A heading and speed cooperative planning method based on A-star algorithm under fixed time

Abstract

The application provides a heading and speed cooperative planning method under fixed time based on an A-star algorithm, relates to the ship route design technical field, and comprises the following steps: extracting environmental data in a specified region under corresponding time, and performing grid processing on the sea area in the range; processing original AIS data, and performing feature extraction on a voyage section; constructing and training a prediction model of oil consumption rate and propeller rotating speed by using characteristic values of the voyage section; improving a conventional A-star algorithm process, adding a speed planning function, and outputting a time-limited minimum oil consumption route. The application can guarantee that a ship is effectively kept away from a strong wind and wave area while reducing oil consumption, and navigation decision is more flexible, and the application can adapt to complex and changeable ocean navigation tasks.

Description

Technical Field

[0001] This invention relates to the field of ship route design technology, and more particularly to a method for co-planning course and speed over a fixed time period based on the A* algorithm. Background Technology

[0002] In ship route design, the key evaluation indicators for route effectiveness are usually the ship's safety and economy. Therefore, a shipping route that meets both the safety and economy requirements of maritime transport will bring significant benefits to the shipping industry and is of great importance. By comprehensively considering the meteorological conditions of the sea areas that may be traversed during the voyage, especially factors such as wind, waves, and currents that can have a significant immediate impact on ship navigation, a reasonable route and speed design can effectively utilize favorable weather conditions and avoid unfavorable weather conditions. This can effectively improve the safety of ship transportation, reduce fuel consumption, and increase the ship's on-time arrival rate.

[0003] Currently, the industry primarily relies on meteorological navigation companies to design and send routes suitable for the vessel type before voyages, based on the departure and destination ports and weather forecasts for a certain period in the future. This method is highly dependent on the accuracy of weather forecasts; inaccurate forecasts can pose unpredictable risks, especially to vessels engaged in long-distance ocean voyages. Furthermore, long-term weather forecasts often have some error, resulting in inflexible navigation decisions and an inability to adjust voyage plans in real time. Additionally, meteorological navigation agencies cannot fully consider the vessel's characteristics and conditions when providing navigation services, lacking comprehensive understanding of the vessel's loading status, main engine characteristics, stability, and structure. Therefore, a route planning method with high computational speed and strong adaptability to the navigation environment, incorporating real-time weather data, is needed to enable vessels to plan their own routes. Meteorological and hydrological conditions, as well as the vessel's own characteristics, have a significant impact on fuel consumption and stall conditions, the latter two being key factors influencing route design. Therefore, accurately predicting fuel consumption and stall conditions is also a crucial part of the route planning process. Currently, mainstream path planning methods can be broadly categorized into three types: traditional graph algorithms, such as the isochronous method, Dijkstra's algorithm, and A* algorithm; swarm optimization algorithms, such as genetic algorithms, ant colony optimization, and particle swarm optimization; and artificial intelligence algorithms, which are data-driven and require extensive data computation and analysis to arrive at the desired path. All algorithms have their limitations; for example, the isochronous method has high computational complexity, swarm optimization algorithms are highly dependent on the initial solution for computational efficiency, and artificial intelligence algorithms have poor generalization ability when dealing with unknown situations.

[0004] The A* algorithm, also known as the A* star algorithm, is characterized by its fast computation speed, high performance, and high accuracy, making it suitable for real-world navigation decision-making conditions. However, a typical A* star algorithm, while providing route design, does not offer speed configuration options. This hinders a ship's ability to flexibly respond to changes in the marine environment and fails to meet the goals of maximizing economic efficiency and safety. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of the aforementioned background technology by proposing a course and speed collaborative planning method based on an improved A-Star algorithm. This method can update the route and speed configuration in real time, meet the requirements of fixed arrival time, take the lowest fuel consumption as the planning objective, has low computational complexity, and is easy to implement. Furthermore, it can accurately predict the ship's fuel consumption throughout the voyage based on weather forecast information and provide recommended propeller speed values ​​to assist navigation.

[0006] The technical means employed in this invention are as follows:

[0007] A course and speed cooperative planning method based on the A* algorithm under fixed time conditions includes the following steps:

[0008] Extract environmental data for the target sea area at the corresponding time and perform rasterization processing on the target sea area; divide the target sea area into obstructed navigation areas and navigable areas according to the ship's draft restrictions and the determination of navigation obstruction areas in the designated sea area;

[0009] The original AIS data is processed, and several uniform speed navigation segments of equal duration are extracted based on the processed AIS data. Feature extraction is performed on the uniform speed navigation segments to obtain the feature values ​​of the segments.

[0010] The feature values ​​of the flight segment are used as input and output data for the prediction model to construct and train the prediction model for fuel consumption rate and propeller speed.

[0011] The design incorporates an evaluation function related to fixed time constraints and voyage fuel consumption. This function uses a fuel consumption rate prediction model to calculate fuel consumption and adds a speed planning function to the A* algorithm. The improved A* algorithm is used to perform path planning within a gridded target sea area and outputs a time-constrained minimum fuel consumption route. Simultaneously, a propeller speed prediction model is used to output a recommended propeller speed value for the entire voyage.

[0012] Furthermore, the environmental data is meteorological and sea state reanalysis data or meteorological and sea state forecast data. The meteorological and sea state reanalysis data is used to train the prediction model, and the meteorological and sea state forecast data is used to design the route.

[0013] Furthermore, the navigation obstruction area includes dynamic navigation obstruction area and static navigation obstruction area; the dynamic navigation obstruction area includes: ice area, area with strong winds and waves that do not meet navigation requirements; the static navigation obstruction area includes land, islands, and areas that do not meet the safe water depth of the vessel itself.

[0014] Furthermore, the process of rasterizing the target sea area and dividing it into navigation-impeded and navigable areas includes the following steps:

[0015] Read environmental data, identify the environmental data, and extract wind, wave, and water depth related data;

[0016] By combining the relevant data with the ship's own strength and performance factors, thresholds for environmental parameters during ship navigation are set;

[0017] Divide the target sea area into grids of the same size according to latitude and longitude;

[0018] The navigation area is divided using a binary method. Based on the threshold values ​​of the environmental parameters, it is determined whether various types of environmental data meet the navigation conditions of the ship. If any type of environmental data in the grid does not meet the navigation requirements, the grid point is set as an unnavigable point with a value of 0. If all types of environmental data in the grid meet the navigation requirements, the grid point is set as a navigable point with a value of 1. The navigable grid is filled with white and the unnavigable grid is filled with black to obtain a binary map of the specified sea area, achieving the effect of gridding the sea area and dividing the sea area into obstructed and navigable areas.

[0019] Further processing of the raw AIS data includes the following steps:

[0020] Delete erroneous data, including data that is missing, has latitude and longitude out of range, or has ship speed or heading out of reasonable range;

[0021] Delete data indicating that the ship is at anchor;

[0022] Extract several sub-segments of uniform speed and duration from the remaining data and perform feature extraction on the uniform speed segments.

[0023] Furthermore, feature extraction for the uniform speed navigation segment includes:

[0024] Principal component analysis was used to select the main factors affecting ship fuel consumption and propeller speed. The mean of the main factors within the sub-segment was taken as the characteristic value of the sub-segment. Meteorological data at the location and time of the sub-segment were retrieved from the reanalysis meteorological data and used as the characteristic value of the sub-segment.

[0025] Furthermore, the prediction models for fuel consumption and propeller speed are constructed using a feedforward neural network method, and the construction steps include:

[0026] Construct a feedforward neural network model with a single hidden layer;

[0027] The main feature values ​​selected from meteorological data and ship routes are used as inputs to the prediction model, and fuel consumption rate or propeller speed is used as the output of the model. Ship data with high correlation to the output data is selected as the input of the model.

[0028] Furthermore, a node in a two-dimensional grid graph is defined as a navigation node, and a navigation node j adjacent to navigation node i in the two-dimensional grid graph is called a neighboring node of navigation node i; the ship's speed at navigation node i is v. w When it is, it is denoted as i w and called i w As an extension node of navigation node i, the ship's speed v during navigation is set within a range to be selected to adapt to different sea conditions and planning requirements.

[0029] Furthermore, the evaluation function of the A-Star algorithm is set as two sets of evaluation functions related to flight time and fuel consumption, as shown below:

[0030] FT iw =GT iw +HT iw

[0031] FF iw =GF iw +HF iw

[0032] Among them, HT iw For ships to move from the current extended node i w Estimated time to reach the destination, HF iw For ships to move from the current extended node i w Estimated fuel consumption upon reaching the destination; GT iw and GF iw For a ship to travel from its starting point to its current point i w The actual sailing time and fuel consumption, of which GF iw The calculation requires calling the trained fuel consumption prediction model.

[0033] Furthermore, the A-Star algorithm incorporates speed planning functionality and outputs a time-limited minimum fuel consumption route, including the following steps:

[0034] Step 1: Initialize Openlist and Closedlist to obtain the time limit for completing the planned voyage mission;

[0035] Step 2: Add all extended nodes belonging to the starting point S to the Openlist, set the G value of these extended nodes to 0, and calculate and record the H value of these nodes;

[0036] Step 3. Check whether the Openlist is an empty set; if the Openlist is an empty set, the problem has no solution, there is no path between the two points, and the operation is terminated; if it is not an empty set, proceed to Step 4;

[0037] Step 4. If the Openlist is not empty, first judge the total sailing time HT values of the extended nodes in the Openlist. If there is an extended node h w whose total sailing time is greater than the limit of the planned time, then the extended node is retained in the Openlist, but all subsequent operations in this step are not performed, that is, the extended node h is screened out in the subsequent operations w ; then find the extended node i with the minimum total sailing fuel consumption among the extended nodes whose total sailing time is within the planned time w , delete the extended node i w from the Openlist, and add it to the Closedlist, and configure the speed record as w;

[0038] Step 5. Traverse all adjacent nodes of the sailing node i. If any extended node of the adjacent node j does not exist in the Openlist and the Closedlist, then add all extended nodes of point j to the Openlist, calculate and record the G and H values of these extended nodes, and set the parent node of these extended nodes to node i w ; if there is an extended node j w subordinate to point j in the Openlist, then compare the GT1 and GF1 values recorded for node j w in the Openlist with the GT2 and GF2 values of node j w in the current operation process: if GT2 < GT1 and GF2 < GF1, then update the F, G, and H values of point j w to the F, G, and H values in the current operation process, and its parent node becomes the current node i w ; conversely, if GT1 and GT2, GF1 and GF2 are not in the above relationship, then all the information stored in node j w remains unchanged;

[0039] Step 6. Check whether there is any extended node subordinate to the end point T in the Openlist. If not, repeat Step 3; if it already exists, trace back the parent node of the current node step by step from this extended node until the starting point is found, and the result path and speed configuration are obtained after mapping, and the algorithm ends;

[0040] Step 7: Using the planned route and speed configuration, deduce the weather forecast data and navigation data of the ship at each extended node on the route; call the trained propeller speed prediction model, use the above data as the input of the prediction model, and calculate the recommended propeller speed value of the ship at each extended node on the route, and form a recommended propeller speed scheme for the whole voyage in sequence, and the algorithm ends.

[0041] Compared with the prior art, the present invention has the following advantages:

[0042] 1. The target sea area rasterization method provided by the present invention uses environmental data at the corresponding time in the target sea area to rasterize the target sea area, and according to the characteristics of the ship and the restrictions of the specified sea area, it realizes the goal of dividing the target sea area into the obstruction area and the navigable area, which can improve the running efficiency of the algorithm and enable the output route to respond to the navigation restriction area in a timely manner.

[0043] 2. The method for constructing a prediction model for fuel consumption rate and propeller speed provided by this invention applies a single-hidden-layer feedforward neural network for system modeling and uses real ship navigation data for model training, making the model highly adaptable, robust, and fault-tolerant, and achieving accurate prediction of fuel consumption rate and propeller speed parameters, providing reliable prediction data support for the subsequent path planning algorithm.

[0044] 3. The fixed-time course and speed collaborative planning method provided by the present invention improves the planning process of the A-Star algorithm by designing an A-Star algorithm evaluation function related to fixed time constraints and navigation fuel consumption. It realizes the timely and targeted optimization of ship course and speed based on the actual characteristics of the ship and weather conditions, and outputs the lowest fuel consumption route under fixed time.

[0045] 4. The propeller speed configuration scheme for the entire flight path provided by the present invention is composed of the predicted propeller speed values ​​on the final flight path obtained by the propeller speed prediction model. This configuration scheme can better fit the actual operation of navigation and provide more technical support for relevant personnel of navigation missions.

[0046] Based on the above reasons, this invention can be widely applied in fields such as ship route design. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart of the method of the present invention.

[0049] Figure 2 This is a gridded binary map of the marine environment according to the present invention.

[0050] Figure 3 This is a prediction result diagram of the propeller speed prediction model of the present invention.

[0051] Figure 4 This is a diagram showing the prediction results of the fuel consumption rate prediction model of this invention.

[0052] Figure 5 This is a schematic diagram of the nodes after incorporating the speed factor in this invention.

[0053] Figure 6 This is the flight path diagram generated after projection according to the present invention.

[0054] Figure 7 This is the speed configuration diagram generated after projection according to the present invention.

[0055] Figure 8 This is a flowchart of the improved algorithm of the present invention.

[0056] Figure 9 This is a diagram showing the route planning results of this invention.

[0057] Figure 10 This is a diagram showing the speed configuration results of the present invention.

[0058] Figure 11 This is a graph showing the predicted fuel consumption for the entire flight according to the present invention.

[0059] Figure 12 This is a diagram showing the recommended propeller speed for the entire flight range according to the present invention. Detailed Implementation

[0060] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0061] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0062] like Figure 1 As shown in the figure, this invention discloses a heading and speed cooperative optimization algorithm based on the A-Star algorithm, which includes the following steps:

[0063] S1. Extract environmental data for the corresponding time period within the designated area and perform rasterization processing on the sea area. Based on the ship's draft restrictions and the determination of navigational obstruction zones within the designated sea area, divide the sea area into navigational obstruction zones and navigable zones;

[0064] S2. Process the original AIS data and extract several uniform speed navigation segments of equal duration based on the ship's AIS data, and then extract the features of the segments.

[0065] S3. Use the characteristic values ​​of the flight segment and wind and wave data as the input and output data of the prediction model to build and train the prediction model of fuel consumption rate and propeller speed.

[0066] S4. Improve the evaluation function and process of the conventional A-star algorithm by adding the function of speed planning and outputting the lowest fuel consumption route with time limit.

[0067] The following is a further explanation of each of the above steps.

[0068] In S1, it is necessary to first extract environmental data for the corresponding time within the specified area. Environmental data for the corresponding time within the specified area refers to meteorological and sea state reanalysis data or meteorological and sea state forecast data. This method obtains data on the above factors from the meteorological and sea state reanalysis database to train the prediction model. During navigation planning, it obtains the above types of data from the meteorological and sea state forecast database to design the route. Most current meteorological and sea state reanalysis data and forecast data are in gridded (GRIB) format. This type of database divides a certain sea area into grids of the same size according to latitude and longitude to store the corresponding meteorological and sea state data, and continuously records meteorological and sea state data at fixed time intervals within a specified time range.

[0069] During navigation, it is necessary to pay attention to areas that may obstruct navigation safety. These areas include dynamic and static obstruction zones. Dynamic obstruction zones include: ice zones and areas with strong winds and waves that do not meet navigation requirements; static obstruction zones include: land, islands, and areas that do not meet the vessel's safe water depth requirements, etc.

[0070] After obtaining the relevant environmental data, the sea area needs to be rasterized and divided into navigation-impeded and navigable areas. In a specific embodiment, the steps of rasterizing the sea area and dividing it into navigation-impeded and navigable areas include:

[0071] Read environmental data, identify and extract data related to wind, waves, and water depth;

[0072] Based on the ship's own strength and performance factors, threshold values ​​for environmental parameters during ship navigation are set.

[0073] The sea area within the specified range is divided into grids of the same size according to latitude and longitude, and the precision is the same as that of the environmental data. For example, if the precision of the environmental data is 0.5°×0.5°, then the sea area will also be divided into grids of 0.5°×0.5°.

[0074] The navigation area is divided using a binary method, and the navigation conditions of various environmental data are determined based on threshold values ​​of environmental parameters. If any type of environmental data in a grid does not meet the navigation requirements, the grid point is set as an unnavigable point with a value of 0. Conversely, if all types of environmental data in a grid meet the navigation requirements, the grid point is set as a navigable point with a value of 1.

[0075] Filling navigable grid cells with white and non-navigable grid cells with black creates a binary map of the designated sea area, achieving a rasterized sea area effect and dividing it into navigable and obstructed zones. Figure 2 As shown.

[0076] In S2, the raw AIS data of a specific vessel is obtained. The data includes at least the sailing time, vessel position, speed, fuel consumption, and propeller speed.

[0077] In S2, the raw AIS data of the ship is processed, specifically including:

[0078] S21. Delete erroneous data, such as data that is missing, latitude and longitude that are out of range, or ship speed and heading that are out of reasonable range.

[0079] S22. Delete data indicating that the ship is at anchor;

[0080] S23. Following the guidance of relevant ISO documents, extract several sub-segments of equal duration from all data to obtain the characteristic values ​​of uniform speed segments. Due to the large amount of data and the long-term forecasting nature of route planning, the average value of various data for each sub-segment will be extracted to represent the ship's navigation status under that segment. Therefore, sub-segments of continuous uniform speed navigation should be extracted. Then, according to the guidance of ISO documents and the rules for judging abnormal data, uniform speed navigation sub-segments containing abnormal data are removed, resulting in several uniform speed navigation sub-segments of equal duration.

[0081] In S23, feature extraction of a voyage segment is performed by principal component analysis to select the main factors affecting ship fuel consumption and propeller speed. Then, the mean of these main factors is taken as the feature value of the voyage segment within the sub-segment. At the same time, the mean of the meteorological data at the location and time of the sub-segment is also retrieved and used as the feature value of the sub-segment.

[0082] In S23, the ISO guidance document refers to part 2 of ISO 19030-2, which defines the default methods for measuring changes in hull and propeller performance and calculating a set of basic performance indicators, including guidelines for cleaning ship navigation data.

[0083] In one specific embodiment, S3 includes:

[0084] S31. Design the prediction model as a feedforward neural network architecture. Since the prediction model needs to be continuously called during flight route planning, it requires a relatively simple model with high prediction accuracy. A single-hidden-layer feedforward neural network model can meet these requirements. A feedforward neural network generally consists of three layers: an input layer, hidden layers, and an output layer. Each layer can contain several neurons. Any neuron in the network is connected to neurons in any adjacent layer, but has no connection to neurons in the same layer or across layers.

[0085] S32. When building the model, wind and wave data of the environment in which the ship is located, as well as key feature values ​​selected from the ship's route, are used as inputs to the prediction model. Fuel consumption rate or propeller speed is used as the model output. In the construction and training of the neural network, ship data with high correlation to the output data should be selected as inputs, and appropriate activation functions, loss functions, and optimizers should be used to obtain better prediction results. For example, using ship-related data such as wind direction angle, wind speed, wave direction angle, wave height, ship draft on both port and starboard sides, and ground speed as inputs, ReLU is used as the activation function of the hidden layer, mean squared error (MSE) is used as the loss function, and the Levenberg-Marquardt algorithm is used as the optimizer. Prediction models for fuel consumption rate and propeller speed are thus established, and the prediction results are as follows: Figure 3 , 4 As shown.

[0086] In S4, for ease of process description, nodes that can be represented in a two-dimensional grid diagram are defined as navigation nodes, and navigation nodes j adjacent to navigation node i in the two-dimensional grid diagram are called neighboring nodes of navigation node i; the ship's speed at navigation node i is v. w When it is, it is denoted as i w and called i w It is an extension node of navigation node i, and the two navigation nodes i and j and the route between them constitute a segment of the ship.

[0087] In S4, the ship's speed during navigation will be maintained within a range based on factors such as its own performance, navigation environment, and economic efficiency. Therefore, the actual speed v can be set within a range to adapt to different sea conditions and operational needs, namely:

[0088] v = [v min ,v min +v cd ,v min +2v cd ,…,v max ]

[0089] As shown in the equation, v is the minimum speed v min Maximum speed v max For and its interval is v cd The constructed row matrix. Correspondingly, since there are multiple choices for the actual speed, each navigation node will also be expanded into a set P, where each element represents the navigation situation at different speeds under the same navigation node, as shown in the equation:

[0090] i w ={P(i,v w )|v w ∈v}

[0091] In the formula, i w P is an extended node representing the actual speed of the ship at the navigation node; P contains all possible actual speeds of the ship at the navigation node. By adding a speed variable and configuring the speed, the expression of the optimized route from the starting point S to the ending point T evolves from the original SABCDET to S 13 -A 12 -B 13 -C 14 -D 12 -E 15 -T, such as Figure 5 As shown, the ship travels from the starting point S at a speed of 13 knots to point A, from point A at a speed of 12 knots to point B, and so on, until it reaches the destination T at a speed of 15 knots at point E. Using this method, the calculated optimized route includes not only optimized waypoints (such as...) Figure 6As shown), and also includes the optimal speed on each segment (e.g. Figure 7 As shown in the figure, this enables the coordinated optimization of ship course and speed.

[0092] In S4, the evaluation function in the A* algorithm is improved. The main form of the A* algorithm evaluation function is as follows:

[0093] F = G + H

[0094] Where G represents the sum of actual travel costs from the starting point to the current point, and H represents the estimated travel cost from the current point to the destination. To meet planning requirements, this is defined as two sets of evaluation functions related to travel time and fuel consumption, namely...

[0095] FT iw =GT iw +HT iw

[0096] FF iw =GF iw +HF iw

[0097] Among them, HT iw For ships to move from the current extended node i w Estimated time to reach the destination, HF iw For ships to move from the current extended node i w Estimated fuel consumption upon reaching the destination. GT iw and GF iw For a ship to travel from its starting point to its current point i w The actual flight time and fuel consumption are shown in the following formula:

[0098]

[0099]

[0100] Assume h w i w For two adjacent extended nodes, h in the flight path w in i w Previously, then t iw Point h w sail to point i w The sailing time, q iw Point h w sail to point i w Fuel consumption. q iw The calculation requires calling the fuel consumption rate prediction model in S3, using ship data and weather forecast data during the voyage as model inputs to calculate the fuel consumption rate from h w Click to i w Fuel consumption q between iw .

[0101] In S4, the A* algorithm process is then improved accordingly, and the process is as Figure 8 shown, including:

[0102] After constructing a suitable evaluation function, the process of the A* algorithm is improved accordingly, and the process is as follows:

[0103] S41. Initialize Openlist (open set, set of nodes to be checked) and Closedlist (closed set, set of nodes that have been checked), and obtain the time limit for completing the navigation task;

[0104] S42. Add all the expanded nodes belonging to the starting point S to Openlist, set the G value of these expanded nodes to 0, and calculate and record the H value of these nodes;

[0105] S43. Check whether Openlist is an empty set. If Openlist is an empty set, the problem has no solution, there is no path between the two points, and the operation is terminated; if it is not an empty set, then proceed to step S44.

[0106] S44. If Openlist is not empty, first judge the total navigation time F t value of each expanded node in Openlist. If there exists an expanded node h w whose total navigation time FT is greater than the planned time limit, then continue to retain this expanded node in Openlist, but do not perform the following operations in this step, that is, screen out this expanded node h w ; then find the expanded node i w with the minimum total navigation fuel consumption among the expanded nodes whose total navigation time is within the planned time, delete this expanded node i w from Openlist, and add it to Closedlist, and configure the recorded speed as w;

[0107] S45. Traverse all the neighboring nodes of the navigation node i. If any expanded node of the neighboring node j does not exist in Openlist and Closedlist, then add all the expanded nodes of point j to Openlist, calculate and record the G and H values of these expanded nodes, and set the parent node of these expanded nodes to node i w ; if there exists an expanded node j w belonging to point j in the neighboring nodes in Openlist, then compare the GT1 and GF1 values recorded for node j w in Openlist with the GT2 and GF2 values of node j w in the current operation process: If GT2 < GT1 and GF2 < GF1, then update jw The F, G, and H values ​​of a point are the F, G, and H values ​​in the current calculation process, and its parent node becomes the current node i. w Conversely, if GT1 and GT2, GF1 and GF2 do not have the above relationships, then node j w All stored information remains unchanged;

[0108] S46. Check if there is any extended node in Openlist that belongs to the endpoint T. If not, repeat step S43. If it already exists, start from the extended node and trace the parent node of the current node step by step until the starting point is found. After mapping, the result path and speed configuration are obtained, and the algorithm ends.

[0109] S47. Using the planned route and speed configuration, deduce the weather forecast data and navigation data (e.g., speed, heading, draft, etc.) for each extended node on the route; call the propeller speed prediction model trained in S3, using the above data as input to the prediction model, thereby calculating the recommended propeller speed values ​​for each extended node on the route, and sequentially assembling the recommended propeller speed scheme for the entire voyage, ending the algorithm. For example, simulating a ship departing from Tainan City, Taiwan Province, China, at 8:00 AM on January 17, 2021, bound for Los Angeles City, USA, with a planned arrival time before 8:00 AM on February 3, 2021, the planned route result is as follows: Figure 9 As shown, the speed configuration scheme is as follows: Figure 10 As shown, the propeller speed value is as follows Figure 11 As shown, the fuel consumption during navigation is as follows: Figure 12 As shown, Figure 10 , 11 The horizontal coordinates of 1 and 2 both refer to segments of the route.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A course and speed cooperative planning method based on the A* algorithm over a fixed time period, characterized in that, Includes the following steps: Extract environmental data for the target sea area at the corresponding time and perform rasterization processing on the target sea area; divide the target sea area into obstructed navigation areas and navigable areas according to the ship's draft restrictions and the determination of navigation obstruction areas in the designated sea area; The original AIS data is processed, and several uniform speed navigation segments of equal duration are extracted based on the processed AIS data. Feature extraction is performed on the uniform speed navigation segments to obtain the feature values ​​of the segments. The feature values ​​of the flight segment are used as input and output data for the prediction model to construct and train the prediction model for fuel consumption rate and propeller speed. The design incorporates an evaluation function related to fixed time constraints and voyage fuel consumption. This function uses a fuel consumption rate prediction model to calculate fuel consumption and adds a speed planning function to the A* algorithm. The improved A* algorithm is used to perform path planning within a gridded target sea area and outputs a time-constrained minimum fuel consumption route. Simultaneously, a propeller speed prediction model is used to output a recommended propeller speed value for the entire voyage.

2. The heading and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 1, characterized in that, The environmental data is meteorological and sea state reanalysis data or meteorological and sea state forecast data. The meteorological and sea state reanalysis data is used to train the prediction model, and the meteorological and sea state forecast data is used to design the route.

3. The heading and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 1, characterized in that, The navigation obstruction areas include dynamic navigation obstruction areas and static navigation obstruction areas; the dynamic navigation obstruction areas include: ice areas and areas with strong winds and waves that do not meet navigation requirements; the static navigation obstruction areas include land, islands, and areas that do not meet the safe water depth of the vessel itself.

4. The heading and speed cooperative planning method based on the A* algorithm under fixed time conditions according to claim 1, characterized in that, The process of rasterizing the target sea area and dividing it into obstructed and navigable zones includes the following steps: Read environmental data, identify the environmental data, and extract wind, wave, and water depth related data; By combining the relevant data with the ship's own strength and performance factors, thresholds for environmental parameters during ship navigation are set; Divide the target sea area into grids of the same size according to latitude and longitude; The navigation area is divided using a binary method. Based on the threshold values ​​of the environmental parameters, it is determined whether various types of environmental data meet the navigation conditions of the ship. If any type of environmental data in the grid does not meet the navigation requirements, the grid is set as an unnavigable point with a value of 0. If all types of environmental data in the grid meet the navigation requirements, the grid is set as a navigable point with a value of 1. Navigable grids are filled with white and unnavigable grids are filled with black to obtain a binary map of the specified sea area, achieving the effect of gridding the sea area and dividing the sea area into obstructed and navigable areas.

5. The heading and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 1, characterized in that, Processing raw AIS data includes the following steps: Delete erroneous data, including data that is missing, has latitude and longitude out of range, or has ship speed or heading out of reasonable range; Delete data indicating that the ship is at anchor; Extract several sub-segments of uniform speed and duration from the remaining data and perform feature extraction on the uniform speed segments.

6. The course and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 5, characterized in that, Feature extraction for the uniform speed navigation segment includes: Principal component analysis was used to select the main factors affecting ship fuel consumption and propeller speed. The mean of the main factors within the sub-segment was taken as the characteristic value of the sub-segment. Meteorological data at the location and time of the sub-segment were retrieved from the reanalysis meteorological data and used as the characteristic value of the sub-segment.

7. The course and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 1, characterized in that, The prediction models for fuel consumption and propeller speed are constructed using a feedforward neural network method, and the construction steps include: Construct a feedforward neural network model with a single hidden layer; The main feature values ​​selected from meteorological data and ship routes are used as inputs to the prediction model, and fuel consumption rate or propeller speed is used as the output of the model. Ship data with high correlation to the output data is selected as the input of the model.

8. The heading and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 1, characterized in that, Nodes defined in a two-dimensional raster graph are called navigation nodes. The relationship between navigation nodes and navigation nodes in a two-dimensional raster graph is... i Adjacent navigation nodes j Called navigation node i The nearest node; the ship at the navigation node i The speed of the ship is v w Time is recorded as i w and said i w For navigation nodes i The extended node, the ship's speed during navigation. v The selection is set within a range to accommodate different sea conditions and planning requirements.

9. The heading and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 1, characterized in that, The evaluation function of the A-Star algorithm is set as two sets of evaluation functions related to flight time and fuel consumption, as shown below: FT iw = GT iw + HT iw FF iw = GF iw + HF iw in, HT iw For ships from the current expansion node i w Estimated time to reach the destination HF iw For ships from the current expansion node i w Estimated fuel consumption upon reaching the destination; GT iw and GF iw For ships to travel from the starting point to the current point i w The actual sailing time and fuel consumption, of which GF iw The calculation requires calling the trained fuel consumption prediction model.

10. The heading and speed cooperative planning method based on the A* algorithm under fixed time as described in claim 8, characterized in that, The A-Star algorithm incorporates speed planning functionality and outputs a time-limited minimum fuel consumption route, including the following steps: Step 1: Initialize Openlist and Closedlist to obtain the time limit for completing the planned voyage mission; Step 2: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] S Add all extended nodes to the Openlist and set the permissions for these extended nodes. G If the value is 0, calculate and record the values ​​of these nodes. H value; Step 3: Check if Openlist is an empty set; if Openlist is an empty set, there is no solution to the problem, there is no path between the two points, and the calculation terminates; if it is not an empty set, proceed to Step 4. Step 4: If the Openlist is not empty, first calculate the total travel time for each extended node in the Openlist. HT The value is used to determine if an extended node exists. h w If the total sailing time exceeds the planned time limit, the extended node will remain in the Openlist, but all subsequent operations in this step will not be performed; that is, the extended node will be removed in the following operations. h w Then, find the extended node with the minimum total fuel consumption among the extended nodes whose total travel time is within the planned time. i w Extend the node i w Remove it from the Openlist and add it to the Closedlist, then configure the speed record as follows. w ; Step 5: Traverse the navigation nodes i All neighboring nodes; if neighboring nodes j If none of the extended nodes exist in the Openlist and Closedlist, then the node will be... j All extended nodes are added to the Openlist, and the values ​​of these extended nodes are calculated and recorded. G , H The value is set, and the parent node of these extended nodes is set to the node. i w If there is a member point among the neighboring nodes. j Extended nodes j w In Openlist, nodes are compared. j w Recorded in Openlist GT 1. GF 1 value and node in the current operation j w of GT 2. GF 2 values: if GT 2< GT 1 and GF 2< GF 1, then update j w Pointed F , G , H The value is in the current operation process F , G , H The value, its parent node becomes the current node. i w Conversely, if GT 1 and GT 2. GF 1 and GF If 2 is not one of the relationships described above, then the node... j w All stored information remains unchanged; Step 6: Check if there are any dependent endpoints in the Openlist. T If any of the extended nodes does not exist, repeat step three; if it already exists, start from that extended node and gradually trace the parent node of the current node until the starting point is found. After mapping, the resulting path and speed configuration are obtained, and the algorithm ends. Step 7: Using the planned route and speed configuration, deduce the weather forecast data and navigation data of the ship at each extended node on the route; call the trained propeller speed prediction model, use the above data as the input of the prediction model, and calculate the recommended propeller speed value of the ship at each extended node on the route, and form a recommended propeller speed scheme for the whole voyage in sequence, and the algorithm ends.

Images