Ocean container intelligent energy-saving tracking system based on multi-mode perception
By combining a multimodal sensing system with satellite positioning and acceleration sensors, intelligent energy-saving tracking of ocean-going containers is achieved, solving the problems of power waste and inaccurate status identification in ocean transportation, and realizing high-efficiency energy saving and accurate status switching of equipment.
Patent Information
- Application Number
- CN202511732477.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-24
- Publication Date
- 2026-02-24
AI Technical Summary
Existing ocean container trackers cannot accurately and flexibly switch to power-saving modes during ocean transportation, resulting in wasted power consumption or increased hardware and labor costs. Furthermore, unstable signals prevent the devices from accurately entering power-saving mode.
By employing a multimodal sensing system that combines satellite positioning, acceleration sensors, and network status monitoring, the system determines the maritime, berthing, and land transport status of containers through multi-source data fusion, thereby implementing a graded energy-saving strategy, including switching between different operating frequencies in offshore, berthing, and land transport scenarios.
Without increasing hardware and labor costs, it can accurately identify the status of containers, optimize power consumption in different scenarios, extend equipment life, and ensure energy saving and accurate location updates during long voyages.
Smart Images
Figure CN121567733A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of container transport monitoring technology, and in particular to an intelligent energy-saving tracking system for ocean-going containers based on multimodal perception. Background Technology
[0002] While existing container trackers are adapted for container installation in terms of hardware structure, most lack specific intelligent energy-saving strategies for the unique scenarios of ocean shipping. Traditional power-saving modes are mostly triggered when the battery is low. However, during ocean shipping, containers are unable to obtain network and location data for most of the time, making it unnecessary for the device to operate at full capacity. The fundamental energy-saving strategy for ocean shipping trackers should involve putting the tracker into power-saving mode when setting sail and then deactivating it upon arrival at port, thereby reducing power consumption during maritime transport. However, current methods, whether through manual control or GPS positioning and matching with built-in electronic fences, cannot achieve accurate and flexible performance.
[0003] Regarding the shortcomings of existing long-haul container tracking technologies: (1) If the equipment is kept running at full capacity, simply increasing the battery capacity to alleviate range anxiety will increase hardware costs and structural installation difficulty, and the problem cannot be completely solved due to the uncertain shipping time.
[0004] (2) If the power saving mode is only entered when the voltage is low, there will be a lot of unnecessary power consumption waste during the sea transport process.
[0005] (3) If the equipment mode is switched manually before and after going to sea, it will increase labor costs.
[0006] (4) By setting up an electronic fence, the tracker can switch to power-saving mode according to the fence entry and exit. However, the GPS signal is unstable near the sea, and there may be cases where the device fails to be located after entering the electronic fence area, resulting in some devices not entering power-saving mode.
[0007] Therefore, existing technologies have shortcomings and need to be improved. Summary of the Invention
[0008] The technical problem to be solved by the present invention is to provide an intelligent energy-saving tracking system for ocean-going containers based on multimodal perception, so as to solve the problems mentioned in the background art.
[0009] The technical solution of the present invention is as follows: A multimodal perception-based intelligent energy-saving tracking system for ocean-going containers is provided, comprising: a terminal device installed on the container and a cloud management platform communicatively connected to the terminal device; the terminal device includes: a sensor layer, a data processing layer communicatively connected to the sensor layer, a core algorithm layer communicatively connected to the data processing layer, a control execution layer communicatively connected to the core algorithm layer, and a communication layer communicatively connected to the data processing layer; the terminal device communicates with the cloud management platform through the communication layer.
[0010] The sensor layer includes: a satellite positioning system and a three-axis accelerometer; the data processing layer includes: a data processing module, a vibration spectrum analysis module, and a network status monitoring module; the core algorithm layer includes: an offshore judgment system and a docking detection system; and the cloud management platform includes: an electronic fence database and a public land mobile network identifier feature library.
[0011] The satellite positioning system is used to obtain the satellite positioning of the container, the triaxial accelerometer is used to collect the vibration spectrum of the container; the signal processing module is used to process the data between the modules, the vibration spectrum analysis module is used to analyze the vibration spectrum, and the network status monitoring module is used to monitor the network connection status of the communication layer.
[0012] The offshore determination system is used to determine whether a container is in a maritime state. It achieves the determination of the offshore scenario through multimodal data fusion, specifically using three sets of information: Electronic fence trigger judgment: When the terminal device enters the predefined electronic fence sea area and continuously obtains valid satellite positioning system positioning for more than ten minutes, it is judged that the ship has left the port and entered the route state. Joint criteria for network status: The communication layer continuously detects no network signal, and the number of base station registration failures is ≥4 times, while the cumulative duration of no network is ≥10 minutes; preferably, the number of base station registration failures is ≥6 times, while the cumulative duration of no network is ≥18 minutes. Motion characteristic analysis criteria: Vibration spectrum is collected by a triaxial accelerometer. If the vibration energy in the 0.1-1Hz frequency band accounts for more than 70% and the peak acceleration is less than 0.5G, it is determined to be a vibration characteristic of ship navigation. Offshore scenario is triggered if any of the following conditions are met: Condition A: The satellite positioning system successfully locates the vessel and records the route of the vessel leaving the port and entering the electronic fence sea area; (i.e., electronic fence trigger judgment alone); Condition B: The satellite positioning system continues to fail and cannot obtain positioning, but the network state criterion and motion feature criterion are both met; (after the electronic fence trigger judgment fails, the network state joint criterion and motion feature analysis joint criterion are used).
[0013] The docking detection system is used to detect whether a container is docked (at the shore or stored in a warehouse). It determines the docking scenario through cross-validation of multi-source sensing data, specifically using three sets of information: Electronic fence detection: After repositioning by the satellite positioning system, the coordinates fall within the range of the port's electronic fence; Network location recovery criteria: The communication layer successfully registers with the local network and resolves the PLMN (Public Land Mobile Network Identifier) or Cell ID (Base Station Cell Identifier) to match the preset port area feature database; Vibration feature identification during loading and unloading: Periodic impact signals with a frequency range of 5-20Hz and a duration of ≥T1 (T1=30 seconds for example) are detected, where T1≥10 seconds, and the feature matches the operation mode of loading and unloading equipment (port hoisting, warehouse gantry crane); The vessel enters the docking scenario if any of the following conditions are met: Condition C: The satellite positioning system successfully locates the vessel and records the route from the electronic fence sea area to the port; Condition D: The network positioning recovery criterion and the loading and unloading vibration feature identification are met simultaneously.
[0014] The control execution layer is used to execute a tiered energy-saving strategy, which includes: Offshore scenario: The communication layer attempts to register with the network every 36-60 hours, and the satellite positioning system samples every 6-10 hours; Dock scenario: The satellite positioning system samples every 8-16 hours, and the communication layer attempts to register with the network every 8-16 hours.
[0015] The process for determining offshore scenarios is as follows: S11: Obtain positioning data from the satellite positioning system, network connectivity data, and data collected by the three-axis accelerometer. S12: Determine whether the satellite positioning system is valid and whether the area has entered the electronic fence sea area; if yes, enter the offshore scenario; if no, proceed to step S13. S13: Determine whether the cumulative number of network registration failures at the communication layer has reached 4 (preferably 6) and whether the continuous network outage time exceeds 10 minutes (preferably 18 minutes); if not, proceed to step S11; if yes, proceed to step S14. S14: Determine whether the vibration spectrum matches the navigation characteristics; if not, proceed to step S11; if yes, proceed to the offshore scenario.
[0016] The process for determining a docking scenario is as follows: S21: Obtain positioning data from the satellite positioning system, network connectivity data, and data collected by the three-axis accelerometer. S22: Determine whether the satellite positioning system is valid and whether the electronic fence port has been entered; if yes, enter the docking scenario; if no, proceed to step S23. S23: Determine whether the communication layer registration network has been restored and whether a matching PLMN and Cell ID have been resolved; if not, proceed to step S21; if yes, proceed to step S24. S24: Determine if the vibration spectrum matches the characteristics of loading and unloading vibration; if not, proceed to step S21; if yes, proceed to the docking scenario.
[0017] The core algorithm layer also includes an anomaly detection system. The trigger conditions for this system are: an impact with an instantaneous acceleration ≥ 1G, or the detection of high-frequency vibration lasting longer than 10 seconds. This anomaly detection system is used to handle sudden situations, such as container falling or container collisions.
[0018] The control execution layer is also used to execute anomaly state strategies, which are as follows: after the anomaly determination system is triggered, the satellite positioning system samples at intervals of 0.5 min to 2 min, and the communication layer attempts network registration every 0.5 min to 2 min. In the event of an anomaly, location information is uploaded as soon as possible for tracking.
[0019] The core algorithm layer also includes a land transport detection system. This system detects whether containers are in land transport status, and its judgment process is as follows: Land transport scenario determination is achieved through cross-validation of multi-source sensing data, specifically involving three sets of information: Electronic fence detection: After GPS repositioning, the coordinates fall within the electronic fence range that is not the docking location or the ocean. Network location recovery criteria: The cellular communication module successfully registers with the local network and resolves the PLMN (Public Land Mobile Network Identifier) or Cell ID (Base Station Cell Identifier) to match the preset non-docking area feature database; Vibration feature identification during transportation: Vibration signals with a frequency range of 5-20Hz and a duration of ≥T2 (T2 = 30 seconds for example) are detected, where T2 ≥ 10 seconds, and the feature matches the land transportation operation mode.
[0020] The following conditions are met to indicate that the vehicle has entered a land transportation scenario: Condition E: GPS positioning is successful and the route of the vehicle from one stop to another is recorded; Condition F: The network positioning recovery criterion and the land transportation vibration feature recognition are met simultaneously.
[0021] When in a land transportation scenario, the control execution layer also executes a land transportation scenario tracking strategy, which is: the satellite positioning system sampling interval is 3min-20min, and the communication layer attempts to register with the network every 3min-20min.
[0022] The satellite positioning system is at least one of GPS, BeiDou, GLONASS, and GALILEO.
[0023] The communication layer comprises 2G cellular modules, 3G cellular modules, 4G cellular modules, and 5G cellular modules. These are also known as 2G communication modules, 3G communication modules, 4G communication modules, and 5G communication modules.
[0024] The sensor layer also includes a temperature and humidity sensor, which is used to detect the temperature and humidity inside the container and the status of the items inside the container.
[0025] When containers are loaded onto the ship from the dock using large lifting equipment, a triaxial accelerometer detects vibrations matching the loading and unloading characteristics, identifying the current process as loading or unloading. After departure, if GPS positioning indicates the ship has entered a marked electronically fenced area or the network cannot register with the local dock network, the system switches to an offshore scenario. When GPS cannot locate the ship and cannot detect its entry into the electronically fenced area, the triaxial accelerometer sensor is used to determine if the ship's navigation characteristics are consistent. In the offshore scenario, due to the distance from the shore and the loss of signal caused by stacked containers, frequent positioning and attempts to search the net are not very meaningful; therefore, the device reduces the frequency of net searches and GPS positioning to conserve power.
[0026] When a vessel docks, GPS locates it after it leaves the designated electronically fenced area and enters the dock, and it registers with the local network, switching to the docking scenario. If either condition is not met—for example, if GPS fails to locate the vessel or network registration fails—a three-axis accelerometer is used to monitor the loading and unloading process to assist in the determination. In the docking scenario, the signal is better than at sea, but the location does not change frequently; therefore, the frequency of network search and GPS positioning can be slightly increased to update the data more promptly.
[0027] When containers are transported from the port to warehouses in various locations, it is necessary to monitor their transport trajectory and other statuses in a timely manner. Equipment should maintain relatively frequent location updates and reporting. A three-axis gravity accelerometer can be used to detect the vehicle's motion. When continuous vibration characteristics of the moving vehicle are detected, the system should be switched to land transport mode to meet the reporting frequency requirements. Once the container arrives at the warehouse and returns to a stationary state, high-frequency reporting is no longer needed, and the system should switch back to the docking scenario.
[0028] Switching between these different scenarios can greatly extend the lifespan of the device while meeting the usage requirements of the device in various environments, and alleviate battery anxiety in long-distance travel scenarios.
[0029] Since most ocean areas are without signal coverage, and voyages typically last for weeks or even longer than a month, this invention can accurately identify when a container has entered a maritime voyage and conserve power as much as possible during long voyages.
[0030] When containers are unloaded at a port during a voyage or after the voyage, they may remain there for several days. This invention can promptly and accurately determine the ship's docking time and update and report information such as its location, as well as adjust the reporting frequency.
[0031] As containers from the dock are distributed to warehouses in various locations via trucks and other land transportation methods, this invention can promptly identify the land transportation scenario, switch to real-time tracking and status monitoring of the trajectory, and ensure the safe and smooth progress of the land transportation process.
[0032] By adopting the above solution, this invention provides an intelligent energy-saving tracking system for ocean-going containers based on multimodal perception. This system enables switching of operating frequencies under different environments throughout the entire transportation process of ocean-going containers, directly addressing the pain point where customers can only choose between increasing battery capacity and reducing user experience. Without modifying the hardware structure or increasing additional labor costs, and while meeting the needs of different usage scenarios, the power consumption of the tracker is compressed to the extreme, achieving optimal resource utilization. Attached Figure Description
[0033] Figure 1 This is a functional block diagram of the present invention; Figure 2 A flowchart for ocean container shipping; Figure 3 Flowchart for determining offshore scenarios; Figure 4 Flowchart for determining the docking scenario. Detailed Implementation
[0034] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.
[0035] Please see Figures 1-4 This invention provides an intelligent energy-saving tracking system for ocean-going containers based on multimodal perception, comprising: a terminal device installed on the container and a cloud management platform communicatively connected to the terminal device; the terminal device includes: a sensor layer, a data processing layer communicatively connected to the sensor layer, a core algorithm layer communicatively connected to the data processing layer, a control execution layer communicatively connected to the core algorithm layer, and a communication layer communicatively connected to the data processing layer; the terminal device communicates with the cloud management platform through the communication layer.
[0036] The sensor layer includes: a satellite positioning system and a three-axis accelerometer; the data processing layer includes: a data processing module, a vibration spectrum analysis module, and a network status monitoring module; the core algorithm layer includes: a departure judgment system, a docking detection system, and a land transport detection system; the cloud management platform includes: an electronic fence database and a public land mobile network identifier feature database.
[0037] The satellite positioning system is used to obtain the satellite positioning of the container, the triaxial accelerometer is used to collect the vibration spectrum of the container; the signal processing module is used to process the data between the modules, the vibration spectrum analysis module is used to analyze the vibration spectrum, and the network status monitoring module is used to monitor the network connection status of the communication layer.
[0038] The offshore determination system is used to determine whether a container is in a maritime state. It achieves the determination of the offshore scenario through multimodal data fusion, specifically using three sets of information: Electronic fence trigger judgment: When the terminal device enters the predefined electronic fence sea area and continuously obtains valid satellite positioning system positioning for more than ten minutes, it is judged that the ship has left the port and entered the route state. Joint criteria for network status: The communication layer continuously detects no network signal, and the number of base station registration failures is ≥4 times, while the cumulative duration of no network is ≥10 minutes; In a preferred embodiment, the number of base station registration failures is ≥6 times, while the cumulative duration of no network is ≥18 minutes.
[0039] Motion characteristic analysis criteria: Vibration spectrum is collected by a triaxial accelerometer. If the vibration energy in the 0.1-1Hz frequency band accounts for more than 70% and the peak acceleration is less than 0.5G, it is determined to be a vibration characteristic of ship navigation. Offshore scenario is triggered if any of the following conditions are met: Condition A: The satellite positioning system successfully locates the vessel and records the route of the vessel leaving the port and entering the electronic fence sea area; (i.e., electronic fence trigger judgment alone); Condition B: The satellite positioning system continues to fail and cannot obtain positioning, but the network state criterion and motion feature criterion are both met; (after the electronic fence trigger judgment fails, the network state joint criterion and motion feature analysis joint criterion are used).
[0040] The docking detection system is used to detect whether a container is docked (at the shore or stored in a warehouse). It determines the docking scenario through cross-validation of multi-source sensing data, specifically using three sets of information: Electronic fence detection: After repositioning by the satellite positioning system, the coordinates fall within the range of the port's electronic fence; Network location recovery criteria: The communication layer successfully registers with the local network and resolves the PLMN (Public Land Mobile Network Identifier) or Cell ID (Base Station Cell Identifier) to match the preset port area feature database; Vibration feature identification during loading and unloading: Periodic impact signals with a frequency range of 5-20Hz and a duration of ≥T1 (T1 = 30 seconds for example) are detected, and the features are matched with the operation mode of loading and unloading equipment (port hoisting, warehouse gantry crane); The vessel enters the docking scenario if any of the following conditions are met: Condition C: The satellite positioning system successfully locates the vessel and records the route from the electronic fence sea area to the port; Condition D: The network positioning recovery criterion and the loading and unloading vibration feature identification are met simultaneously.
[0041] The land transport detection system is used to detect whether a container is in a land transport state. Its judgment process is as follows: Land transport scenario determination is achieved through cross-validation of multi-source sensing data, specifically using three sets of information: Electronic fence detection: After GPS repositioning, the coordinates fall within the electronic fence range that is not the docking location or the ocean. Network location recovery criteria: The cellular communication module successfully registers with the local network and resolves the PLMN (Public Land Mobile Network Identifier) or Cell ID (Base Station Cell Identifier) to match the preset non-docking area feature database; Vibration feature identification during transportation: Vibration signals with a frequency range of 5-20Hz and a duration of ≥T2 (T2 = 30 seconds for example) are detected, and the features are matched with the land transportation operation mode.
[0042] The following conditions are met to indicate that the vehicle has entered a land transportation scenario: Condition E: GPS positioning is successful and the route of the vehicle from one stop to another is recorded; Condition F: The network positioning recovery criterion and the land transportation vibration feature recognition are met simultaneously.
[0043] The control execution layer is used to execute a tiered energy-saving strategy, which includes: Offshore scenario: The communication layer attempts network registration every 36-60 hours, and the satellite positioning system samples every 6-10 hours; in a preferred embodiment, the communication layer attempts network registration every 48 hours, and the satellite positioning system samples every 8 hours. Docked scenario: The satellite positioning system samples every 8-16 hours, and the communication layer attempts network registration every 8-16 hours. In a preferred embodiment, the satellite positioning system samples every 12 hours, and the communication layer attempts network registration every 12 hours.
[0044] When in a land transportation scenario, the control execution layer also executes a land transportation scenario tracking strategy, which is: a satellite positioning system sampling interval of 3-20 minutes, and a communication layer attempting network registration every 3-20 minutes. In a preferred embodiment, the satellite positioning system sampling interval is 10 minutes, and the communication layer attempts network registration every 10 minutes.
[0045] The process for determining offshore scenarios is as follows: S11: Obtain positioning data from the satellite positioning system, network connectivity data, and data collected by the three-axis accelerometer. S12: Determine whether the satellite positioning system is valid and whether the area has entered the electronic fence sea area; if yes, enter the offshore scenario; if no, proceed to step S13. S13: Determine whether the cumulative number of network registration failures at the communication layer has reached 6, and whether the continuous network outage time exceeds 18 minutes; if not, proceed to step S11; if yes, proceed to step S14. S14: Determine whether the vibration spectrum matches the navigation characteristics; if not, proceed to step S11; if yes, proceed to the offshore scenario.
[0046] The process for determining a docking scenario is as follows: S21: Obtain positioning data from the satellite positioning system, network connectivity data, and data collected by the three-axis accelerometer. S22: Determine whether the satellite positioning system is valid and whether the electronic fence port has been entered; if yes, enter the docking scenario; if no, proceed to step S23. S23: Determine whether the communication layer registration network has been restored and whether a matching PLMN and Cell ID have been resolved; if not, proceed to step S21; if yes, proceed to step S24. S24: Determine if the vibration spectrum matches the characteristics of loading and unloading vibration; if not, proceed to step S21; if yes, proceed to the docking scenario.
[0047] The core algorithm layer also includes an anomaly detection system. The trigger conditions for this system are: an impact with an instantaneous acceleration ≥ 1G, or the detection of high-frequency vibration lasting longer than 10 seconds. This anomaly detection system is used to handle sudden situations, such as container falling or container collisions.
[0048] The control execution layer is also used to execute an abnormal state strategy, which is as follows: after the abnormal situation determination system is triggered, the satellite positioning system samples at an interval of 0.5 min to 2 min, and the communication layer attempts to register with the network every 0.5 min to 2 min. In the event of an abnormal situation, location information is uploaded as soon as possible for tracking. In a preferred embodiment, the satellite positioning system samples at an interval of 1 min, and the communication layer attempts to register with the network every 1 min.
[0049] The satellite positioning system is at least one of GPS, BeiDou, GLONASS, and GALILEO. In this embodiment, GPS is used as an example.
[0050] The communication layer is a 4G cellular module.
[0051] The sensor layer also includes a temperature and humidity sensor, which is used to detect the temperature and humidity inside the container and the status of the items inside the container.
[0052] When containers are loaded onto the ship from the dock using large lifting equipment, a triaxial accelerometer detects vibrations matching the loading and unloading characteristics, identifying the current process as loading or unloading. After departure, if GPS positioning indicates the ship has entered a marked electronically fenced area or the network cannot register with the local dock network, the system switches to an offshore scenario. When GPS cannot locate the ship and cannot detect its entry into the electronically fenced area, the triaxial accelerometer sensor is used to determine if the ship's navigation characteristics are consistent. In the offshore scenario, due to the distance from the shore and the loss of signal caused by stacked containers, frequent positioning and attempts to search the net are not very meaningful; therefore, the device reduces the frequency of net searches and GPS positioning to conserve power.
[0053] When a vessel docks, GPS locates it after it leaves the designated electronically fenced area and enters the dock, and it registers with the local network, switching to the docking scenario. If either condition is not met—for example, if GPS fails to locate the vessel or network registration fails—a three-axis accelerometer is used to monitor the loading and unloading process to assist in the determination. In the docking scenario, the signal is better than at sea, but the location does not change frequently; therefore, the frequency of network search and GPS positioning can be slightly increased to update the data more promptly.
[0054] When containers are transported from the port to warehouses in various locations, it is necessary to monitor their transport trajectory and other statuses in a timely manner. Equipment should maintain relatively frequent location updates and reporting. A three-axis gravity accelerometer can be used to detect the vehicle's motion. When continuous vibration characteristics of the moving vehicle are detected, the system should be switched to land transport mode to meet the reporting frequency requirements. Once the container arrives at the warehouse and returns to a stationary state, high-frequency reporting is no longer needed, and the system should switch back to the docking scenario.
[0055] Switching between these different scenarios can greatly extend the lifespan of the device while meeting the usage requirements of the device in various environments, and alleviate battery anxiety in long-distance travel scenarios.
[0056] Since most ocean areas are without signal coverage, and voyages typically last for weeks or even longer than a month, this invention can accurately identify when a container has entered a maritime voyage and conserve power as much as possible during long voyages.
[0057] When containers are unloaded at a port during a voyage or after the voyage, they may remain there for several days. This invention can promptly and accurately determine the ship's docking time and update and report information such as its location, as well as adjust the reporting frequency.
[0058] As containers from the dock are distributed to warehouses in various locations via trucks and other land transportation methods, this invention can promptly identify the land transportation scenario, switch to real-time tracking and status monitoring of the trajectory, and ensure the safe and smooth progress of the land transportation process.
[0059] In summary, this invention provides an intelligent energy-saving tracking system for ocean-going containers based on multimodal perception. It enables the switching of operating frequencies under different environments throughout the entire transportation process of ocean-going containers, directly addressing the pain point where customers can only choose between increasing battery capacity and reducing user experience. Without modifying the hardware structure or increasing additional labor costs, and while meeting the needs of different usage scenarios, the power consumption of the tracker is compressed to the extreme, achieving optimal resource utilization.
[0060] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A smart energy-saving tracking system for ocean-going containers based on multimodal perception, characterized in that, include: Terminal equipment installed on the container and a cloud management platform communicating with the terminal equipment; The terminal device includes: a sensor layer, a data processing layer communicatively connected to the sensor layer, a core algorithm layer communicatively connected to the data processing layer, a control execution layer communicatively connected to the core algorithm layer, and a communication layer communicatively connected to the data processing layer; the terminal device communicates with a cloud management platform through the communication layer. The sensor layer includes: a satellite positioning system and a triaxial accelerometer; the data processing layer includes: a data processing module, a vibration spectrum analysis module, and a network status monitoring module; the core algorithm layer includes: an offshore judgment system and a docking detection system; the cloud management platform includes: an electronic fence database and a public land mobile network identifier feature library. The satellite positioning system is used to obtain the satellite positioning of the container, the triaxial accelerometer is used to collect the vibration spectrum of the container; the signal processing module is used to process the data between the modules, the vibration spectrum analysis module is used to analyze the vibration spectrum, and the network status monitoring module is used to monitor the network connection status of the communication layer. The offshore determination system is used to determine whether a container is in a maritime state. It achieves the determination of the offshore scenario through multimodal data fusion, specifically using three sets of information: Electronic fence trigger judgment: When the terminal device enters the predefined electronic fence sea area and continuously obtains valid satellite positioning system positioning for more than ten minutes, it is judged that the ship has left the port and entered the route state. Joint criteria for network status determination: The communication layer continuously detects no network signal, and the number of base station registration failures is ≥4 times, while the cumulative duration of no network is ≥10 minutes; Motion characteristic analysis criteria: Vibration spectrum is collected by a triaxial accelerometer. If the vibration energy in the 0.1-1Hz frequency band accounts for more than 70% and the peak acceleration is less than 0.5G, it is determined to be a vibration characteristic of ship navigation. An offshore scenario is triggered if any of the following conditions are met: Condition A: The satellite positioning system successfully located the vessel and recorded its route from the port into the electronically fenced sea area; Condition B: The satellite positioning system continues to fail and cannot obtain positioning, but the network status criterion and motion characteristic criterion are both met simultaneously; The docking detection system is used to detect whether a container is docked. It determines the docking scenario through cross-validation of multi-source sensing data, specifically using three sets of information: Electronic fence detection: After repositioning by the satellite positioning system, the coordinates fall within the range of the port's electronic fence; Network location recovery criteria: The communication layer successfully registers with the local network and resolves the public land mobile network identifier or base station cell identifier to match the preset port area feature database; Vibration feature identification during loading and unloading: Periodic impact signals with a frequency range of 5-20Hz and a duration of ≥T1 are detected, where T1 ≥ 10 seconds, and the feature matches the operation mode of the loading and unloading equipment. Meeting any of the following conditions means that the device has entered the docking scenario: Condition C: The satellite positioning system successfully located the vessel and recorded the route of the vessel from the electronic fenced sea area to the port; Condition D: The network location recovery criterion and the loading and unloading vibration feature identification are both met simultaneously; The control execution layer is used to execute a tiered energy-saving strategy, which includes: Offshore scenario: The communication layer attempts to register with the network every 36-60 hours, and the satellite positioning system samples every 6-10 hours; Dock scenario: The satellite positioning system samples every 8-16 hours, and the communication layer attempts to register with the network every 8-16 hours.
2. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 1, characterized in that, The process for determining offshore scenarios is as follows: S11: Obtain positioning data from the satellite positioning system, network connectivity data, and data collected by the three-axis accelerometer. S12: Determine whether the satellite positioning system is valid and whether the area has entered the electronic fence sea area; if yes, enter the offshore scenario; if no, proceed to step S13. S13: Determine whether the cumulative number of network registration failures at the communication layer has reached 4, and whether the continuous network outage time exceeds 10 minutes; if not, proceed to step S11; if yes, proceed to step S14. S14: Determine whether the vibration spectrum matches the navigation characteristics; if not, proceed to step S11; if yes, proceed to the offshore scenario.
3. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception as described in claim 1, characterized in that: The process for determining a docking scenario is as follows: S21: Obtain positioning data from the satellite positioning system, network connectivity data, and data collected by the three-axis accelerometer. S22: Determine whether the satellite positioning system is valid and whether the electronic fence port has been entered; if yes, enter the docking scenario; if no, proceed to step S23. S23: Determine whether the communication layer registration network has been restored and whether a matching public land mobile network identifier or base station cell identifier has been resolved; if not, proceed to step S21; if yes, proceed to step S24. S24: Determine if the vibration spectrum matches the characteristics of loading and unloading vibration; if not, proceed to step S21; if yes, proceed to the docking scenario.
4. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 1, characterized in that, The core algorithm layer also includes an abnormal situation determination system; the triggering conditions for the abnormal situation determination system are: an impact with an instantaneous acceleration ≥ 1G, or the detection of high-frequency vibration lasting longer than 10s.
5. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 4, characterized in that, The control execution layer is also used to execute an abnormal state strategy, which is as follows: after the abnormal situation determination system is triggered, the satellite positioning system sampling interval is 0.5min-2min, and the communication layer attempts to register with the network every 0.5min-2min.
6. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 1, characterized in that, The core algorithm layer also includes a land transport detection system. This system detects whether containers are in land transport status, and its judgment process is as follows: Land transport scenario determination is achieved through cross-validation of multi-source sensing data, specifically involving three sets of information: Electronic fence detection: After GPS repositioning, the coordinates fall within the electronic fence range that is not the docking location or the ocean. Network location recovery criteria: The cellular communication module successfully registers with the local network and resolves the public land mobile network identifier or base station cell identifier to match the preset non-stopping area feature database; Vibration feature identification during transportation: Vibration signals with a frequency range of 5-20Hz and a duration of ≥T2 were detected, where T2 ≥ 10 seconds, and the features matched the land transportation operation mode. Meeting any of the following conditions means that the player has entered the land transport scenario: Condition E: GPS positioning is successful, and the route of the vehicle from one stop to another is recorded; Condition F: The network location recovery criterion and the land transport vibration feature identification are both met simultaneously.
7. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 6, characterized in that, When in a land transportation scenario, the control execution layer also executes a land transportation scenario tracking strategy, which is: the satellite positioning system sampling interval is 3-20 minutes, and the communication layer attempts to register with the network every 3-20 minutes.
8. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 1, characterized in that, The satellite positioning system is at least one of GPS, BeiDou, GLONASS, and GALILEO.
9. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 1, characterized in that, The communication layer includes 2G cellular modules, 3G cellular modules, 4G cellular modules, and 5G cellular modules.
10. The intelligent energy-saving tracking system for ocean-going containers based on multimodal perception according to claim 1, characterized in that, The sensor layer also includes a temperature and humidity sensor, which is used to detect the temperature and humidity inside the container and the status of the items inside the container.