Weather forecasts and navigation for vehicle routes

Abstract

This provides methods for weather forecasting and navigation along vehicle routes. [Solution] According to one aspect of the embodiment, the method may include acquiring first real-time observation data corresponding to the current conditions within the vehicle's operating area, the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the vehicle's operating area may be generated based on the first real-time observation data. The first prediction may be adjusted according to the operating area for the vehicle's operating area. The method may include generating a vehicle navigation route within the operating area based on the first prediction, the vehicle navigating the operating area according to the navigation route.

Description

[Technical Field]

[0001] The embodiments discussed in this disclosure relate to vehicle route determination based on customized weather forecasts. [Background technology]

[0002] Vehicle weather routing means planning the vehicle's route based on current and / or predicted weather conditions to avoid adverse weather conditions and / or optimize fuel efficiency. The quality of weather forecasts in the operating area covered by the vehicle route planning can significantly impact the effectiveness of the resulting route.

[0003] The subject matter claimed in this disclosure is not limited to embodiments that resolve any defects or embodiments that operate only in the environments described above. Rather, this background art is provided solely to illustrate an example of the technical field in which some of the embodiments described in this disclosure may be carried out. [Overview of the Initiative]

[0004] According to one aspect of the embodiment, the method may include acquiring first real-time observational data corresponding to the current conditions within the vehicle's operating area, the first real-time observational data being acquired based on the first real-time observational data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the vehicle's operating area may be generated based on the first real-time observational data. The first prediction may be adjusted for the vehicle's operating area according to that operating area. The method may include generating a vehicle navigation route within the operating area based on the first prediction, the vehicle navigating the operating area according to the navigation route.

[0005] The objectives and advantages of the embodiments will be realized and achieved by at least the elements, features, and combinations specifically indicated in the claims. Both the above summary and the following detailed description are illustrative and explanatory, and do not limit the claimed disclosures.

[0006] Exemplary embodiments will be described and explained with further identification and detail using the attached drawings. [Brief explanation of the drawing]

[0007] [Figure 1] This represents an exemplary system configured to generate navigation routes for a vessel, according to at least one embodiment of the present disclosure. [Figure 2] This is a flowchart illustrating an exemplary method of data assimilation arranged according to one or more embodiments of the present disclosure. [Figure 3] This diagram shows a flowchart illustrating an exemplary method for weather forecasting and route determination, arranged according to at least one embodiment of the present disclosure. [Figure 4] This diagram shows a block diagram of an exemplary computing system that may be used in conjunction with a weather forecasting and route determination system according to one or more embodiments of the present disclosure. [Modes for carrying out the invention]

[0008] A vehicle may travel along a navigation route to reach its destination. A navigation route is defined as the path or direction that leads a vehicle or its driver from one place to another. A vehicle includes any means of transport designed to carry people or goods from one place to another. For example, a vehicle can include various types such as ships, boats, automobiles, trucks, buses, bicycles, trains, airplanes, motorcycles, and off-highway vehicles. Navigation routes can vary depending on the type of vehicle. For example, a navigation route can include waterways, roads, railroads, air routes, and off-road trails. Navigation routes are determined based on various factors such as weather conditions, traffic conditions, road conditions, and vehicle type.

[0009] As an example, ships and other vessels may navigate waters based on navigation routes. Navigation routes can correspond to various routes and / or paths that a vessel may take to reach its destination. In some situations, navigation routes may be determined based on meteorological conditions such as storms, wind, waves, and tides. Determining navigation routes and other maritime operations based on meteorological conditions can help vessels avoid hazardous areas, reduce fuel consumption, and minimize the risk of delays and cargo damage. Furthermore, navigation routes determined based on meteorological conditions can help improve the efficiency of the navigation route so that greenhouse gas emissions may be reduced.

[0010] By forecasting or predicting weather conditions in future or upcoming areas and / or times, navigation routes can be determined taking weather conditions into account. For example, weather conditions corresponding to the area between the vessel and its destination can be predicted for the time of the vessel's voyage. Such predicted weather conditions can be used in route planning to identify and / or determine the optimal navigation route.

[0011] However, predicting weather conditions in various locations (e.g., at sea) can limit how efficiently and accurately real-time weather conditions can be predicted at a particular location. For example, ships and / or route-finding systems generally rely on external sources for weather condition predictions. Such external sources provide weather forecasts for the entire area, rather than weather forecasts specific to the area related to the ship. Furthermore, incorporating new observations and / or data in real time can be difficult due to the use of complex models that require enormous computing resources. Moreover, the prediction intervals from such external sources can be in the range of hours or days (e.g., due to the enormous computing resources used), which can limit real-time predictions based on real-time conditions.

[0012] In accordance with one or more embodiments of this disclosure, a weather forecasting and route determination system may be configured to enable the determination and / or adaptive modification of navigation routes based on real-time observational data. In particular, as described in detail in this disclosure, a weather forecasting and route determination system may be configured to utilize an artificial intelligence (AI) forecasting model configured to generate weather condition forecasts. In particular, the AI ​​forecasting model may generate weather condition forecasts based on real-time observational data. Such use of an AI forecasting model may help to generate timely and localized weather condition forecasts.

[0013] Weather forecasts can be used to improve route planning. For example, weather forecasts can lead to hazard recognition (e.g., identification of hazardous areas ahead), route adjustments (e.g., route changes to avoid areas with bad or hazardous weather conditions), and fuel efficiency (e.g., avoiding areas that require more fuel to pass through, such as areas with high waves). Weather forecasts generated based on real-time observational data can help determine safer and more fuel-efficient navigation routes.

[0014] Embodiments of this disclosure will be described with reference to the accompanying drawings.

[0015] Figure 1 shows an exemplary system 100 configured to generate navigation routes for a vessel according to at least one embodiment of the present disclosure. Generally, system 100 may be configured to generate navigation routes 116 ("Route 116") for a maritime vessel such as a ship based on weather condition forecasts. System 100 may be configured to generate Route 116 for any type of ship or vessel. System 100 may be implemented in-situ on a vessel. Additionally or alternatively, system 100 may be implemented remotely.

[0016] In some embodiments, system 100 may include an artificial intelligence (AI) predictive model 102 trained on a historical weather conditions dataset 104 ("historical dataset 104"). In some embodiments, the AI ​​predictive model 102 may be trained on entities associated with the operation of a particular vessel. In other embodiments, the AI ​​predictive model 102 may be trained on third-party entities not directly related to the operation of a particular vessel. In some embodiments, the AI ​​predictive model 102 may be trained for use by different vessels and / or entities. In some embodiments, the AI ​​predictive model 102 may include any suitable machine learning model, such as deep learning accelerators (DLA), neural networks, convolutional neural networks (CNN), regional convolutional neural networks (RCNN), generative models, physics-informed neural networks, transformers, and neural operator models.

[0017] In some embodiments, the historical dataset 104 may include historical ocean and meteorological analysis data obtained from various sources. For example, the historical dataset 104 may include historical ocean reanalysis data from WAVERYS and GLORYS data products provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). The historical dataset 104 may include such data over a period of time. For example, the historical dataset 104 may include meteorological condition data over a period of 5, 10, 15, 20 years, etc. In some embodiments, the historical dataset 104 may include changes in meteorological conditions over time at various locations. Such changes in meteorological conditions over time may include atmospheric pressure, direction of atmospheric pressure (rising or falling), rainfall, cloud cover, sea surface temperature, swell magnitude, wave height, current velocity, wave direction, storms, wind direction, wind speed, ocean currents and tides, lightning, precipitation, temperature, humidity, etc. In some embodiments, such changes in weather conditions over time may show how certain conditions at a particular location at a particular time are related to and / or influence conditions at a particular location at different times and / or conditions at different locations at various times.

[0018] In some embodiments, the AI ​​prediction model 102 may be tested and validated using a second historical dataset. In these and other embodiments, the second historical dataset may include historical weather condition data over periods that do not overlap with the period corresponding to the historical dataset 104. For example, in one case, the AI ​​prediction model 102 may be trained using a historical dataset containing weather condition data over 15 years (e.g., 1993–2018). In such a case, the second historical dataset may include weather condition data from years excluding the 15-year period (e.g., 2018–2022). Such a testing and validation process may help improve the accuracy and / or efficiency of the AI ​​prediction model 102.

[0019] The AI prediction model 102 can be trained and / or configured to receive data representing the current meteorological or oceanic state, such as wave height, flow velocity, wave direction, storm, wind direction, wind speed, ocean current and tide, rainfall, lightning, precipitation amount, air temperature, humidity, cloud cover, etc. The AI prediction model 102 can be configured to generate a prediction 108 that includes a forecast of the meteorological or oceanic conditions at the next time step (e.g., 1 hour later, 2 hours later, 3 hours later, 10 hours later, etc.) based on the current meteorological or oceanic state.

[0020] In some embodiments, the AI prediction model 102 can be trained to generate the prediction 108 considering the relationships between different geographical locations with respect to meteorological conditions. The above relationships can represent how the meteorological conditions at one geographical location affect the meteorological conditions at other geographical locations. For example, the rainfall and / or high waves at the first location may lead to rainfall, high waves, and / or other related meteorological conditions at the second location after a certain period of time. The AI prediction model 102 can predict the meteorological conditions at the second location at a future time based on the current meteorological conditions at the first location in addition to the current meteorological conditions at the second location.

[0021] In some embodiments, the historical dataset 104 can include meteorological condition data corresponding to the geographical area where the ship has sailed and / or the geographical area associated with the ship. For example, a specific ship may generally sail in a specific area (e.g., across the Pacific Ocean). The AI prediction model 102 associated with or used for a specific ship can be trained using the historical dataset 104 associated with the Pacific Ocean. Additionally, or alternatively, the AI prediction model 102 can be trained using historical datasets 104 from various locations in a general or geographically independent manner.

[0022] In some embodiments, the AI ​​prediction model 102 may be any type of suitable AI model, or may include any type of suitable AI model. In some embodiments, the AI ​​prediction model 102 may include a grid-independent and resolution-independent AI model. A grid-independent AI model learns mappings between function spaces rather than depending on a fixed grid. The AI ​​model may also transform spatial information into the frequency domain, where the representation of the function is independent of the underlying grid structure, allowing it to capture global features without being tied to specific grid points. A resolution-independent AI model learns to represent the function at any resolution, so it can approximate a solution regardless of the level of detail of the input data. Resolution independence may be useful in cases where high-resolution data may not be available. In these and other embodiments, the AI ​​prediction model 102 can take real-world observational data from various sources and generate predictions on a regularly structured grid. A regularly structured grid is a systematic arrangement of points in a multidimensional space where the spacing between points is uniform. The AI ​​prediction model 102 can take structured data on a regular grid and / or unstructured data on a regular grid and generate weather condition forecasts in a specific grid format suitable for route planning. As an example, the AI ​​prediction model 102 may include an FNO-based AI model.

[0023] In some embodiments, the AI ​​prediction model 102 may be configured to acquire real-time observational data 106. The real-time observational data 106 may include observational data relating to a specific location at a specific time (e.g., the current time). For example, the real-time observational data 106 may include observational data relating to the planned operating area of ​​a vessel. The planned operating area may include the area between the vessel and its destination. The area between the vessel and its destination is not limited to the direct route between the vessel and its destination, but includes any area that the vessel may pass through before reaching its destination. In these and other embodiments, the real-time observational data 106 may include current or real-time observations and meteorological conditions that are specifically relevant to the area between the vessel and its destination. Additionally or alternatively, the real-time observational data 106 may include observations or meteorological conditions at locations that may influence the meteorological conditions in the area between the vessel and its destination.

[0024] In some embodiments, the real-time observation data 106 may include various types of data acquired using various types of sensors or devices. For example, the real-time observation data 106 may include data acquired using buoys, floats, sail drones, gliders, satellite observations, offshore sites, etc. In these and other embodiments, the real-time observation data 106 may include sea surface temperature, wave height and period, wind speed and direction, atmospheric pressure, ocean current speed and direction, pressure at different depths, sea temperature, cloud volume and type, precipitation, etc. In some embodiments, the real-time observation data 106 may be acquired from one or more third-party sources associated with such devices or sensors (e.g., marine meteorological services, marine data providers, government and research institutions, coastal services, etc.). Additionally or alternatively, a vessel may include one or more onboard sensors configured to acquire real-time observation data 106 for an area near the vessel.

[0025] The AI ​​prediction model 102 may be configured to generate a prediction 108 based on real-time observational data 106. Using real-time observational data 106 that specifically corresponds to the planned operating area of ​​the vessel can improve the accuracy of the prediction 108 compared to obtaining a pre-generated generalized prediction for a wider area (e.g., a wider area including the planned operating area of ​​the vessel) and estimating a prediction for a specific part of that wider area (e.g., the planned operating area of ​​the vessel) from the generalized prediction. For example, estimating a prediction for a specific part from a generalized prediction may not provide accurate predictions or information because weather conditions may change within a wide area. The real-time observational data 106 can provide information to the AI ​​prediction model 102 to generate a curated prediction 108 for a specific part, i.e., the planned operating area of ​​the vessel.

[0026] Furthermore, real-time observational data 106 can help the AI ​​prediction model 102 generate predictions 108 in real time. For example, predictions 108 may be updated in real time based on the real-time observational data 106. In some embodiments, predictions 108 may include predicted weather conditions at various time steps in the planned operating area of ​​the vessel. For example, the AI ​​prediction model 102 may be configured to perform autoregressive predictions at specific time steps to periodically generate predictions 108 at those time steps. For example, the AI ​​prediction model 102 may be configured to periodically generate predictions 108 at regular time intervals (e.g., every 5 minutes, every 10 minutes, every 30 minutes, every hour, every 2 hours, etc.). Such predictions 108 generated based on real-time observational data 106 may allow predictions 108 to be periodically updated to reflect any changes in real-time weather conditions. Such predictions 108 are more accurate and reliable compared to predictions 108 generated at less frequent intervals (e.g., predictions 108 generated every few hours).

[0027] In some embodiments, the AI ​​prediction model 102 may generate predictions 108 with a first resolution or a first grid. For example, a prediction 108 for a particular area may be represented using a grid corresponding to that particular area. For example, a particular area may be covered using a grid with tiles of a particular size. In this case, each tile represents a prediction 108 for the area corresponding to each tile. In some embodiments, the prediction 108 may have a first resolution. In some embodiments, the first resolution may be coarse or relatively low. For example, the first resolution may be 20 km. 2 This may relate to low-resolution grids or tiles, such as the grid sizes mentioned above. In these and other embodiments, the AI ​​prediction model 102 may generate predictions 108 at such low resolution because there is insufficient high-resolution training data and / or to reduce the computational load on the AI ​​prediction model 102.

[0028] In some embodiments, the coarse-resolution prediction 108 may not be sufficiently adequate or detailed enough to generate the route 116 at the target accuracy level. In these and other embodiments, the system 100 may include an AI downscaling model 110. The AI ​​downscaling model 110 may include any suitable machine learning model, such as a deep learning accelerator (DLA), a neural network, a convolutional neural network (CNN), a region-based convolutional neural network (RCNN), a generative model, a neural network incorporating physical laws, a transformer, or a neural operator model.

[0029] The AI ​​downscaling model 110 may be configured to obtain a prediction 108 with a first resolution and generate a downscaled prediction 113. The downscaled prediction 113 may include a prediction 108 with a second resolution. In this case, the second resolution is higher than the first resolution. For example, the second resolution is 1 km. 2 The resolution is relatively higher than the first resolution, for example, with the following grid sizes: 2The grid is 1km 2 It can be divided into multiple grids of different sizes, so that each grid is 20km long. 2 It can represent changing forecasts or weather condition predictions within a grid.

[0030] The AI ​​downscaling model 110 may be trained using a first historical training dataset 111 containing weather condition data with a first resolution and a second historical training dataset 112 containing weather condition data with a second resolution. In some embodiments, the first historical training dataset 111 and the second historical training dataset 112 may correspond to historical weather condition data corresponding to roughly the same location and / or time. For example, the first historical training dataset 111 may represent weather condition data for a specific location at the first resolution, and the second historical training dataset 112 may represent weather condition data for that specific location at the second resolution. The AI ​​downscaling model 110 may be trained to acquire weather condition data and / or forecasts 108 at the first resolution and generate downscaled forecasts 113 at the second resolution.

[0031] In some embodiments, the system 100 may include a routing module 114 configured to generate routes 116 based on a downscaled forecast 113. For example, the routing module 114 may be configured to determine one or more routes 116 between a vessel and a destination. The routing module 114 may be configured to generate routes 116 such that one or more routes 116 are optimized with respect to the downscaled forecast 113. For example, a route 116 may avoid areas with weather conditions that could affect the vessel's operation. For example, a route 116 may avoid areas experiencing high waves, storms, etc., that could affect the vessel's operation. In these and other embodiments, the vessel may operate within a planned operating area based on one or more routes 116.

[0032] In some embodiments, the routing module 114 may include code and routines configured to enable a computing system to perform one or more operations corresponding to the routing module 114. Additionally or alternatively, the routing module 114 may be implemented using hardware including one or more processors, microprocessors, microcontrollers, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other digital or analog circuits configured to interpret and / or execute program instructions and / or process data. Furthermore, references to operations performed by the routing module 114 may also include operations that the routing module 114 causes other components to perform.

[0033] In some embodiments, route 116 may be modified and / or regenerated based on newly acquired real-time observational data. For example, weather conditions may change after the routing module 114 has determined route 116, which may necessitate a new and / or modified route. In such cases, the AI ​​prediction model 102 may be configured to regenerate the prediction 108 based on newly acquired real-time observational data, thereby enabling the routing module 114 to modify and / or regenerate the route. The operation relating to route regeneration and / or modification may be described in more detail in reference to Figure 2 of this disclosure.

[0034] System 100 may be modified, added to, or deleted without departing from the scope of this disclosure. For example, in some embodiments, System 100 may include any number of other components that are not expressly illustrated or described.

[0035] For example, Figure 1 is generally described in relation to a maritime vessel, but the same principles described herein may be used for route planning of other vehicles. For example, forecast 108 may be generated for cars, trucks, aircraft, etc. Forecast 108 may be used to generate specific routes that may be safe and / or efficient for cars, trucks, aircraft, etc. For example, in response to forecast 108 indicating the presence of a severe storm, a navigation route for an aircraft may be determined or modified to avoid the storm.

[0036] Figure 2 shows a flowchart of an exemplary method 200 for navigation route planning based on real-time weather data, arranged according to at least one embodiment of the present disclosure. In particular, method 200 may include deciding to supply or take in additional real-time weather data. In some embodiments, one or more operations of method 200 may be performed by any suitable element of a weather forecasting and route determination system, such as system 100 in Figure 1. Although represented as separate steps, the various steps of method 200 may be divided into further steps, combined into fewer steps, or omitted, depending on the desired implementation. Furthermore, the order in which the different steps are performed may vary depending on the desired implementation.

[0037] In block 202, first real-time observation data may be acquired. In some embodiments, the first real-time observation data may correspond to the real-time observation data 106 in Figure 1. The first real-time observation data may include meteorological condition data corresponding to the ship's planned operating area. In some embodiments, the first real-time observation data may also include meteorological condition data outside the planned operating area that may affect the meteorological conditions in the planned operating area.

[0038] In block 204, the AI ​​prediction model may be configured to generate predictions or weather condition forecasts based on first real-time observation data. The AI ​​prediction model may correspond to AI prediction model 102 in Figure 1. The AI ​​prediction model may be trained and / or configured to generate predictions or weather condition forecasts for a certain area based on first real-time observation data. In some embodiments, the prediction may correspond to prediction 108 in Figure 1.

[0039] In block 206, an AI downscaling model (AI downscaling model 110 in Figure 1) may acquire predictions generated using an AI prediction model and downscale the predictions. For example, the AI ​​downscaling model may be configured to increase and / or improve the resolution of the predictions so that the predictions are more suitable for route generation. The AI ​​downscaling model may be trained using a first historical training dataset with a first resolution and a second historical training dataset with a second resolution, where the second resolution is higher than the first resolution. In some embodiments, the first and second historical training datasets may be associated with roughly the same geographical location and time, so that the second historical training dataset contains substantially the same weather condition data as the first historical training dataset, but with a higher resolution.

[0040] In Block 208, one or more ship settings may be obtained. A ship setting may include various factors of the ship that may influence route determination, such as the ship's design, capabilities, and / or operational characteristics. Different ship settings may impose different limitations and / or constraints on route determination. For example, different types and sizes of maritime vessels may have different draft limits, bridge clearances, and / or port access restrictions. Different ships may have different engine types and fuel efficiencies. Ships may also have various navigational constraints, such as shallow waters and narrow straits. Additionally or alternatively, different ships may be subject to different regulations and / or safety requirements. A ship setting may include any other characteristics of the ship that may influence the ship's navigation route determination.

[0041] In block 210, one or more navigation routes may be generated based on downscaled forecasts and one or more ship settings. For example, the downscaled forecast may provide predicted weather conditions in different areas that may affect route planning. The downscaled forecast may provide such predicted weather conditions in more detail or granularity. In some embodiments, one or more navigation routes may be generated using a routing module, such as routing module 114 in Figure 1. In some embodiments, one or more navigation routes may represent routes that a ship may take to reach its destination. One or more navigation routes may represent routes that are optimized in terms of the safety and efficiency of navigation by the ship, based on forecasts and ship settings. For example, a navigation route may include areas related to forecasts that help or do not adversely affect the operation of the ship. For example, a navigation route may avoid areas that would adversely affect the operation of the ship. In these and other embodiments, the downscaled forecast may allow the routing module to avoid areas in smaller bits, by providing various forecasts for smaller areas.

[0042] In block 212, user interaction with one or more navigation routes may be enabled by the user interface. For example, one or more navigation routes may be presented on the user interface to enable user interaction. In some embodiments, the user interface may include a computing device with a display for displaying one or more navigation routes. In some embodiments, the user interface may be installed on a ship. In other embodiments, the user interface may be installed in a remote location. In some embodiments, user interaction may include actions to change and / or modify one or more navigation routes.

[0043] For example, in block 214, the user may introduce and / or add constraints and / or additional data to be supplied to the routing module or AI prediction model via the user interface. For example, the user may supply one or more spatial domain constraints to the routing module. Spatial domain constraints may include constraints specific to the domain or area that the routing module should consider. For example, spatial domain constraints may include specific areas to be avoided (e.g., areas unsuitable for navigation and / or known dangerous areas) and / or priority areas. In response to the user providing new spatial domain constraints, the routing module may regenerate and / or modify one or more routes to take the new spatial domain constraints into account.

[0044] In some embodiments, the user may determine that the initial observational data has become invalid due to changes in weather conditions and / or the passage of time. In these and other embodiments, in block 216, the user may decide whether to supply new observational data to the AI ​​prediction model. In response to deciding to supply new observational data, the user may specify a time step in which the new observational data may be acquired. The time step may specify a particular time for acquiring the new observational data. For example, the particular time may include every hour, every three hours, every five hours, etc. In some embodiments, the time step may specify a one-time opportunity to acquire the new observational data. In response to deciding that new observational data is not needed (e.g., no substantial change in weather is observed), one or more navigation routes may not need to be changed.

[0045] In response to a user instructing the acquisition of new observational data, new or second real-time observational data may be acquired in block 220. In some embodiments, the second real-time observational data may be acquired from the same source as the first real-time observational data. In other embodiments, the second real-time observational data may be acquired from a different source than the first real-time observational data.

[0046] In block 222, model-free data assimilation may be performed. Model-free data assimilation may include incorporating second real-time observational data into an AI prediction model to improve predictions or forecasts without relying on complex computational models. In these and other embodiments, any suitable data assimilation technique or method may be used. For example, in some embodiments, data fusion using ensemble Kalman filtering (EnKF) techniques may be used to feed second real-time observational data into an AI prediction model, thereby enabling the AI ​​prediction model to regenerate new fields to maintain the quality of autoregressive predictions for a longer period. In some embodiments, other data assimilation techniques such as ensemble prediction, Bayesian data fusion, and kernel methods may be used.

[0047] In response to the assimilation of the second real-time observation data into the AI ​​prediction model, in block 224, the user may initiate a rerun of the AI ​​prediction model via the user interface. For example, the AI ​​prediction model may be rerun to generate new and / or modified weather condition forecasts based on the second real-time observation data.

[0048] In some embodiments, the steps or blocks of Method 200 may be repeated until a satisfactory navigation route is generated, the user is no longer aware of significant changes in weather conditions, and / or the vessel reaches its destination. In some embodiments, the steps or blocks of Method 200 may be performed without user interaction. For example, in some embodiments, the steps of Method 200 may be automatically repeated at specific time intervals. For example, new real-time data (e.g., second real-time observation data) may be periodically acquired at regular time intervals (e.g., every 5 minutes, every 10 minutes, every 15 minutes, every 30 minutes, every hour, every 2 hours, etc.).

[0049] In addition, or alternatively, in some embodiments, routes may be assigned a safety and / or efficiency score (a number from 0 to 100, where a higher number indicates greater safety and / or efficiency). For example, a scoring module or scoring AI model may be used to assign scores to one or more navigation routes. The scores may be compared to predefined safety thresholds and / or predefined efficiency thresholds. In response to a score not meeting the threshold (e.g., below the threshold), method 200 may proceed to (e.g., block 216) supply new observational data to an AI prediction model.

[0050] Method 200 may be modified, added to, or deleted without departing from the scope of this disclosure. For example, the operations of Method 200 may be performed in a different order. Additionally or alternatively, two or more operations may be performed simultaneously. Furthermore, the operations and activities described are given as examples, and some of the operations and activities may be optional, combined into fewer operations and activities, or expanded into further operations and activities, without deviating from the essence of the disclosed embodiments.

[0051] For example, Figure 2 is described in relation to a vessel (e.g., a sea vessel), but the same principles described herein may be used for route planning of other vehicles. For example, forecast 108 may be generated in relation to cars, trucks, aircraft, etc. Forecast 108 may be used to generate specific routes that may be safe and / or efficient for cars, trucks, aircraft, etc. For example, in response to forecast 108 indicating the presence of a severe storm, the navigation route for an aircraft may be determined or modified to avoid the storm.

[0052] Figure 3 shows a flowchart of an exemplary method 300 for weather forecasting and route determination, arranged according to at least one embodiment of the present disclosure. One or more operations of method 300 may be performed by any suitable system, such as system 100 in Figure 1 and / or computing system 400 in Figure 4. Although represented as separate steps, the various steps of method 300 may be divided into further steps, combined into fewer steps, or omitted, depending on the desired implementation. Furthermore, the order in which the different steps are performed may vary depending on the desired implementation. In some embodiments, method 300 may be described in relation to vehicles. Vehicles described in method 300 may include various types of vehicles, such as ships, vessels, aircraft, automobiles, and motorcycles.

[0053] In some embodiments, method 300 may include block 302. In block 302, an artificial intelligence (AI) prediction model may be trained to predict weather using a historical weather conditions dataset. The AI ​​prediction model may correspond to the AI ​​prediction model 102 in Figure 1, and the historical weather conditions dataset may correspond to the first historical training dataset 111 in Figure 1. In some embodiments, weather conditions may include wave height, current velocity, wave direction, storms, wind direction, wind speed, ocean currents and tides, rain, thunderstorms, precipitation, temperature, humidity, or cloud cover.

[0054] In block 304, first real-time observational data can be acquired that corresponds to the current conditions within the vehicle's planned operating area. The current conditions may include various meteorological and / or oceanic conditions at the time of observation. In some embodiments, the first real-time observational data may be acquired based on first real-time data that specifically corresponds to the vehicle's planned operating area. The planned operating area may include the area between the vehicle and its target destination. For example, the planned operating area may include the area to which the vehicle may travel in order to reach its target destination.

[0055] In some embodiments, real-time observational data may include current conditions at locations that may not be within the vehicle's planned operating area but could influence weather conditions within that area. For example, weather conditions at a first location (e.g., outside the vehicle's operating area) could influence a second location (e.g., within the vehicle's operating area) after a certain period of time. For instance, a storm at the first location might generally move to the second location after a certain period. In such cases, real-time observational data could include current conditions at the first location.

[0056] In block 306, a first forecast of one or more weather conditions associated with the vehicle's planned operating area may be generated using an AI forecasting model based on first real-time observation data. In some embodiments, the first forecast generated using the AI ​​forecasting model may be specifically tailored to the vehicle's planned operating area.

[0057] In some embodiments, the first forecast may include meteorological condition data at a first resolution. In some embodiments, the first resolution may be relatively low or coarse. For example, the level of detail in the data for the first forecast may be low or sparse. In particular, the first forecast may be generated based on a large grid size or dimension where each cell in the grid represents a specific geographic coordinate. For example, the first forecast may be based on a grid where each cell represents 20 km 2 It may be represented by a grid showing the geographical area and / or corresponding coordinates.

[0058] In some embodiments, the first prediction may be downscaled. For example, the grid size of the first prediction may be reduced so that the resolution of the first prediction may be improved. In some embodiments, an AI downscaling model may be used to downscale the first prediction. The AI ​​downscaling model may be trained using a first historical training dataset containing meteorological condition data with a first resolution and a second historical training dataset containing meteorological condition data with a second resolution. In these and other embodiments, the first resolution is lower than the second resolution. For example, the second resolution may be represented using a smaller grid size than the first resolution to represent meteorological conditions with more detail. For example, the second resolution may be 1 km. 2 It can be displayed with a grid size of [size].

[0059] In these and other embodiments, the AI ​​downscaling model may be configured to obtain a first prediction at a first resolution from the AI ​​prediction model. The AI ​​downscaling model may convert the first resolution of the first prediction to a second resolution. The downscaled prediction is more suitable for route generation because a more accurate route is determined based on prediction data with a higher resolution or more detail.

[0060] In block 308, one or more navigation routes for a vehicle in an operating area may be generated based on a first prediction, and the vehicle navigates the planned operating area according to one or more navigation routes. In some embodiments, one or more navigation routes may be presented in a user interface to allow the user to interact with one or more navigation routes. In some embodiments, the user interface may present additional information about one or more navigation routes, such as estimated time and fuel efficiency.

[0061] Method 300 may be modified, added to, or deleted without departing from the scope of this disclosure. For example, the operations of Method 300 may be performed in a different order. Additionally or alternatively, two or more operations may be performed simultaneously. Furthermore, the operations and activities described are given as examples, and some of the operations and activities may be optional, combined into fewer operations and activities, or expanded into further operations and activities, without deviating from the essence of the disclosed embodiments.

[0062] For example, Figure 3 may further include obtaining user input via a user interface indicating a selection to acquire second real-time observation data. In some embodiments, the second real-time observation data may correspond to current conditions within one or more navigation routes. The user may instruct to acquire the second real-time observation data in response to identifying a significant or noticeable change in weather conditions that necessitates a new forecast.

[0063] In response to user selection, second real-time observational data corresponding to the current conditions within one or more navigation routes may be acquired. An AI prediction model may generate second predictions for one or more weather conditions associated with one or more navigation routes, based on or considering the second real-time observational data. In these and other embodiments, based on the second predictions, one or more updated navigation routes may be generated for a vehicle in a planned area of ​​operation, and the vehicle navigates the area according to one or more updated navigation routes.

[0064] Method 300 may be applied to various types of vehicles. For example, Method 300 may be applied to boats, boards, automobiles, trucks, buses, bicycles, trains, aircraft, motorcycles, off-highway vehicles, etc. The navigation route may vary depending on the type of vehicle. For example, the navigation route may include waterways, roads, railways, sea routes, off-road trails, etc.

[0065] Figure 4 shows a block diagram of an exemplary computing system 400 that may be used in relation to a weather forecasting and route determination system according to at least one embodiment of the present disclosure. For example, the computing system 400 may be used to generate a vehicle navigation route.

[0066] The computing system 400 may include a processor 410, memory 412, data storage 414, and a user interface 416. The processor 410, memory 412, data storage 414, and user interface 416 may be connected in a communicative manner.

[0067] Generally, the processor 410 may include any suitable special-purpose or general-purpose computer, computing entity, or processing device, including various computer hardware or software modules, and may be configured to execute instructions stored in any applicable computer-readable storage medium. For example, the processor 410 may include a microprocessor, microcontroller, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other digital or analog circuitry configured to interpret and / or execute program instructions and / or process data. Although represented as a single processor in Figure 4, the processor 410 may include any number of processors configured to individually or collectively perform or direct any number of the operations described herein. Furthermore, one or more processors may reside in one or more different electronic devices, such as different servers.

[0068] In some embodiments, the processor 410 may be configured to interpret and / or execute program instructions stored in memory 412, data storage 414, or both memory 412 and data storage 414, and / or process stored data. In some embodiments, the processor 410 may fetch program instructions from data storage 414 and load them into memory 412. After the program instructions are loaded into memory 412, the processor 410 may execute the program instructions.

[0069] The memory 412 and data storage 414 may include computer-readable storage media that carry or store computer-executable instructions or data structures. Such computer-readable storage media may include any available media accessible by a general-purpose or special-purpose computer, such as the processor 410. Such computer-readable storage media may include, but are not limited to, tangible or non-temporary computer-readable storage media, including random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), compact disk read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage or other magnetic storage devices, flash memory devices (e.g., solid-state memory devices), or other storage media accessible by a general-purpose or special-purpose computer that can be used to carry or store specific program code in the form of computer-executable structures or data structures. Combinations of the above may also be included within the scope of computer-readable storage media. Computer-executable instructions may include, for example, instructions and data configured to cause the processor 410 to perform a specific operation or group of operations.

[0070] The user interface 416 may include any device that enables the user to interface with the computing system 400. For example, the user interface 416 may include, among many other devices, a mouse, a trackpad, a keyboard, buttons, a camera, and / or a touchscreen. The user interface 416 may receive user input and supply the input to the processor 410.

[0071] Modifications, additions, or deletions can be made to the computing system 400 without exceeding the scope of this disclosure. For example, in some embodiments, the computing system 400 may include any number of other components that are not expressly illustrated or described.

[0072] In this disclosure, the terms used in particular in the attached claims (e.g., the text of the attached claims) are generally intended to be “open” terms (for example, the word “including” should be interpreted as “including but not limited to,” the word “having” should be interpreted as “at least having,” and the word “includes” should be interpreted as “including but not limited to,” etc.).

[0073] Furthermore, if a specific number is intended in an introduced claim recitation, that intention must be clearly stated in the claim; if there is no such statement, then no such intention exists. For example, to facilitate understanding, subsequent appended claims may use introductory phrases such as "at least one" and "one or more" to introduce a claim. However, the use of such phrases should not be interpreted as suggesting that a particular claim containing the introduced claim recitation is limited to cases that include only one instance of that item, even if the claim includes both an introductory phrase such as "one or more" or "at least one" and an indefinite article such as "a" or "an" (for example, "a" and / or "an" should be interpreted as meaning "at least one" or "one or more"). The same applies when introducing a claim recitation using a definite article.

[0074] Furthermore, even if a specific number is explicitly stated in the introduced claim description, it will be understood by those skilled in the art that such description should generally be interpreted to mean at least the number stated (for example, if there is a description of only "two descriptions" without any other modifiers, this description means at least two descriptions, or two or more descriptions). Furthermore, when a description similar to "at least one of A, B, and C, etc." or "one or more of A, B, and C, etc." is used, such a structure is generally intended to include A only, B only, C only, both A and B, both A and C, both B and C, and / or all of A, B, and C, etc. For example, the use of the term "and / or" is intended to be interpreted in this way.

[0075] Furthermore, any disjunctions and / or disjunctions representing two or more selectable terms, whether in the specification, claims, or drawings, should be understood as intended to include the possibility of including one of those terms, either of those terms, or both of those terms. For example, the phrase "A or B" should be understood to include the possibility of "A or B" or "A and B".

[0076] All examples and conditional language cited herein are intended for educational purposes to help readers understand the concepts and inventions to which the inventors have contributed to the advancement of the art, and should be construed as not being limited to such specifically cited examples and conditions. While embodiments of this disclosure have been described in detail, various modifications, substitutions, and alternatives may be made without departing from the spirit and scope of this disclosure.

[0077] In addition to the embodiments described above, the following additional information is disclosed. (Note 1) The process involves acquiring first real-time observation data corresponding to the current situation within the vehicle's operating area, and the first real-time observation data is acquired based on the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the operating area of ​​the vehicle is generated based on the first real-time observation data, and the first prediction is adjusted for the operating area of ​​the vehicle according to the operating area. Based on the first prediction, one or more navigation routes for the vehicle in the operating area are generated, and the vehicle navigates the operating area according to the one or more navigation routes. A method of having. (Note 2) Before acquiring the first real-time observation data corresponding to the current status of the vehicle within the operating area, The method further includes training the AI ​​prediction model to predict weather conditions using a dataset of past weather conditions. The method described in Appendix 1. (Note 3) To acquire second real-time observation data corresponding to the current situation within one or more of the aforementioned navigation routes, Using the AI ​​prediction model, generate a second prediction of one or more weather conditions associated with one or more navigation routes based on the second real-time observation data, Based on the second prediction, generate one or more updated navigation routes for the vehicle in the operating area, and the vehicle navigates the operating area according to the one or more updated navigation routes. The method described in Appendix 1, further comprising the above. (Note 4) Before acquiring the second real-time observation data, The system further includes obtaining user input via a user interface indicating a selection to acquire the second real-time observation data. The method described in Appendix 3. (Note 5) The user input indicates a time step for supplying the second real-time observation data to the AI ​​prediction model. The method described in Supplementary Note 4. (Supplementary Note 6) Before generating the one or more navigation routes, further comprising downscaling a first set of predictions using an AI downscaling model, The method described in Supplementary Note 1. (Supplementary Note 7) Said downscaling of the first set of predictions is training the AI downscaling model using a first past training dataset including weather condition data having a first resolution and a second past training dataset including weather condition data having a second resolution, wherein the first resolution is lower than the second resolution, and obtaining the first set of predictions at the first resolution from the AI prediction model, and converting the first resolution of the first set of predictions to the second resolution using the AI downscaling model including, The method described in Supplementary Note 6. (Supplementary Note 8) The weather condition data having the first resolution includes data having a grid size equal to or exceeding 20 km 2 equal to 20 km 2 including data having a grid size equal to or exceeding 20 km The method described in Supplementary Note 7. (Supplementary Note 9) The weather condition data having the second resolution includes data having a grid size equal to or smaller than 1 km 2 equal to 1 km 2 including data having a grid size equal to or smaller than 1 km The method described in Supplementary Note 7. (Supplementary Note 10) The weather conditions include one or more of wave height, flow velocity, wave direction, storm, wind direction, wind speed, ocean current and tide, rain, thunder, precipitation, temperature, humidity, cloud cover. The method described in Supplementary Note 1. (Supplementary Note 11) having one or more processors, the one or more processors being The process involves acquiring first real-time observation data corresponding to the current situation within the vehicle's operating area, and the first real-time observation data is acquired based on the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the operating area of ​​the vehicle is generated based on the first real-time observation data, and the first prediction is adjusted for the operating area of ​​the vehicle according to the operating area. Based on the first prediction, one or more navigation routes for the vehicle in the operating area are generated, and the vehicle navigates the operating area according to the one or more navigation routes. A system configured to perform actions that include the following. (Note 12) The above operation is performed before acquiring the first real-time observation data corresponding to the current status of the vehicle within the operating area, This further includes training the AI ​​prediction model to predict weather conditions using a dataset of past weather conditions. The system described in Appendix 11. (Note 13) The aforementioned operation is, To acquire second real-time observation data corresponding to the current situation within one or more of the aforementioned navigation routes, Using the AI ​​prediction model, generate a second prediction of one or more weather conditions associated with one or more navigation routes based on the second real-time observation data, Based on the second prediction, generate one or more updated navigation routes for the vehicle in the operating area, and the vehicle navigates the operating area according to the one or more updated navigation routes. The system described in Appendix 11, further including the above. (Note 14) The above operation is performed before acquiring the second real-time observation data. The further includes obtaining user input via a user interface indicating a selection to acquire the second real-time observation data, The system described in Appendix 13. (Note 15) The user input indicates a specific time step for supplying the second real-time observation data to the AI ​​prediction model. The system described in Appendix 14. (Note 16) The above operation is performed before generating the one or more navigation routes. This further includes downscaling the first set of predictions using an AI downscaling model. The system described in Appendix 11. (Note 17) Downscaling the aforementioned first set of predictions is The method involves training the AI ​​downscaling model using a first historical training dataset containing weather condition data with a first resolution and a second historical training dataset containing weather condition data with a second resolution, wherein the first resolution is lower than the second resolution. Obtaining the first set of predictions from the AI ​​prediction model with the first resolution, Using the AI ​​downscaling model, the first resolution of the first set of predictions is converted to the second resolution. including, The system described in Appendix 16. (Note 18) The weather condition data having the first resolution is 20km 2 Equivalent to or 20km 2 Including data with grid sizes exceeding, The system described in Appendix 17. (Note 19) The meteorological condition data having the second resolution is 1 km 2 Equivalent to or 1 km 2 Includes data with a grid size smaller than, The system described in Appendix 17. (Note 20) One or more non-temporary computer-readable media storing instructions, When the aforementioned instruction is executed by one or more processors, it causes the system to perform an action. The aforementioned operation is, The process involves acquiring first real-time observation data corresponding to the current situation within the vehicle's operating area, and the first real-time observation data is acquired based on the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the operating area of ​​the vehicle is generated based on the first real-time observation data, and the first prediction is adjusted for the operating area of ​​the vehicle according to the operating area. Based on the first prediction, one or more navigation routes for the vehicle in the operating area are generated, and the vehicle navigates the operating area according to the one or more navigation routes. One or more non-temporary computer-readable media, including [the specified text]. [Explanation of Symbols]

[0078] 100 Systems 102 Artificial Intelligence (AI) Predictive Models 104 Historical weather condition datasets 106 Real-time observation data 108 predictions 110 AI Downscaling Models 111 First historical training dataset 112 Second historical training dataset 113 Downscaled Predictions 114 Routing Modules 116 Navigation Route 400 Computing Systems 410 Processor 412 memory 414 Data Storage 416 User Interface

Claims

1. The process involves acquiring first real-time observation data corresponding to the current conditions within the vehicle's operating area, and the first real-time observation data is acquired based on the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the operating area of ​​the vehicle is generated based on the first real-time observation data, and the first prediction is adjusted for the operating area of ​​the vehicle according to the operating area. Based on the first prediction, one or more navigation routes for the vehicle in the operating area are generated, and the vehicle navigates the operating area according to the one or more navigation routes. A method of having.

2. Before acquiring the first real-time observation data corresponding to the current status of the vehicle within the operating area, The method further comprises training the AI ​​prediction model to predict weather conditions using a dataset of past weather conditions. The method according to claim 1.

3. To acquire second real-time observation data corresponding to the current situation within one or more of the aforementioned navigation routes, Using the AI ​​prediction model, generate a second prediction of one or more weather conditions associated with one or more navigation routes based on the second real-time observation data, Based on the second prediction, generate one or more updated navigation routes for the vehicle in the operating area, and the vehicle navigates the operating area according to the one or more updated navigation routes. The method according to claim 1, further comprising:

4. Before acquiring the second real-time observation data, The system further includes obtaining user input via a user interface indicating a selection to acquire the second real-time observation data. The method according to claim 3.

5. The user input indicates the time step for supplying the second real-time observation data to the AI ​​prediction model. The method according to claim 4.

6. Before generating one or more of the aforementioned navigation routes, The method further involves downscaling the first set of predictions using an AI downscaling model. The method according to claim 1.

7. Downscaling the first set of predictions mentioned above is: The method involves training the AI ​​downscaling model using a first historical training dataset containing weather condition data with a first resolution and a second historical training dataset containing weather condition data with a second resolution, wherein the first resolution is lower than the second resolution. Obtaining the first set of predictions from the AI ​​prediction model at the first resolution, Using the AI ​​downscaling model, the first resolution of the first set of predictions is converted to the second resolution. including, The method according to claim 6.

8. The weather condition data having the first resolution is 20 km ２ Equivalent to or 20 km ２ Including data with grid sizes exceeding, The method according to claim 7.

9. The meteorological condition data having the second resolution is 1 km ２ Equivalent to or 1 km ２ Includes data with a grid size smaller than, The method according to claim 7.

10. The aforementioned weather conditions include one or more of the following: wave height, current velocity, wave direction, storm, wind direction, wind speed, ocean currents and tides, rain, thunderstorms, precipitation, temperature, humidity, cloud cover, The method according to claim 1.

11. It has one or more processors, and these one or more processors are The process involves acquiring first real-time observation data corresponding to the current conditions within the vehicle's operating area, and the first real-time observation data is acquired based on the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the operating area of ​​the vehicle is generated based on the first real-time observation data, and the first prediction is adjusted for the operating area of ​​the vehicle according to the operating area. Based on the first prediction, one or more navigation routes for the vehicle in the operating area are generated, and the vehicle navigates the operating area according to the one or more navigation routes. A system configured to perform actions that include the following.

12. The above operation is performed before acquiring the first real-time observation data corresponding to the current status of the vehicle within the operating area, The further includes training the AI ​​prediction model to predict weather conditions using a dataset of past weather conditions, The system according to claim 11.

13. The aforementioned operation is, To acquire second real-time observation data corresponding to the current situation within one or more of the aforementioned navigation routes, Using the AI ​​prediction model, generate a second prediction of one or more weather conditions associated with one or more navigation routes based on the second real-time observation data, Based on the second prediction, generate one or more updated navigation routes for the vehicle in the operating area, and the vehicle navigates the operating area according to the one or more updated navigation routes. The system according to claim 11, further comprising:

14. The above operation is performed before acquiring the second real-time observation data. The further includes obtaining user input via a user interface indicating a selection to acquire the second real-time observation data, The system according to claim 13.

15. The user input indicates a specific time step for supplying the second real-time observation data to the AI ​​prediction model. The system according to claim 14.

16. The above operation is performed before generating the one or more navigation routes. This further includes downscaling the first set of predictions using an AI downscaling model. The system according to claim 11.

17. Downscaling the first set of predictions mentioned above is: The method involves training the AI ​​downscaling model using a first historical training dataset containing weather condition data with a first resolution and a second historical training dataset containing weather condition data with a second resolution, wherein the first resolution is lower than the second resolution. Obtaining the first set of predictions from the AI ​​prediction model at the first resolution, Using the AI ​​downscaling model, the first resolution of the first set of predictions is converted to the second resolution. including, The system according to claim 16.

18. The weather condition data having the first resolution is 20 km ２ Equivalent to or 20 km ２ Including data with grid sizes exceeding, The system according to claim 17.

19. The meteorological condition data having the second resolution is 1 km ２ Equivalent to or 1 km ２ Includes data with a grid size smaller than, The system according to claim 17.

20. One or more non-temporary computer-readable media storing instructions, When the aforementioned instruction is executed by one or more processors, it causes the system to perform an action. The aforementioned operation is, The process involves acquiring first real-time observation data corresponding to the current conditions within the vehicle's operating area, and the first real-time observation data is acquired based on the first real-time observation data corresponding to the vehicle's operating area. Using an artificial intelligence (AI) prediction model, a first prediction of one or more weather conditions associated with the operating area of ​​the vehicle is generated based on the first real-time observation data, and the first prediction is adjusted for the operating area of ​​the vehicle according to the operating area. Based on the first prediction, one or more navigation routes for the vehicle in the operating area are generated, and the vehicle navigates the operating area according to the one or more navigation routes. One or more non-temporary computer-readable media, including [the specified text].

Images