through a preset state transition equation, and compares the predicted position with the fused measurement position using the observation equation.

[0023] Finally, the state estimate is updated based on the prediction error, thus obtaining a continuous and smooth final positioning result.

[0024] When the system detects that the number of available location data sources for voting is less than a preset threshold, the server determines that the current system is in an emergency or abnormal state and automatically triggers the emergency mode. In the emergency mode, the server estimates the location of personnel only based on LoRa4MHz RSSI ranging and positioning data to ensure that a rough positioning result for the entire ship can still be provided even if WiFi access points, UHF RFID readers, or ship monitoring equipment fail.

[0025] The server determines whether the wristband is worn on a living person based on heart rate data. When a heart rate is continuously missing or in an abnormal range, the location data corresponding to the wristband is considered unreliable and its weight is reduced or removed during the voting and fusion process. In addition, when the computer vision module identifies a passenger with high confidence and the identification information is inconsistent with that uploaded by the wristband, the server marks the wristband's location data as abnormal data and gives the computer vision location result a higher priority, directly using the computer vision location result as the final location result.

[0026] The multi-source coupled personnel localization method for cruise ships includes the following steps:

[0027] Step 1: When crew members and passengers board the cruise ship, they wear wristband terminals, receive room cards, and enter their basic personal information. The system automatically binds the ID number of the wristband terminal and the passenger information.

[0028] Step 2: The wristband terminal periodically sends wireless signals containing the terminal's identification to the LoRa base station and WiFi access point through its built-in LoRa communication module, WiFi communication module and UHF RFID tag, and passively responds to read the data when passing through the coverage area of ​​the UHF RFID reader placed in the passage or stairwell; the ship monitoring equipment collects video images of public areas and uploads them to the server for computer vision processing.

[0029] Step 3: The server obtains multi-source positioning data sets from various parts, performs voting filtering, weighted fusion, and Kalman filtering, and outputs the final positioning results of the personnel.

[0030] When the emergency mode management module in the server detects that the types of location data sources that can participate in voting have decreased significantly, it determines that the current system may be in an emergency or abnormal state. Once the emergency or abnormal state is confirmed, it automatically switches to a simplified location strategy mode.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. By coupling multiple positioning technologies such as LoRa, WiFi, UHF RFID, computer vision, anchored cooperative positioning, and event logging positioning, the system effectively suppresses errors and failures from a single data source through a voting mechanism and weighted fusion, significantly improving the robustness and reliability of the positioning system.

[0033] 2. By introducing heart rate detection and computer vision identity recognition mechanisms, it is possible to identify situations where the wristband is not being worn or is not in the wearer's hand, reducing the risk of mislocation caused by mismatch between the person and the terminal.

[0034] 3. In normal mode, it can provide high-precision and continuous personnel location trajectories. In emergency mode, even if only LoRa base stations are retained, it can still achieve rough positioning of the entire ship, providing necessary data support for emergency evacuation and search and rescue.

[0035] 4. The multi-source coupling positioning results can be used as the basic data input for the upper-level safety management platform and emergency evacuation guidance system, and have good scalability and application value. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the overall structure of the cruise ship multi-source coupled personnel positioning system of the present invention.

[0037] Figure 2 This is a schematic diagram of the hardware structure of the wristband terminal of the present invention.

[0038] Figure 3 This is a schematic diagram illustrating the arrangement of a LoRa base station, a WiFi access point, a UHF RFID reader, and a video surveillance camera on a cruise ship deck, as provided in this invention example.

[0039] Figure 4 This is a schematic diagram of the multi-source positioning data voting screening and weighted fusion processing flow of the present invention.

[0040] Figure 5 This is a schematic diagram of the anchored cooperative positioning based on WiFi RSSI according to the present invention. Detailed Implementation

[0041] The structure and working process of the present invention will be further described below with reference to the accompanying drawings.

[0042] The purpose of this invention is to provide a multi-source coupled personnel positioning method, device, and system for cruise ships. By coupling LoRa 4MHz positioning, WiFi fingerprint and TDOA / RSSI positioning, UHF RFID presence detection positioning, computer vision positioning, WiFi RSSI-based anchored cooperative positioning, and event logging positioning, and combining voting mechanisms, weighted fusion, and Kalman filtering, the system can reliably obtain the location of personnel on cruise ships. At the same time, in the event of failure of some positioning subsystems or personnel not wearing the terminal correctly, the overall availability of the positioning system is ensured through anomaly identification and emergency mode switching.

[0043] A cruise ship multi-source coupled personnel positioning system includes a positioning system comprising a wristband terminal, shipboard infrastructure, and a server. The wristband terminal is equipped with a LoRa communication module, a WiFi communication module, a UHF RFID tag, and a heart rate sensor. It is worn on the wrist of passengers or crew members on the cruise ship to collect relevant information about the wearer and to communicate wirelessly with the shipboard infrastructure.

[0044] The shipboard infrastructure includes several LoRa base stations, WiFi access points, WiFi anchor points, UHF RFID readers and ship monitoring equipment distributed in different locations on the hull, which collect video images of public areas and upload them to the server for computer vision processing, and card swiping devices installed at cabin doors, boarding gates and entrances to certain service facilities.

[0045] The server is located inside the ship's safety control center or computer room to receive and process various types of data. It performs voting filtering, weighted fusion, and Kalman filtering on the received multi-source positioning data to output the final positioning result of the personnel.

[0046] Specific embodiments, such as Figures 1 to 5 As shown:

[0047] This embodiment provides a multi-source coupled personnel positioning system for cruise ships. The system can generally include a wristband terminal, shipboard infrastructure, and a server. The wristband terminal is worn on the wrist of passengers or crew members to collect relevant information about the wearer and communicate wirelessly with the shipboard infrastructure. The wristband terminal integrates a LoRa communication module, a WiFi communication module, a UHF RFID tag, a heart rate sensor, a display and prompt unit, and control circuitry. The LoRa base station, WiFi access point, UHF RFID reader, and video surveillance camera are used to receive signals from the wristband terminal or collect personnel images. The server processes the received multi-source positioning data through voting filtering, weighted fusion, Kalman filtering, etc., and outputs the final personnel positioning result.

[0048] Shipboard infrastructure may include several LoRa base stations, several WiFi access points, WiFi anchor points, several UHF RFID readers, several video surveillance cameras, and card readers distributed at different locations on the hull. The locations of all types of equipment can be pre-recorded in the server during system deployment.

[0049] The server can be located inside the ship's safety control center or computer room, and can generally include data acquisition modules, voting and anomaly detection modules, fusion and filtering modules, emergency mode management modules, and interface and display modules. Those skilled in the art can divide the above functional modules into multiple program units or have them implemented by a single program, depending on the software implementation method; this does not affect the essence of the present invention.

[0050] Through the above structure, the present invention forms an overall architecture of "terminal acquisition + multiple infrastructure perception + central server fusion processing", laying the foundation for subsequent multi-source coupling positioning.

[0051] To achieve the functions of this system, the specific processing steps for each part are as follows:

[0052] First, the wristband terminal periodically sends wireless signals containing the terminal's identification to the LoRa base station and WiFi access point through its built-in LoRa communication module, WiFi communication module, and UHF RFID tag. It passively responds to read the signals when passing through the coverage area of ​​UHF RFID readers placed in passageways or stairwells. Simultaneously, the ship's monitoring equipment (including video capture cameras) captures video images of public areas and uploads them to the server for computer vision processing. The server obtains RSSI positioning data based on LoRa 4MHz from the LoRa base station, fingerprint positioning, TDOA positioning, or RSSI ranging positioning data based on WiFi from the WiFi access point, presence detection positioning data from the UHF RFID reader, and personnel positioning data based on human detection and facial recognition from the computer vision module. It also generates event record positioning data based on card-swiping events uploaded by the card-swiping device, thus forming a multi-source positioning data set.

[0053] Secondly, the server performs voting filtering on the multi-source positioning data. Specifically, the server selects three positioning points from the multi-source positioning data at the current moment, calculates the geometric mean position of the three positioning points, and uses this geometric mean position as a candidate position; then, it calculates the distance between the remaining positioning points and the candidate position, and considers positioning points whose distance is less than a preset distance threshold as voting support for the candidate position; when the number of positioning points supporting the vote accounts for more than half of the total number of positioning points participating in the vote, the server determines the candidate position as the winning position for this positioning, and marks the positioning data corresponding to the data source that is consistent with or similar to the winning position as valid positioning data; for positioning data sources that significantly deviate from the candidate position in multiple voting rounds, the server marks them as abnormal data sources and reduces their weight or temporarily removes their positioning data.

[0054] Next, the server performs weighted fusion of the valid location data. The server assigns basic weights to various location data sources based on their hardware accuracy and environmental adaptability. Human location data based on computer vision has a higher weight, followed by location data based on WiFi fingerprinting or TDOA, then WiFi RSSI anchored cooperative positioning data has a moderate weight, and location data based on LoRa RSSI and UHF RFID presence detection has a relatively lower weight. For location data obtained based on UHF RFID presence detection and card swipe events, the server also dynamically adjusts its weight over time using an exponential decay function based on the difference between the event's occurrence time and the current time, so that the influence of earlier event location data on the result gradually weakens during the fusion process.

[0055] Next, the server inputs the weighted fusion measurement positions as observations into the Kalman filter model to smooth the personnel's position trajectory. The server constructs a state vector containing position and velocity components, predicts the personnel's position at the next moment using a preset state transition equation, and compares the predicted position with the fused measurement position using the observation equation. The state estimate is updated based on the prediction error, thus obtaining a continuous and smooth final positioning result.

[0056] During the above process, the wristband terminal collects the wearer's heart rate data in real time through its built-in heart rate sensor and uploads it to the server. The server determines whether the wristband terminal is worn on a living person based on the heart rate data. When a persistently missing or abnormal heart rate is detected, the corresponding location data of the wristband terminal is considered unreliable data and its weight is reduced or removed during the voting and fusion process. Furthermore, when the computer vision module identifies a passenger with high confidence, but this identification does not match the identity information uploaded by the wristband terminal, the server marks the wristband terminal's location data as abnormal data and assigns a higher priority to the computer vision location result. If necessary, the computer vision location result is directly used as the final location result.

[0057] Additionally, when the system detects that the number of available location data sources (excluding event log locations) is less than a preset threshold, the server determines that the system is in an emergency or abnormal state and automatically triggers emergency mode. In emergency mode, the server estimates personnel location solely based on LoRa 4MHz RSSI ranging data to ensure that a rough location result for the entire ship can still be provided even if WiFi access points, UHF RFID readers, or video surveillance equipment fail.

[0058] The overall working principle and working process of the system and its individual components are as follows:

[0059] Working principles of the wristband terminal and its subsystems:

[0060] See Figure 2 In a typical implementation, the wristband terminal periodically transmits wireless signals containing its own identification information to multiple LoRa base stations via its LoRa wireless module. The LoRa base stations receive the signals and measure the received signal strength, thereby obtaining a set of distance-related RSSI data. In practical deployments, a logarithmic distance path loss model can be used.

[0061]

[0062] in:

[0063] Distance between the terminal and the base station;

[0064] : Signal strength at a reference distance (1m);

[0065] Path loss index (commonly taken as 2-3.5 for shipboard indoor environments);

[0066] : Actual measured signal strength.

[0067] Because LoRa signals have strong coverage capabilities, even when the ship's structure is complex or power supply to some areas is affected, basic coverage can still be obtained over a large area. This feature plays an important role in subsequent emergency modes.

[0068] Meanwhile, the wristband terminal communicates with the ship's WiFi access point via a WiFi communication module. During daily operation, the WiFi access point can report the signal strength information of the wristband terminal. The server can use this information to match with a pre-established fingerprint database, or combine time information from multiple access points to calculate TDOA (Time Difference of Arrival), thereby obtaining relatively high-precision WiFi positioning information. The TDOA positioning equation is as follows:

[0069] For three reference base stations The time difference is measured as follows:

[0070]

[0071] Forming the hyperbola positioning equation system:

[0072]

[0073] Solving for the terminal position .

[0074] This invention does not limit the specific implementation of WiFi positioning algorithms, as long as they can provide location-related measurement results.

[0075] When a person wearing a wristband passes through a passageway, stairwell, or doorway equipped with a UHF RFID reader, the reader can read the UHF RFID tag's identification information and report the event "a tag has appeared in the reader's area" to the server. The server can then update the location of the person corresponding to that tag to the coordinates or area code of the reader's area. Although this method can only provide a rough location at the area level, it has high reliability in scenarios such as determining whether a person has passed through a key passageway or entered a secure area.

[0076] Video surveillance cameras can transmit monitoring video to servers or specialized video analysis equipment. Video analysis can utilize existing human detection and tracking (YOLO) and facial recognition algorithms. By applying the principle of similar triangles, the image location corresponding to the person's identity can be determined. Combined with the camera's installation location, orientation, and field of view, the position in the image can be converted into the ship's planar coordinate system, thereby obtaining highly accurate visual positioning data.

[0077] In addition, several fixed wireless anchor points can be set up in open areas such as the ship's main hall, such as... Figure 5As shown, the positions of each anchor point are pre-recorded on the server during deployment. The wristband terminal detects the signal strength of surrounding anchor points and all surrounding wristbands via its WiFi communication module. The server calculates the relative position of the wristband terminal within the open area based on the approximate changes in these signal strengths. Although a single RSSI measurement can be affected by the environment, by combining the constraints of multiple anchor points and correcting with data from other wristbands, a relatively reasonable approximate position within the area can be obtained. This has a positive effect on filling the positioning gaps in visual blind spots and areas where RFID deployment is difficult.

[0078] Card readers installed at cabin doors, boarding gates, and entrances to certain service facilities can report card-swiping events when passengers swipe their room cards or make payments. The server can then update the location of the corresponding person to the location of that card reader based on its fixed location. Although this location is "event-based" and instantaneous, it can provide an important constraint for determining "where a person has been" when inferring travel patterns.

[0079] Through the above methods, the server can aggregate candidate location information from multiple subsystems such as LoRa, WiFi, UHF RFID, vision, anchored collaboration, and event logging within each time slice, thereby overcoming the problem in the background technology that "single location technology is severely interfered with and the whole system is unusable when a certain system fails".

[0080] Voting and Anomaly Detection of Multi-Source Data

[0081] See Figure 4 In this embodiment of the invention, in order to avoid individual data from deviating significantly from the true location and affecting the overall positioning accuracy, the voting and anomaly identification module in the server can perform a kind of "group consensus" screening on the candidate locations given by each data source.

[0082] Specifically, at any given moment, the server can obtain several candidate location points from different data sources. The voting and anomaly detection module can select several candidate location points, for example, three randomly selected, to obtain a representative average point. Then, it compares the distances between other candidate location points and this average point. If the distance is within a certain preset range, these location points are considered to have "voted in favor" of the average point. If the number of candidate location points voting in favor is significantly greater than the number voting against, then the average point can be regarded as the "consensus point" for this round of voting. The mathematical expression of the voting mechanism is as follows:

[0083] Candidate position set:

[0084]

[0085] Find the geometric mean of any three points:

[0086]

[0087] Calculate the distances to the remaining points:

[0088]

[0089] If the following conditions are met:

[0090]

[0091] Then the The winning position in this round of voting.

[0092] Through multiple rounds or continuous voting processes, if a certain type of data source is observed to almost always fall outside the voting consensus, the voting and anomaly identification module can mark the data source as an anomaly and automatically reduce its weight or even temporarily disregard its data in subsequent fusion stages. This allows for the elimination of erroneous localizations caused by environmental anomalies, equipment malfunctions, or certain special interferences without manual intervention. This process effectively avoids the common problem in the background technology where "a single faulty device drags down the overall localization results."

[0093] On the other hand, the heart rate sensor in the wristband can continuously report the wearer's heart rate data. The server can determine whether the wristband is being worn correctly on a living person based on the presence and trend of the heart rate signal. When the heart rate signal is missing or exhibits abnormal static characteristics for a period of time, there is reason to believe that the wristband may have been removed, placed on a table, or in another location. In this case, the voting and anomaly recognition module can reduce the reliability of the location data corresponding to the terminal, reducing false positioning caused by mismatch between the wristband and the person.

[0094] Furthermore, when the vision subsystem identifies a specific person and that identification does not match the identification information of a nearby wristband terminal, the server can determine that there is a "person-wristband mismatch". In this case, the vision positioning can be given higher weight, or the positioning of the wristband terminal can be "rejected" to ensure that the final positioning result can reflect the real location of the person as accurately as possible, rather than just following the location of the terminal.

[0095] Fusion and Filtering and Their Technical Effects

[0096] After completing the voting and anomaly identification, the fusion and filtering module in the server can weight and combine the voted data. Different data sources can be assigned different basic weights based on their positioning accuracy, coverage, and sensitivity to the environment. For example, visual positioning data generally has high accuracy and can be given a larger weight; LoRa RSSI positioning data generally has coarser accuracy but good coverage and can be given a smaller weight; RFID presence detection and card-swipe event-based positioning can be given a certain weight for a period of time after the event occurs, and its influence can be gradually reduced over time.

[0097] After weighted fusion, a comprehensive measurement position can be obtained. The fusion and filtering module can then use filtering methods to smooth the personnel's trajectory. The filtering process can use Kalman filtering or other similar state estimation algorithms. The basic idea is: given the position and velocity at the previous moment, predict the position at the next moment, and then correct the prediction using the current fused measurement value, thereby obtaining a continuous, smooth trajectory that is less susceptible to noise from single measurements. The reference equation is as follows:

[0098] Equations of state:

[0099]

[0100] Observation equation:

[0101]

[0102] predict:

[0103]

[0104] Kalman gain:

[0105]

[0106] renew:

[0107]

[0108] In this way, the present invention inherits the advantages of high-precision data while utilizing the redundancy of multi-source data. Even in complex wireless environments, with multipath effects and local obstruction, it can still maintain stable and reliable overall positioning results. This is of great significance for subsequent heat map analysis, hazardous area early warning, and emergency evacuation route generation.

[0109] Emergency Mode and its Advantages

[0110] When the emergency mode management module in the server detects a significant reduction in the types of location data sources that can participate in the voting process, such as a large-scale offline status of WiFi access points or video surveillance cameras, it can determine that the current system may be in an emergency or abnormal state. At this time, the emergency mode management module can automatically switch the system to a simplified location strategy, that is, to rely mainly on the RSSI information received by the LoRa base station for coarse location.

[0111] Although this positioning method has lower accuracy, it still possesses high survivability in situations where power is limited or communication is compromised due to the coverage and low power consumption characteristics of LoRa signals. By retaining this "minimum" positioning capability, ship managers can still see the approximate distribution of personnel throughout the ship in emergencies, thereby enabling them to issue targeted evacuation orders or organize search and rescue operations. This is a significant improvement over the "complete failure of the positioning system in emergency situations" mentioned in the background technology.

[0112] In summary, this invention enables the positioning system to maintain high reliability and robustness in complex environments through the collaborative cooperation between the wristband terminal and various infrastructures, as well as mechanisms such as voting, fusion, and emergency mode management in the server, and provides high-value basic data support for upper-layer applications such as emergency evacuation systems.

[0113] Based on the above system, the present invention also discloses a multi-source coupled personnel positioning method for cruise ships, comprising the following steps:

[0114] Step 1: When crew members and passengers board the cruise ship, they wear wristband terminals, receive room cards, and enter their basic personal information. The system automatically binds the ID number of the wristband terminal and the passenger information.

[0115] Step 2: The wristband terminal periodically sends wireless signals containing the terminal's identification to the LoRa base station and WiFi access point through its built-in LoRa communication module, WiFi communication module and UHF RFID tag, and passively responds to read the data when passing through the coverage area of ​​the UHF RFID reader placed in the passage or stairwell; the ship monitoring equipment collects video images of public areas and uploads them to the server for computer vision processing.

[0116] Step 3: The server obtains multi-source positioning data sets from various parts, performs voting filtering, weighted fusion, and Kalman filtering, and outputs the final positioning results of the personnel.

[0117] When the emergency mode management module in the server detects that the types of location data sources that can participate in voting have decreased significantly, it determines that the current system may be in an emergency or abnormal state. Once the emergency or abnormal state is confirmed, it automatically switches to a simplified location strategy mode.

[0118] It should be understood that this solution is not limited to the specific embodiments described above. Devices and structures not described in detail herein should be understood as being implemented in a manner common to the art. Any person skilled in the art can make many possible variations and modifications to this solution, or modify it into equivalent embodiments, without departing from the scope of this solution, using the methods and techniques disclosed above. This does not affect the substantive content of this solution. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this solution, without departing from its scope, still fall within the protection scope of this solution.

Claims

1. A multi-source coupled personnel positioning system for cruise ships, characterized in that: The positioning system includes a wristband terminal, shipboard infrastructure, and a server. The wristband terminal has a built-in LoRa communication module, WiFi communication module, UHF RFID tag, and heart rate sensor. It is worn on the wrists of passengers or crew members on the cruise ship to collect relevant information about the wearer and communicate wirelessly with the shipboard infrastructure. The shipboard infrastructure includes several LoRa base stations, WiFi access points, WiFi anchor points, UHF RFID readers and ship monitoring equipment distributed in different locations on the ship. It collects video images of public areas and uploads them to the server for computer vision processing. It also includes card swiping devices installed at cabin doors, boarding gates and entrances to certain service facilities. The server is located inside the ship's safety control center or computer room to receive and process various types of data. It performs voting filtering, weighted fusion, and Kalman filtering on the received multi-source positioning data to output the final positioning result of the personnel.

2. The cruise ship multi-source coupled personnel positioning system according to claim 1, characterized in that: The wristband terminal periodically sends wireless signals containing the terminal's identification to the LoRa base station and WiFi access point through its built-in LoRa communication module, WiFi communication module, and UHF RFID tag, and passively responds to read the data when passing through the coverage area of ​​UHF RFID readers placed in passageways or stairwells; it also collects the wearer's heart rate data in real time through its built-in heart rate sensor and uploads it to the server.

3. The cruise ship multi-source coupled personnel positioning system according to claim 1, characterized in that: The server obtains RSSI positioning data based on LoRa 4MHz from the LoRa base station, fingerprint positioning, TDOA positioning, or RSSI ranging positioning data based on WiFi from the WiFi access point, presence detection positioning data from the UHF RFID reader, and personnel positioning data based on human detection and facial recognition from the ship monitoring equipment. It also generates event record positioning data based on the card swiping events uploaded by the card swiping device, forming a multi-source positioning data set.

4. The cruise ship multi-source coupled personnel positioning system according to claim 3, characterized in that: The process by which the server performs a voting filter on multi-source location data is as follows: First, the server selects three positioning points from the multi-source positioning data at the current moment, calculates the geometric mean position of the three positioning points, and uses the geometric mean position as the candidate position. Then, the distance between the remaining positioning points and the candidate position is calculated, and positioning points whose distance is less than a preset distance threshold are regarded as voting support for the candidate position; when the number of positioning points voting support accounts for more than half of the total number of positioning points participating in the vote, the server determines the candidate position as the winning position for this positioning, and marks the positioning data corresponding to the data source that is consistent with or similar to the winning position as valid positioning data. Finally, for location data sources that significantly deviate from the candidate positions in multiple voting rounds, the server marks them as abnormal data sources and reduces their weight or temporarily removes their location data.

5. The cruise ship multi-source coupled personnel positioning system according to claim 4, characterized in that: The server performs weighted fusion of valid location data as follows: The server assigns basic weights to various positioning data sources based on their hardware accuracy and environmental adaptability. The weights are arranged from highest to lowest as follows: human positioning data based on computer vision, positioning data based on WiFi fingerprint or TDOA, anchored cooperative positioning data based on WiFi RSSI, and positioning data based on LoRa RSSI and UHF RFID presence detection have relatively lower weights. For location data obtained based on UHF RFID presence detection and card swipe events, the server also dynamically adjusts the weight of the event over time using an exponential decay function based on the difference between the event time and the current time, so that the influence of earlier event location data on the result gradually weakens during the fusion process.

6. The cruise ship multi-source coupled personnel positioning system according to claim 5, characterized in that: The process by which the server obtains the final location result is as follows: First, the server inputs the weighted fusion measurement positions as observations into the Kalman filter model to smooth the personnel position trajectories; Then, the server constructs a state vector containing position and velocity components, predicts the position of the person at the next moment through a preset state transition equation, and compares the predicted position with the fused measurement position using the observation equation. Finally, the state estimate is updated based on the prediction error, thus obtaining a continuous and smooth final positioning result.

7. The cruise ship multi-source coupled personnel positioning system according to claim 4, characterized in that: When the system detects that the number of available location data sources for voting is less than a preset threshold, the server determines that the current system is in an emergency or abnormal state and automatically triggers emergency mode. In emergency mode, the server estimates personnel location based solely on LoRa 4MHz RSSI ranging data to ensure a rough location result across the entire ship is still available even if WiFi access points, UHF RFID readers, or ship monitoring equipment fail.

8. The cruise ship multi-source coupled personnel positioning system according to claim 1, characterized in that: The server determines whether the wristband is being worn on a living human body based on heart rate data. When a heart rate is continuously missing or in an abnormal range, the location data corresponding to the wristband is considered untrusted data and its weight is reduced or removed during the voting and fusion process. Furthermore, when the computer vision module identifies a passenger with high confidence and the identification information is inconsistent with that uploaded by the wristband terminal, the server marks the wristband terminal's location data as abnormal data and assigns a higher priority to the computer vision location result, directly using the computer vision location result as the final location result.

9. A cruise ship multi-source coupled personnel positioning method based on the system described in any one of claims 1 to 8, characterized in that: Includes the following steps: Step 1: When crew members and passengers board the cruise ship, they wear wristband terminals, receive room cards, and enter their basic personal information. The system automatically binds the ID number of the wristband terminal and the passenger information. Step 2: The wristband terminal periodically sends wireless signals containing the terminal's identification to the LoRa base station and WiFi access point through its built-in LoRa communication module, WiFi communication module and UHF RFID tag, and passively responds to read the data when passing through the coverage area of ​​the UHF RFID reader placed in the passage or stairwell; the ship monitoring equipment collects video images of public areas and uploads them to the server for computer vision processing. Step 3: The server obtains multi-source positioning data sets from various parts, performs voting filtering, weighted fusion, and Kalman filtering, and outputs the final positioning results of the personnel.

10. The cruise ship multi-source coupled personnel positioning method according to claim 9, characterized in that: When the emergency mode management module in the server detects that the types of location data sources that can participate in voting have decreased significantly, it determines that the current system may be in an emergency or abnormal state. Once the emergency or abnormal state is confirmed, it automatically switches to a simplified location strategy mode.

Images