Skiing control method and device, computer program product and electronic equipment

Abstract

The invention relates to the technical field of data processing, and relates to a skiing control method and device, a computer program product and electronic equipment. The method comprises the steps that multi-modal information of a skier is collected, wherein the multi-modal information comprises main body information of the skier and environment interaction information of the skier in the skiing process; carrying out object tracking on the skier according to the multi-modal information and predicting a skiing trajectory to obtain a first predicted skiing trajectory of the target skier and a second predicted skiing trajectory of skiers around the target skier; and performing analysis and calculation based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multi-modal information to obtain a skiing control strategy, and performing collision early warning and / or path recommendation on the target skier according to the skiing control strategy. According to the invention, a skiing control strategy can be provided for the skier according to the multi-modal information of the skier, and the risk of skiing collision is reduced.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and more particularly, to a ski control method, a ski control device, a computer program product, and an electronic device. Background Art

[0002] While skiing is widely popular for its thrill and fun, it also carries a high risk of injury, particularly from collisions between skiers. Current ski safety measures, such as protective gear and safety education, have helped, but collisions are still difficult to prevent.

[0003] Related technologies mark the skier's position on a map for reference during skiing, but rely on the skier to actively maintain distance and observe the surrounding environment. When skiing at high speed or with obstructed vision, it is difficult for the skier to react in time to avoid collision, which affects the skier's safety to a certain extent.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0005] The purpose of the present disclosure is to provide a ski control method and device, a computer program product and an electronic device, which can at least provide a ski control strategy for a skier based on the skier's multimodal information to reduce the risk of ski collision.

[0006] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by practice of the present disclosure.

[0007] According to one aspect of the present disclosure, a skiing control method is provided, comprising: collecting multimodal information of a skier, the multimodal information including the skier's main body information and the skier's environmental interaction information during skiing; tracking the skier and predicting the skiing trajectory based on the multimodal information to obtain a first predicted skiing trajectory of a target skier and a second predicted skiing trajectory of skiers surrounding the target skier; performing analysis and calculation based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a skiing control strategy, and performing collision warning and / or path recommendation for the target skier based on the skiing control strategy.

[0008] In an exemplary embodiment of the present disclosure, analysis and calculation are performed based on the first predicted ski trajectory, the second predicted ski trajectory and multimodal information to obtain a ski control strategy, and collision warning and / or path recommendation are performed for the target skier according to the ski control strategy, including: if it is determined that the target skier meets the predetermined collision condition according to the first predicted ski trajectory, the second predicted ski trajectory and multimodal information, collision warning information is generated and provided to the target skier; and / or, path generation is performed according to the first predicted ski trajectory, the second predicted ski trajectory and multimodal information, and a recommended ski path is obtained and provided to the target skier.

[0009] In an exemplary embodiment of the present disclosure, the subject information includes the skier's biometric data, equipment status data and identity attribute data; the environmental interaction information includes the skier's motion status data, environmental perception data and behavioral decision data; the skier is tracked and the skiing trajectory is predicted based on the multimodal information, including: generating the skier's motion endurance evaluation coefficient based on the biometric data, equipment status data and identity attribute data; determining the trajectory search area based on the environmental perception data, behavioral decision data and motion endurance evaluation coefficient; predicting the trajectory position in the trajectory search area based on the running status data, and determining the skier's predicted skiing trajectory based on the trajectory position.

[0010] In an exemplary embodiment of the present disclosure, the environmental interaction information includes skiing weather information; collecting multimodal information of skiers also includes: real-time monitoring of the confidence level of the skiing weather information; if the confidence level of the skiing weather information is lower than a preset confidence level, correcting the skiing weather data.

[0011] In an exemplary embodiment of the present disclosure, if the confidence level of the skiing weather information is lower than a preset confidence level, the skiing weather data is corrected, including at least one of the following: generating alternative data for the skiing weather data based on the historical weather data of the current ski slope during the same period; obtaining weather data of adjacent areas of the current ski slope, and performing spatial interpolation based on the weather data to obtain alternative data for the skiing weather data; sending an alarm signal to the ski resort management end, and receiving the alternative data sent by the ski resort management end.

[0012] In an exemplary embodiment of the present disclosure, determining that the confidence level of ski weather information is lower than a preset confidence level includes: comparing the ski weather information with weather station data and data differences from a third-party weather application; if the data difference exceeds a preset difference threshold, determining that the confidence level of the ski weather information is lower than the preset confidence level.

[0013] In an exemplary embodiment of the present disclosure, if it is determined based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information that the target skier meets the predetermined collision condition, collision warning information is generated and provided to the target skier, including: if the first predicted skiing trajectory and the second predicted skiing trajectory meet the first collision condition, and / or the target multimodal information indicates that the distance between the target skier and the obstacle object meets the second collision condition, collision warning information is generated, and the obstacle object includes the surrounding skiers and surrounding obstacles of the target skier.

[0014] In an exemplary embodiment of the present disclosure, a path is generated based on a first predicted skiing trajectory, a second predicted skiing trajectory and multimodal information to obtain a recommended skiing path and provide it to a target skier, including: determining a target skiing power of the target skier based on the multimodal information and a first relationship model, the first relationship model being used to characterize the correspondence between the multimodal information and the skiing power; determining a target skiing turning radius based on the target skiing power, the multimodal information and the second relationship model, the second relationship model being used to characterize the correspondence between the skiing power and the turning radius under different skiing methods; using the second predicted skiing trajectory as a roadblock, constructing a search space based on the snow trail, and searching for path points in the search space based on the target skiing turning radius to determine a recommended skiing path based on the obtained target path points.

[0015] In an exemplary embodiment of the present disclosure, a second predicted ski trajectory is used as a roadblock, a search space is constructed according to a snow track, and a path point search is performed in the search space based on a target ski turning radius to generate a recommended ski path based on the obtained target path points, including: constructing a search space according to the snow track, wherein each grid node in the search space includes a state vector for characterizing the facing angle and ski turning radius of the skier at the grid node, and marking the second predicted ski trajectory as a roadblock node in the search space; determining the grid node corresponding to the current position of the target skier as the initial node, and searching for path points in the search space based on the state vectors of each grid node under the constraints of the target ski turning radius and the snow track radius; if the end position of the target skier is found to be a path point, then determining the target path point based on each path point between the end position and the initial node to generate a recommended ski path based on the obtained target path point.

[0016] In an exemplary embodiment of the present disclosure, path generation is performed based on a first predicted skiing trajectory, a second predicted skiing trajectory and multimodal information to obtain a recommended skiing path and provide it to a target skier, including: obtaining a pre-trained path generation model; generating a recommended skiing path based on the first predicted skiing trajectory, the second predicted skiing trajectory and multimodal information through the pre-trained path generation model.

[0017] In an exemplary embodiment of the present disclosure, multimodal information of a target skier is collected, including: obtaining the skier's registration information; obtaining sensor information from the skier's wearable device; performing feature extraction on the collected image of the skier to obtain image feature information; and determining multimodal information based on the registration information, sensor information, and image feature information.

[0018] According to one aspect of the present disclosure, a ski control device is provided, comprising: an information acquisition module for acquiring multimodal information of a skier, the multimodal information including the skier's main body information and the skier's environmental interaction information during skiing; a trajectory prediction module for tracking the skier and predicting the ski trajectory based on the multimodal information, thereby obtaining a first predicted ski trajectory of a target skier and a second predicted ski trajectory of skiers surrounding the target skier; a control scheduling module for performing analysis and calculation based on the first predicted ski trajectory, the second predicted ski trajectory and the multimodal information to obtain a ski control strategy, and performing collision warning and / or path recommendation for the target skier according to the ski control strategy.

[0019] According to one aspect of the present disclosure, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, any one of the above methods is implemented.

[0020] According to one aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform any one of the above methods by executing the executable instructions.

[0021] The skiing control method in the exemplary embodiment of the present disclosure, on the one hand, can fully perceive the skiing situation by simultaneously collecting the skier's main information and environmental interaction information, thus solving the problem of misjudgment caused by traditional single-dimensional monitoring and improving the accuracy of identifying potential risks. On the other hand, the dual prediction mechanism (the first predicted trajectory is used for the skier's own path prediction, and the second predicted trajectory is used for behavior analysis of surrounding skiers) can realize collision analysis in the future, and can detect dangers earlier than existing reactive warning systems, thereby guiding active obstacle avoidance during skiing and reducing the risk of skiing collisions. Furthermore, by coupling and analyzing multimodal information with predicted trajectories, the generated skiing control strategy includes collision warning and / or path recommendation. This not only provides early warning of collisions, but also provides skiers with better obstacle avoidance routes or braking suggestions through path recommendations, which helps improve the skier's decision-making efficiency in complex skiing processes.

[0022] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and other objects, features and advantages of the exemplary embodiments of the present disclosure will become readily understood by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present disclosure are shown by way of example and not limitation.

[0024] Figure 1 An application environment according to an exemplary embodiment of the present disclosure is shown.

[0025] Figure 2 A flowchart of a ski control method according to an exemplary embodiment of the present disclosure is shown.

[0026] Figure 3 A flowchart illustrating an implementation method for generating a recommended ski route according to an exemplary embodiment of the present disclosure is shown.

[0027] Figure 4 A flowchart of generating a recommended ski route according to an exemplary embodiment of the present disclosure is shown.

[0028] Figure 5 A flowchart illustrating an implementation method for predicting a skiing trajectory according to an exemplary embodiment of the present disclosure is shown.

[0029] Figure 6 A flowchart of collecting multimodal information of a skier according to an exemplary embodiment of the present disclosure is shown.

[0030] Figure 7 A schematic diagram showing the composition of a ski control device according to an exemplary embodiment of the present disclosure is shown.

[0031] Figure 8 A block diagram of an electronic device according to an exemplary embodiment of the present disclosure is shown.

[0032] In the drawings, the same or corresponding reference numerals denote the same or corresponding parts. DETAILED DESCRIPTION

[0033] The exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these examples are provided so that this disclosure will be thorough and complete and will fully convey the concepts of the exemplary embodiments to those skilled in the art. Identical reference numerals in the figures represent identical or similar structures, and thus detailed descriptions thereof will be omitted.

[0034] In addition, the described features, structures or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, well-known structures, methods, devices, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0035] The blocks shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. Specifically, these functional entities may be implemented in software, or in one or more software-hardened modules, or in different networks and / or processor devices and / or microcontroller devices.

[0036] While skiing is widely popular for its thrill and fun, it also carries a high risk of injury, particularly from collisions between skiers. Current ski safety measures, such as protective gear and safety education, have helped, but collisions are still difficult to prevent.

[0037] Current ski control methods mark the positions of skiers in front of or behind the skier on a map for the skier's reference during skiing. However, these methods rely on the skier to actively maintain distance and observe the surrounding environment. When skiing at high speeds or with obstructed vision, it is difficult for the skier to react in time to avoid a collision, which affects the skier's safety to a certain extent.

[0038] Based on this, an exemplary embodiment of the present disclosure provides a skiing control method, which can provide a skier with a skiing control strategy based on the skier's multimodal information, thereby reducing the risk of skiing collisions.

[0039] It should be noted that the gaze point estimation method of the exemplary embodiment of the present disclosure can be applied to skiing scenes, competitive training grounds, ski resort management scenes, etc. The exemplary embodiment of the present disclosure does not impose specific restrictions on the application scenarios.

[0040] The ski control method provided by the exemplary embodiments of the present disclosure can be applied to Figure 1 In the application environment shown, the terminal 101 communicates with the server 102 via a network. The data storage system can store data that the server 102 needs to process. The data storage system can be integrated on the server 102 or placed on the cloud or other network servers.

[0041] In one exemplary embodiment, the ski control method provided by the exemplary embodiments of the present disclosure can be executed by server 102, with the corresponding ski control device being disposed in server 102. Accordingly, in this mode of execution by server 102, server 102 can begin executing the steps of the technical solution of the exemplary embodiments of the present disclosure in response to a trigger command, wherein the trigger command can be sent by a terminal used by a user or can be triggered locally by the server in response to some automated event.

[0042] The server 102 may be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server 102 may perform background tasks.

[0043] Furthermore, in another exemplary embodiment, the terminal 101 may also have similar functions to the server 102 , thereby executing the ski control method provided by the exemplary embodiment of the present disclosure.

[0044] The terminal 101 may be a ski device, such as ski goggles, a ski helmet, etc. A ski control device may be integrated into the terminal 101 , and the exemplary embodiments of the present disclosure do not limit the type of the terminal 101 .

[0045] In addition, the technical solutions of the exemplary embodiments of the present disclosure can also be collaboratively executed by the terminal 101 and the server 102. In this collaborative execution mode by the terminal 101 and the server 102, some steps of the technical solutions provided by the exemplary embodiments of the present disclosure are executed by the terminal 101, while other steps are executed by the server 102. For example, the terminal 101 collects multimodal information of a skier and sends the multimodal information to the server 102. The server 102 then tracks the skier and predicts the skiing trajectory based on the multimodal information, obtaining a first predicted skiing trajectory of the target skier and a second predicted skiing trajectory of the target skier's surrounding skiers. The server 102 then analyzes and calculates the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information to obtain a skiing control strategy. The server then performs collision warnings and / or route recommendations for the target skier based on the skiing control strategy. It should be noted that in this collaborative execution mode by the terminal 101 and the server 102, the steps respectively executed by the terminal 101 and the server 102 can be dynamically adjusted according to actual conditions, and there are no special restrictions on this.

[0046] The terminal 101 and the server 102 may be connected directly or indirectly via wireless communication, and the exemplary embodiments of the present disclosure do not impose any special restrictions thereon.

[0047] It should be noted that the collection / aggregation, updating, analysis, use, transmission, and storage of user personal information involved in the technical solutions disclosed herein are all in compliance with relevant laws and regulations, are used for legitimate and reasonable purposes, are not shared, disclosed, or sold outside of these legitimate uses, and are subject to supervision and management by national regulatory authorities. Necessary measures should be taken with respect to user personal information to selectively block the use or access of personal information data to prevent unauthorized access to such personal information data, ensure that persons with access to personal information data comply with relevant laws and regulations, and ensure the security of user personal information. Furthermore, once such user personal information data is no longer needed, risks should be minimized by restricting or even prohibiting data collection and / or deleting the data.

[0048] like Figure 2 FIG. 1 is a flow chart of a ski control method according to an exemplary embodiment of the present disclosure. The ski control method includes steps S210 to S230, which are specifically as follows:

[0049] In step S210 , multimodal information of the skier is collected, where the multimodal information includes the skier's main information and the skier's environmental interaction information during the skiing process.

[0050] In the exemplary embodiments of the present disclosure, a skier's subject information refers to static or dynamic information directly related to the skier's own attributes, which can be collected through wearable devices, embedded sensors, etc. Subject information may include the skier's biometric data, equipment status data, and identity attribute data. Biometric data includes information such as the skier's heart rate, body temperature, muscle condition, and fatigue level; equipment status data includes data from the ski board's pressure sensor and feedback from wearable devices (such as smart protective gear); and identity attribute data includes basic user attributes such as age, weight, and skill level.

[0051] Environmental interaction information is dynamic information that reflects the interaction between the skier and the external environment. It can be collected by environmental sensors (such as cameras), positioning systems (such as the Beidou Positioning System, GPS (Global Positioning System), etc.), meteorological equipment, etc. Environmental interaction information can include the skier's motion state data, environmental perception data, and behavioral decision data. Motion state data includes position, speed, acceleration, posture (such as tilt angle), etc. Environmental perception data includes snow quality, temperature, wind speed, obstacle location, surrounding skier information (such as distance), etc. Behavioral decision data includes turning actions, emergency stop records, path selection records, etc.

[0052] In some optional embodiments, collecting multimodal information of a target skier may include: obtaining the skier's registration information, obtaining sensor information from the skier's wearable device, and performing feature extraction on the collected image of the skier to obtain image feature information, and determining multimodal information based on the registration information, sensor information, and image feature information.

[0053] Specifically, registration information can be collected by the ski resort management system or mobile application when the user registers, providing benchmark parameters for subsequent ski control decisions. The sensors in the wearable device may include IMU (Inertial Measurement Unit), pressure sensor, biosensor, and positioning module. The IMU can be embedded in the snow boot or snowboard binding to obtain three-axis acceleration / angular velocity (attitude, sliding speed). The pressure sensor can be implemented through the thin film pressure array of the snow boot lining to obtain the force distribution of the foot. The biosensor can be implemented through the fabric electrodes of the ski suit to collect heart rate and muscle tension. The positioning module can be integrated on the top of the helmet, or it can be used with the ski resort differential base station to improve accuracy and obtain the absolute position of the skier.

[0054] When collecting images of skiers, they can be collected through image acquisition devices on the skier's helmet or glasses, or through image acquisition devices set up at the ski resort, such as drone aerial photography and wide-angle camera acquisition.

[0055] Among them, the image feature information extracted from the skier's image may include the skier's body features, such as skeletal key points, ski-snow contact area, texture, shape, color, etc., and may also include environmental interaction features, such as snow track texture, obstacles, etc.

[0056] In some optional embodiments, a pre-trained target recognition model may be used to perform feature extraction on the collected images of the skier. For example, a machine learning model or a deep learning model may be used to extract image features, such as a lightweight MobileNet network to process the collected images to extract image feature information, and then part or all of the subject information and / or environmental interaction information can be determined based on the image feature information.

[0057] Before extracting features from the skier images, the images can be preprocessed to improve the accuracy of feature extraction. For example, image resizing, contrast adjustment, denoising, and lighting compensation can be performed.

[0058] As an example, for images captured by a helmet, a physical polarizer can be used to filter out the mirror-reflected light from the snow surface to avoid overexposure of the image due to strong reflections from the snow.

[0059] Of course, multimodal information can also be obtained through other methods, such as detecting surrounding skiers through radar, or extracting and identifying features from frame images collected through radar images, or performing target detection and identification on three-dimensional depth maps, etc. Exemplary embodiments of the present disclosure include but are not limited to the above-mentioned methods for collecting multimodal information.

[0060] It should be understood that, for a skier, a skier may be a target skier for themselves, but may be a surrounding skier for other skiers. In other words, the target skier and surrounding skiers in the exemplary embodiments of the present disclosure are relative. When acquiring multimodal information about a skier, all skiers at the ski resort are considered as information collection targets, and this will not be repeated hereafter.

[0061] In step S220 , the skier is tracked and the skiing trajectory is predicted based on the multimodal information to obtain a first predicted skiing trajectory of the target skier and second predicted skiing trajectories of skiers around the target skier.

[0062] In an exemplary embodiment of the present disclosure, object tracking of a skier is performed by determining the position of the skier in real time based on multimodal data. Optionally, the appearance of the skier can be modeled based on multimodal information, and the target can be found in subsequent frames through the appearance modeling results, such as area matching, feature point matching or optical flow method. Optionally, the search area can be narrowed down based on the search method to search in the narrowed search area to achieve object tracking, for example, using Kalman filtering, ion filtering, etc. to narrow the search area. Alternatively, the gradient descent method can be used to determine the search direction to find the most likely target position, such as using Mean Shift (mean shift) algorithm, Cam shift (Continuously Adaptive Mean-Shift) algorithm, etc. The current object tracking methods can all be applied to this solution. The surrounding skiers of the target skiing circle can be determined through object tracking, and the exemplary embodiments of the present disclosure do not impose special restrictions on this.

[0063] Predicting a skier's trajectory involves predicting their location within the next few seconds or even more than ten seconds. This prediction requires predicting both the target skier's trajectory and the trajectories of their surrounding skiers. Optionally, a pre-trained deep learning model can be used for prediction. Specifically, the model takes the skier's multimodal information as input and outputs a predicted trajectory. Pre-trained deep learning models can include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), LSTM (Long Short-Term Memory), or Transformer-based network models. Alternatively, by assuming a linear relationship between the future position and the current multimodal information, linear regression can be used to predict the trajectory. Alternatively, a Markov model can be used to predict the skier's trajectory based on the skier's multimodal information, using the current state to predict the next state. Of course, other methods for predicting the skier's trajectory can also be used, such as Kalman filtering and particle filtering. The exemplary embodiments of this disclosure can customize the method for predicting ski trajectories based on actual needs.

[0064] It is worth noting that the exemplary embodiments of the present disclosure can obtain multimodal information of each skier, thereby predicting a first predicted skiing trajectory based on the multimodal information of the target skier, and at the same time predicting a second predicted skiing trajectory of the skiers surrounding the target skier.

[0065] In step S230, analysis and calculation are performed based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a skiing control strategy, and collision warning and / or path recommendation are performed for the target skier according to the skiing control strategy.

[0066] In the exemplary embodiments of the present disclosure, as mentioned above, the first predicted ski trajectory is the future sliding path predicted by the target skier (i.e., the skier currently being tracked and analyzed) based on the current motion state (such as speed, direction, posture) and environmental information (such as the slope of the snow track, snow quality), and the second predicted ski trajectory is the predicted sliding path of other skiers around the target skier (peripheral skiers), which is usually obtained through object tracking and motion modeling. The ski control strategy is a decision-making plan generated based on the analysis results, which may include collision warnings, path adjustment suggestions, etc. Among them, when it is predicted that the target skier may collide with surrounding skiers or obstacles, a warning may be issued through visual, auditory or tactile means (such as vibration of a smart helmet), and the path recommendation is to provide skiers with optimized path suggestions (such as avoiding crowded areas or dangerous terrain), which may be executed through AR (Augmented Reality) glasses, smart watches or voice prompts.

[0067] The system can analyze and calculate the presence of a collision risk based on the first and second predicted ski trajectories and multimodal information, and issue a collision warning to the target skier if a collision risk exists. For example, a collision warning can be issued if the first and second predicted ski trajectories intersect, and / or if the distance between the target skier and an obstacle is determined to be below a safety threshold based on multimodal information.

[0068] When the snow road conditions are determined to be risky based on multimodal information, the range of route recommendations can be limited. For example, if the snow road is icy or has steep slopes, it can be recommended to slow down or choose a flatter route.

[0069] The skiing control method in the exemplary embodiment of the present disclosure, on the one hand, achieves complete perception of skiing conditions by simultaneously collecting information about the skier's main body and environmental interaction, resolving the misjudgment problem caused by traditional single-dimensional monitoring and improving the accuracy of identifying potential risks. On the other hand, the dual prediction mechanism (the first predicted trajectory is used for self-path prediction, and the second predicted trajectory is used for behavior analysis of surrounding skiers) enables analysis of collision scenarios in the future, enabling earlier detection of dangers than existing reactive warning systems, thereby guiding active obstacle avoidance during skiing and reducing the risk of skiing collisions. Furthermore, by coupling and analyzing multimodal information with predicted trajectories, the generated skiing control strategy includes collision warnings and / or path recommendations. This not only provides early warning of collisions, but also provides skiers with optimal obstacle avoidance routes or braking suggestions through path recommendations, thereby improving the skier's decision-making efficiency during complex skiing processes.

[0070] In an exemplary embodiment, analyzing and calculating the first predicted ski trajectory, the second predicted ski trajectory, and the multimodal information to obtain a ski control strategy, and providing a collision warning and / or a path recommendation to a target skier based on the ski control strategy, includes:

[0071] If the target skier is determined to meet predetermined collision conditions based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information, a collision warning message is generated and provided to the target skier. Alternatively, a route is generated based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information to obtain a recommended skiing route, which is then provided to the target skier.

[0072] Specifically, the preset collision conditions may include the following:

[0073] If the first predicted skiing trajectory and the second predicted skiing trajectory meet the first collision condition, and / or the target multimodal information indicates that the distance between the target skier and the obstacle object meets the second collision condition, then collision warning information is generated, and the obstacle object includes the surrounding skiers and surrounding obstacles of the target skier.

[0074] The first collision condition can be that there is at least one intersecting track location (i.e., a potential collision point) between the first predicted skiing trajectory and the second predicted skiing trajectory, indicating a risk of collision between the target skier and surrounding skiers. The second collision condition can be that the distance between the target skier and an obstacle object, determined based on multimodal information, is less than a distance threshold, such as 0.5 meters, indicating a risk of collision between the target skier and the obstacle object. The distance threshold can also be flexibly adjusted based on the skill level of the target skier. When the target skier's skill level is high, their ability to avoid obstacles is greater, and the distance threshold can be appropriately reduced. Conversely, when the target skier's skill level is low (such as a novice), the distance threshold can be increased.

[0075] Of course, the predicted skiing trajectory and multimodal information may also be combined to determine whether the target skier meets a predetermined collision condition. That is, when the first predicted skiing trajectory and the second predicted skiing trajectory meet a first collision condition, and the target multimodal information indicates that the distance between the target skier and the obstacle meets a second collision condition, a collision risk is determined to exist.

[0076] After the collision warning information is generated, the collision warning information can be provided to the target skier, including but not limited to prompts in the form of voice broadcast, light adjustment or vibration through helmets, AR glasses, smart watches, etc.

[0077] In an exemplary embodiment of the present disclosure, a method for generating a recommended ski route is also provided. Figure 3As shown, if a path is generated based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information, and a recommended skiing path is obtained and provided to a target skier, the method may include:

[0078] Step S310: Determine the target skiing power of the target skier based on the multimodal information and the first relationship model, where the first relationship model is used to represent the correspondence between the multimodal information and the skiing power.

[0079] Skiing force is the mechanical parameter applied to the ski by the skier during skiing, including the mechanical parameters applied to the skis and the skier's grip strength. This can be obtained through questionnaires, laboratory tests, or public datasets. Furthermore, multimodal information related to skiing force is obtained. This multimodal information may affect skiing force, for example, skiers of different ages and heights may have different skiing forces.

[0080] Based on the collected data, historical skiing data can be trained using machine learning (e.g., neural networks or random forests) to establish a mapping between multimodal information and skiing power, thereby generating a first relational model. Thus, the multimodal information of a target skier can be input into the first model, and a target skiing power can be output.

[0081] Step S320: Determine a target skiing turning radius based on the target skiing power, the multimodal information, and the second relationship model, where the second relationship model is used to characterize the correspondence between skiing power and turning radius under different skiing styles.

[0082] The ski turn radius is the ideal radius of curvature of the path required for a skier to complete a turn in their current motion state. It is determined by factors such as skiing power and skiing style (speed, ski angle, and trail conditions). The second relationship model is used to establish a mathematical or data-driven model of the correspondence between skiing power and turn radius, taking into account the impact of different skiing styles.

[0083] A machine learning model or deep learning model (e.g., random forest or neural network) can be trained using skiing power, multimodal information, and turn radius from historical skiing data to generate a second relationship model. The multimodal information includes information related to different skiing styles. The target skiing power and multimodal information are then input into the second relationship model to determine the target skiing turn radius.

[0084] By establishing the second relationship model, it is possible to provide skiers with scientific turning constraints that are consistent with their abilities, environmental conditions, and sports goals, thereby optimizing skiing safety and smoothness.

[0085] Step S330: Using the second predicted ski trajectory as a roadblock, constructing a search space based on the ski trail, and performing a path point search in the search space based on the target ski turning radius to determine a recommended ski path based on the obtained target path points.

[0086] The search space refers to the navigable area within the skiing environment. It is a virtual path planning range determined by dynamic constraints such as trail boundaries, obstacles, and the positions of other skiers. It can be understood as a discretized spatial representation used for path planning. Waypoint search is the process of finding feasible waypoints within the search space based on motion constraints such as the ski turn radius, ultimately forming a continuous and safe recommended path.

[0087] By deeply integrating the principles of skiing mechanics, real-time environmental perception and intelligent algorithms, a leap from "passive response" to "active prevention" has been achieved, improving skiing safety and user experience.

[0088] Among them, such as Figure 4 As shown, the second predicted ski trajectory is used as a roadblock, a search space is constructed according to the ski trail, and a path point search is performed in the search space based on the target ski turning radius to generate a recommended ski path based on the obtained target path points, including:

[0089] Step S410: Constructing a search space based on the ski trail, wherein each grid node in the search space includes a state vector for characterizing the facing angle and skiing turning radius of the skier at the grid node, and marking the second predicted skiing trajectory as a roadblock node in the search space.

[0090] The data of the snow track may include static data and dynamic data. The static data is a pre-stored electronic map of the snow track, including but not limited to width, slope, curve position, fixed obstacles such as trees, cable car poles, etc. The dynamic data can be updated in real time through GPS, cameras, etc. A search space is constructed based on the data of the snow track. Each grid node in the search space represents a passable location on the snow track and contains state information (i.e., state vector).

[0091] The second predicted ski trajectory (the future path of surrounding skiers) is treated as a dynamic obstacle, and its location is marked in the search space to indicate the area it occupies. For example, if a skier is predicted to reach position (x, y) in 3 seconds, this area is marked as temporarily impassable. When searching for feasible pathpoints, these marked points are excluded or avoided. For example, if the target skier is expected to reach a grid cell at time t, and this grid cell has been marked as a roadblock node, this is considered a conflict and is not used as a pathpoint. In addition, grid cells outside the ski slope can be marked as permanent obstacles.

[0092] The state vector of each grid node in the constructed search space is used to describe the skier's dynamic state at that grid node, including at least the facing angle and the ski turn radius. The facing angle is the skier's current direction of travel (e.g., 0° represents due north, 90° represents due east), and the ski turn radius is the minimum turn radius allowed at that location (determined by the curvature of the ski slope, snow quality, etc.).

[0093] As an example, the ski slope can be divided into a two-dimensional grid with a fixed resolution (such as 1 meter × 1 meter). Each grid corresponds to a node. The state vector of each node includes the coordinates (x, y) reflecting the physical position, the facing angle θ (the initial value is set according to the direction of the ski slope (such as along the center line of the ski slope)) and the ski turning radius, which is determined by the local curvature of the ski slope.

[0094] Step S420: Determine the grid node corresponding to the current position of the target skier as the initial node, and search for path points in the search space according to the state vector of each grid node under the constraints of the target skiing turning radius and the ski trail radius.

[0095] Step S430: If the destination position of the target skier is found to be a path point, a target path point is determined based on each path point between the destination position and the initial node, so as to generate a recommended skiing route based on the obtained target path point.

[0096] The initial node represents the current grid location of the target skier and can contain the position coordinates (x, y) and the initial state vector (facing angle θ0, current turning radius R0). The target ski turning radius constraint refers to the minimum radius limit for the skier to safely complete a turn in the current motion state, and the local maximum allowable turning radius determined by the geometry of the ski trail itself (the ski trail radius constraint). These two constraints ensure that when generating the search for subsequent nodes, the turning radius is not less than the target ski turning radius and the path does not deviate from the ski trail.

[0097] The Hybrid A* algorithm can be used to generate recommended ski routes for target skiers. Of course, the algorithm can also be flexibly adjusted according to actual scenario requirements, and recommended ski routes can be generated under the constraints of turning radius.

[0098] An exemplary embodiment of the present disclosure, by marking the second predicted skiing trajectory as a roadblock node, ensures that the future positions of surrounding skiers are actively avoided during path planning, significantly reduces the risk of collision, and strictly verifies the feasibility of the turning radius and ski trail width constraints, avoids recommending sharp turning paths that are beyond the skier's control ability, and prevents loss of control or falls.

[0099] In an exemplary embodiment, another implementation method for generating a recommended ski route is provided. If a route is generated based on the first predicted ski trajectory, the second predicted ski trajectory, and multimodal information, and a recommended ski route is obtained and provided to a target skier, the method may include:

[0100] Get the pre-trained path generation model;

[0101] Through the pre-trained path generation model, a recommended ski path is generated according to the first predicted ski trajectory, the second predicted ski trajectory and multimodal information.

[0102] Among them, the pre-trained path generation model is obtained by training the path planning model to be trained based on the skier data samples, including sample data of the first predicted ski trajectory, sample data of the second predicted ski trajectory and sample data of multimodal information, as well as corresponding sample labels, that is, sample data of recommended ski paths.

[0103] The pre-trained path generation model can be a deep learning model, a large model, etc., and the exemplary embodiment of the present disclosure does not limit the model structure.

[0104] In some optional embodiments, if the second predicted ski trajectory is used as a roadblock, a search space is constructed according to the snow trail, and a path point search is performed in the search space based on the target ski turning radius, but the target skier's end position is not found as a path point, or the open list is empty, then at this time, a recommended ski path can be generated based on a pre-trained path generation model as a supplementary method to ensure that the recommended ski path can be provided to the skier in a timely manner during the skiing process.

[0105] In an exemplary embodiment, a method for predicting skiing trajectories is also provided. Figure 5 As shown in FIG, the skier is tracked and the skiing trajectory is predicted based on multimodal information, including:

[0106] Step S510: Generate a skier's sports endurance evaluation coefficient based on the biometric data, equipment status data, and identity attribute data.

[0107] Among them, the sports endurance assessment coefficient is a dimensionless dynamic parameter used to quantify the skier's ability to continue exercising in the current state. The range can be 0.1-1.0, and it can also be set according to actual scenario requirements.

[0108] A model of biometric data-equipment status data-identity attribute data and sports endurance coefficient can be established in advance. The model to be trained can be trained using the collected relevant sample data to obtain a sports endurance evaluation model. Based on the sports endurance evaluation model, the biometric data, equipment status data and identity attribute data are used as input to output the skier's sports endurance evaluation coefficient.

[0109] For example, multimodal information of at least 500 skiers may be collected as multi-scenario data to train the sports endurance assessment model to be trained, so as to be used for assessing the skiers' sustained exercise ability.

[0110] Specifically, during model training, the biometric data, equipment status data, and identity attribute data of the collected sample data can be input into the sports endurance assessment model to be trained to obtain a predicted sports endurance assessment coefficient. A loss value is then calculated based on the predicted sports endurance assessment coefficient and the actual measured sports endurance assessment coefficient. The parameters of the sports endurance assessment model to be trained are adjusted based on the obtained loss value. Training is repeated multiple times until a training stop condition is reached, such as reaching the training number or model convergence, resulting in a completed sports endurance assessment model. The loss value can be calculated using methods such as cross-entropy loss.

[0111] By converting the complex physiology-equipment-environment relationship into a real-time quantitative endurance index, the model can provide basic data for intelligent decision-making for skiing safety and performance improvement.

[0112] Step S520: Determine a trajectory search area based on the environmental perception data, the behavior decision data, and the exercise endurance evaluation coefficient.

[0113] The trajectory search area refers to the candidate path space that meets safety and accessibility requirements. The initial range can be determined based on environmental perception data and the sports endurance assessment coefficient. Specifically, a pre-trained range determination model can be used to determine the initial range based on the environmental perception data and the sports endurance assessment coefficient. Specifically, the pre-trained range determination model is trained based on behavioral environment data and sports endurance data samples. It can be a deep learning model or a large model, and there is no restriction on this.

[0114] Then, the initial range is modified according to the terrain constraints in the environmental perception data, for example, areas with slopes greater than a safety threshold (such as 35°) are eliminated from the initial range, and areas where obstacles are located are eliminated from the initial range to obtain a reference area.

[0115] Next, based on the behavioral decision data, a pre-trained behavioral decision model is used to predict the skier's turn probability. Furthermore, based on the reference area and turn probability, multiple candidate paths are generated. The area covered by these candidate paths becomes the final trajectory search area. The pre-trained behavioral decision model is trained based on behavioral decision data samples and can be a deep learning model or a large model. The exemplary embodiments of this disclosure do not impose any particular limitations on this model.

[0116] Step S530: predicting a track position in the track search area according to the running status data, and determining the predicted skiing track of the skier according to the track position.

[0117] Predicting ski trajectories involves outputting the probability distribution of a skier's skiing locations over the next few seconds. This can be achieved using an LSTM (Long Short-Term Memory) time series prediction network, which includes an input layer, an LSTM layer, and an output layer. The LSTM layer can use a two-layer stacked LSTM structure with 64 neurons per layer. The forget gate bias is initialized to 1.5 to enhance long-term memory capabilities. The output layer can use a fully connected layer to predict the two-dimensional coordinates of future trajectory points. Of course, other prediction models can also be selected based on actual needs and scenarios.

[0118] During the training phase, the continuous time series is divided into sliding windows of fixed length (to ensure time series continuity) to generate training samples, wherein the operating status data (speed, acceleration) is standardized, and the LSTM time series prediction network is trained based on the training samples, wherein the weighted MSE (Mean Squared Error) can be used as the loss function. The exemplary embodiments of the present disclosure do not specifically limit the process of training the LSTM time series prediction network.

[0119] The motion state data is processed using the trained model. When predicting the trajectory position, it is necessary to make the obtained trajectory position fall within the trajectory search area under the constraint of the trajectory search area to obtain the predicted skiing trajectory.

[0120] It should be understood that the above-mentioned method of predicting skiing trajectories is applicable to both target skiers and surrounding skiers, and no distinction is made here.

[0121] An exemplary embodiment of the present disclosure generates a personalized endurance coefficient through real-time fusion of biometric data and equipment status data, quantifies the trend of skier's athletic ability decline, and preferentially generates a trajectory search area that conforms to user habits based on the endurance coefficient, environmental perception data, and behavioral decision data. Then, it can further search for predicted skiing trajectories in the trajectory search area based on the motion status data, thereby improving the correlation between the predicted skiing trajectory and multimodal information such as users and ski resorts, and improving the accuracy of the predicted skiing trajectory.

[0122] In an exemplary embodiment, the environmental interaction information includes skiing weather information, which refers to meteorological data used to evaluate the impact of weather conditions on the safety, comfort and snow quality of skiing, such as temperature, wind speed, snowfall, snow quality, etc.

[0123] Based on this, Figure 6As shown, the multimodal information of the skier is collected, which also includes:

[0124] Step S610: monitor the confidence level of ski weather information in real time.

[0125] Step S620: If the confidence level of the ski weather information is lower than a preset confidence level, the ski weather data is corrected.

[0126] The confidence level of ski weather information is a quantitative measure of its reliability in skiing scenarios. This information can be compared with weather station data and data from third-party weather applications. If the data discrepancy exceeds a preset threshold, the confidence level of the ski weather information is determined to be lower than the preset level.

[0127] Weather station data can be from professional monitoring equipment deployed at ski resorts (such as automatic weather stations and laser wind radars), providing high-precision real-time data. Third-party weather applications refer to weather platforms that rely on global data model output.

[0128] Specifically, the ski weather information can be time-aligned with the weather station data and the data from the third-party weather application, for example, by interpolating the data onto the same grid. Then, the data differences between the ski weather information and the weather station data, and between the ski weather information and the data from the third-party weather application, are obtained. If at least one of the obtained data differences exceeds its corresponding preset difference threshold, the confidence level of the ski weather information is determined to be below a preset confidence level.

[0129] As an example, if a third-party application predicts a temperature of -5°C, the weather station actually measures -8°C, and the ski weather information indicates a temperature of -3°C, the temperature difference between the ski weather information and the temperature forecast by the third-party application is 2°C, and the temperature difference between the ski weather information and the temperature actually measured by the weather station is 5°C (exceeding the threshold of ±2°C), then it is determined that the confidence level of the ski weather information is lower than the preset confidence level.

[0130] It should be understood that different categories of ski weather information have their own corresponding preset difference thresholds, which can be flexibly set based on actual experience and ski resort conditions.

[0131] Through cross-validation of multi-source data, it can provide precise meteorological risk management for skiing scenes and provide accurate meteorological data support for formulating skiing control strategies.

[0132] In some optional embodiments, if the confidence level of the ski weather information is lower than a preset confidence level, the ski weather data is modified, including at least one of the following:

[0133] 1) Generate alternative data for the skiing weather data based on historical weather data for the same period of the current ski slope.

[0134] Historical meteorological data for the same period refers to meteorological observation records of the same ski slope in the same time period in the past (such as the same date and time period in the past three years), which is used to establish a benchmark reference for data correction. For example, the wind speed, temperature and other factors of the past 24 hours can be used to calculate the mean to obtain replacement data. Specifically, if the real-time wind speed sensor of the ski slope fails and the current wind speed shows 0m / s, the wind speed data at 10 am on the same day in the past three years (mean 4.2m / s, standard deviation 0.8m / s) is retrieved to generate replacement data of 4.2±0.8m / s.

[0135] 2) Obtain meteorological data of the adjacent areas of the current ski slope and perform spatial interpolation based on the meteorological data to obtain alternative data for skiing meteorological data.

[0136] Among them, the Kriging interpolation method can be used to interpolate the data of adjacent areas based on the meteorological data of the current ski slope to obtain replacement data. For example, if the temperature sensor in Area A of the alpine ski track is damaged and the -8°C outlier needs to be repaired, the adjacent areas B (-12°C), C (-11°C), and D (-10°C) are obtained. Kriging interpolation is performed based on the meteorological data of the adjacent areas, and the interpolated temperature result in Area A is -11.2°C (error ±0.6°C) as the replacement data.

[0137] 3) Send an alarm signal to the ski resort management terminal and receive the replacement data sent by the ski resort management terminal. In other words, it can also receive the correction value entered by manual input.

[0138] By correcting ski weather information, we can reduce the risk of incorrect ski control strategies due to data errors and avoid ski accidents.

[0139] The skiing control method in the exemplary embodiment of the present disclosure, on the one hand, constructs a complete skiing situational awareness model by simultaneously collecting information about the skier's main body and the environment's interaction with it. This model addresses the misjudgment problem caused by traditional single-dimensional monitoring and improves the accuracy of identifying potential risks. Furthermore, the dual prediction mechanism (a first predicted trajectory for the skier's own path prediction and a second predicted trajectory for analyzing the behavior of surrounding skiers) enables analysis of collision scenarios in the future, enabling earlier detection of dangers than existing reactive warning systems, thereby guiding proactive obstacle avoidance during skiing and reducing the risk of skiing collisions. Furthermore, by coupling multimodal information with predicted trajectories for analysis, the generated skiing control strategy includes collision warnings and / or path recommendations. This not only provides early warning of collisions but also provides skiers with optimal obstacle avoidance routes or braking suggestions through path recommendations, thereby improving the skier's decision-making efficiency during complex skiing situations.

[0140] In an exemplary embodiment of the present disclosure, a ski control device is also provided. Figure 7As shown, the ski control device 700 may include an information acquisition module 710, a trajectory prediction module 720, and a control scheduling module 730. Specifically:

[0141] The information collection module 710 is used to collect multimodal information of the skier, which includes the skier's main information and the skier's environmental interaction information during the skiing process; the trajectory prediction module 720 is used to track the skier and predict the skiing trajectory based on the multimodal information, thereby obtaining a first predicted skiing trajectory of the target skier and a second predicted skiing trajectory of the target skier's surrounding skiers; the control scheduling module 730 is used to perform analysis and calculation based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a skiing control strategy, and to provide collision warnings and / or path recommendations to the target skier based on the skiing control strategy.

[0142] In an exemplary embodiment of the present disclosure, the control scheduling module 730 is configured to execute: if it is determined based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information that the target skier meets the predetermined collision condition, then collision warning information is generated and provided to the target skier; and / or, path generation is performed based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a recommended skiing path and provide it to the target skier.

[0143] In an exemplary embodiment of the present disclosure, the subject information includes the skier's biometric data, equipment status data and identity attribute data; the environmental interaction information includes the skier's motion status data, environmental perception data and behavioral decision data; the trajectory prediction module 720 is configured to perform: generating the skier's motion endurance evaluation coefficient based on the biometric data, equipment status data and identity attribute data; determining the trajectory search area based on the environmental perception data, behavioral decision data and motion endurance evaluation coefficient; predicting the trajectory position in the trajectory search area based on the running status data, and determining the skier's predicted skiing trajectory based on the trajectory position.

[0144] In an exemplary embodiment of the present disclosure, the environmental interaction information includes ski weather information; the information collection module 710 is further configured to perform: real-time monitoring of the confidence level of the ski weather information; if the confidence level of the ski weather information is lower than a preset confidence level, correcting the ski weather data.

[0145] In an exemplary embodiment of the present disclosure, if the confidence level of the skiing weather information is lower than a preset confidence level, the skiing weather data is corrected, including at least one of the following: generating alternative data for the skiing weather data based on the historical weather data of the current ski slope during the same period; obtaining weather data of adjacent areas of the current ski slope, and performing spatial interpolation based on the weather data to obtain alternative data for the skiing weather data; sending an alarm signal to the ski resort management end, and receiving the alternative data sent by the ski resort management end.

[0146] In an exemplary embodiment of the present disclosure, determining that the confidence level of ski weather information is lower than a preset confidence level includes: comparing the ski weather information with weather station data and data differences from a third-party weather application; if the data difference exceeds a preset difference threshold, determining that the confidence level of the ski weather information is lower than the preset confidence level.

[0147] In an exemplary embodiment of the present disclosure, the control scheduling module 730 is configured to execute: if the first predicted skiing trajectory and the second predicted skiing trajectory meet the first collision condition, and / or the target multimodal information indicates that the distance between the target skier and the surrounding skiers meets the second collision condition, then generate collision warning information; and provide the collision warning information to the target skier.

[0148] In an exemplary embodiment of the present disclosure, the control scheduling module 730 is configured to perform: determining the target skiing power of the target skier based on multimodal information and a first relationship model, the first relationship model being used to characterize the correspondence between the multimodal information and the skiing power; determining the target skiing turning radius based on the target skiing power, the multimodal information and the second relationship model, the second relationship model being used to characterize the correspondence between the skiing power and the turning radius under different skiing methods; using the second predicted skiing trajectory as a roadblock, constructing a search space based on the snow trail, and performing a path point search in the search space based on the target skiing turning radius to determine a recommended skiing path based on the obtained target path points.

[0149] In an exemplary embodiment of the present disclosure, a second predicted ski trajectory is used as a roadblock, a search space is constructed according to a snow track, and a path point search is performed in the search space based on a target ski turning radius to generate a recommended ski path based on the obtained target path points, including: constructing a search space according to the snow track, wherein each grid node in the search space includes a state vector for characterizing the facing angle and ski turning radius of the skier at the grid node, and marking the second predicted ski trajectory as a roadblock node in the search space; determining the grid node corresponding to the current position of the target skier as the initial node, and searching for path points in the search space through a heuristic strategy based on the state vectors of each grid node under the constraints of the target ski turning radius and the snow track radius; if the end position of the target skier is found to be a path point, then determining the target path point based on each path point between the end position and the initial node to generate a recommended ski path based on the obtained target path point.

[0150] In an exemplary embodiment of the present disclosure, the control scheduling module 730 is configured to execute: obtaining a pre-trained path generation model; and generating a recommended ski path according to the first predicted ski trajectory, the second predicted ski trajectory and multimodal information through the pre-trained path generation model.

[0151] In an exemplary embodiment of the present disclosure, the information acquisition module 710 is configured to: obtain registration information of a skier; obtain sensor information from a wearable device of the skier; perform feature extraction on the collected image of the skier to obtain image feature information; and determine multimodal information based on the registration information, sensor information, and image feature information.

[0152] Since the details of the functional modules of the ski control device of the exemplary embodiment of the present disclosure have been described in the exemplary embodiment of the ski control method described above, they will not be repeated here.

[0153] It should be noted that while the above detailed description refers to several modules or units of the ski control device, this division is not mandatory. In fact, depending on the embodiments of the present disclosure, the features and functions of two or more modules or units described above may be embodied in a single module or unit. Conversely, the features and functions of a single module or unit described above may be further divided and embodied by multiple modules or units.

[0154] The exemplary embodiments of the present disclosure further provide a computer program product, which includes a computer program, and when the computer program is executed by a processor, implements the above-mentioned ski control method.

[0155] In one embodiment, a computer program product may be a tangible product containing a computer program, such as a computer-readable storage medium storing the computer program. The computer-readable storage medium may be a storage medium based on electrical, magnetic, optical, electromagnetic, infrared, or other signals, including but not limited to random access memory (RAM), read-only memory (ROM), magnetic tape, floppy disk, flash memory (Flash), mechanical hard disk drive (HDD), solid-state drive (SSD), and the like. Exemplarily, the computer program product may be implemented as a non-volatile storage medium storing the computer program, such as a read-only memory, NAND flash memory, and the like.

[0156] In one embodiment, the computer program product may be an intangible product containing a computer program. For example, the computer program product may be implemented as a virtual digital product, such as a digital file such as an executable file or installation package storing the computer program.

[0157] The code of the computer program can be written in one or more programming languages. Programming languages ​​include C, Java, C++, etc. The program code can be executed entirely on the user computing device, partially on the user computing device, or as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device via any type of network, such as a local area network (LAN), a wide area network (WAN), etc., or can be connected to an external computing device (e.g., via an Internet connection provided by a carrier).

[0158] Computer programs can be carried or transmitted via electrical, magnetic, optical, electromagnetic, infrared, or other signals. Electronic devices can convert signals carrying the computer program into digital signals, thereby executing the computer program. When the computer program is executed on an electronic device, its code causes the electronic device (more specifically, the processor of the electronic device) to execute the method steps of various exemplary embodiments of the present disclosure, such as the ski control method described above.

[0159] In addition, in an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided. Those skilled in the art will appreciate that various aspects of the present disclosure can be implemented as a system, method, or program product. Therefore, various aspects of the present disclosure can be specifically implemented in the following forms, namely: a complete hardware embodiment, a complete software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, which can be collectively referred to herein as a "circuit," "module," or "system."

[0160] Refer to the following Figure 8800 according to this embodiment of the present disclosure will be described. Figure 8 The electronic device 800 shown is merely an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0161] like Figure 8 As shown, electronic device 800 is implemented as a general-purpose computing device. Components of electronic device 800 may include, but are not limited to, the aforementioned at least one processing unit 810, the aforementioned at least one storage unit 820, a bus 830 connecting various system components (including storage unit 820 and processing unit 810), and a display unit 840.

[0162] The storage unit stores program code, which can be executed by the processing unit 810, so that the processing unit 810 performs the steps according to various exemplary embodiments of the present disclosure described in the "Exemplary Methods" section above. For example, the processing unit 810 can perform the steps shown above.

[0163] The storage unit 820 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 821 and / or a cache memory unit 822 , and may further include a read-only memory unit (ROM) 823 .

[0164] The storage unit 820 may also include a program / utility 824 having a set (at least one) of program modules 825, such program modules 825 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0165] Bus 830 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0166] The electronic device 800 can also communicate with one or more external devices 900 (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 800, and / or any device that enables the electronic device 800 to communicate with one or more other computing devices (e.g., a router, a modem, etc.). Such communication can occur via an input / output (I / O) interface 850. Furthermore, the electronic device 800 can also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network such as the Internet) via a network adapter 860. As shown, the network adapter 860 communicates with other modules of the electronic device 800 via a bus 830. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 800, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0167] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0168] Furthermore, the figures above are merely illustrative of the processes included in the methods according to exemplary embodiments of the present disclosure and are not intended to be limiting. It is readily understood that the processes illustrated in the figures above do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0169] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow from the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the claims.

Claims

1. A ski control method, characterized in that: include: Collecting multimodal information of the skier, wherein the multimodal information includes the skier's main information and the skier's environmental interaction information during the skiing process; Tracking the skier and predicting a skiing trajectory based on the multimodal information to obtain a first predicted skiing trajectory of the target skier and second predicted skiing trajectories of skiers around the target skier; An analysis and calculation is performed based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a skiing control strategy, and a collision warning and / or a path recommendation is performed for the target skier according to the skiing control strategy.

2. The method according to claim 1, characterized in that The analyzing and calculating based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information to obtain a skiing control strategy, and performing collision warning and / or path recommendation for the target skier according to the skiing control strategy, includes: If it is determined based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information that the target skier satisfies a predetermined collision condition, generating collision warning information and providing it to the target skier; And / or, a path is generated based on the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a recommended skiing path and provide it to the target skier.

3. The method according to claim 1, characterized in that The subject information includes biometric data, equipment status data, and identity attribute data of the skier; the environmental interaction information includes motion status data, environmental perception data, and behavioral decision data of the skier; and tracking the skier and predicting the skiing trajectory based on the multimodal information includes: generating a skier's sports endurance assessment coefficient based on the biometric data, equipment status data, and identity attribute data; determining a trajectory search area based on the environmental perception data, the behavioral decision data, and the exercise endurance evaluation coefficient; A trajectory position is predicted in the trajectory search area according to the running state data, and a predicted skiing trajectory of the skier is determined according to the trajectory position.

4. The method according to claim 3, characterized in that The environmental interaction information includes skiing weather information; and the multimodal information collected from skiers further includes: monitoring the confidence level of the ski weather information in real time; If the confidence level of the ski weather information is lower than a preset confidence level, the ski weather data is corrected.

5. The method according to claim 4, characterized in that If the confidence level of the ski weather information is lower than a preset confidence level, correcting the ski weather data includes at least one of the following: generating alternative data for the skiing weather data based on historical weather data for the same period of the current ski slope; Acquiring meteorological data of an area adjacent to the current ski slope, and performing spatial interpolation based on the meteorological data to obtain substitute data for the ski meteorological data; Send an alarm signal to the ski resort management end, and receive replacement data sent by the ski resort management end.

6. The method according to claim 4, characterized in that Determining that the confidence level of the ski weather information is lower than a preset confidence level includes: Comparing the ski weather information with weather station data and third-party weather application data differences; If the data difference exceeds a preset difference threshold, it is determined that the confidence level of the ski weather information is lower than a preset confidence level.

7. The method according to claim 2, characterized in that If it is determined based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information that the target skier meets a predetermined collision condition, generating collision warning information and providing it to the target skier, including: generating collision warning information if the first predicted skiing trajectory and the second predicted skiing trajectory satisfy a first collision condition, and / or if the target multimodal information indicates that the distance between the target skier and an obstacle object satisfies a second collision condition, the obstacle object including surrounding skiers and surrounding obstacles of the target skier; The collision warning information is provided to the target skier.

8. The method according to claim 2, characterized in that Generating a path based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information to obtain a recommended skiing path and providing the path to the target skier includes: determining a target skiing power of the target skier based on the multimodal information and a first relationship model, wherein the first relationship model is used to represent a correspondence between the multimodal information and the skiing power; determining a target skiing turning radius based on the target skiing power, the multimodal information, and a second relationship model, wherein the second relationship model is used to represent a correspondence between skiing power and turning radius under different skiing styles; The second predicted ski trajectory is used as a roadblock, a search space is constructed according to the snow track, and a path point search is performed in the search space based on the target ski turning radius to determine the recommended ski path according to the obtained target path points.

9. The method according to claim 8, characterized in that The step of using the second predicted ski trajectory as a roadblock, constructing a search space according to a snow track, and searching for path points in the search space based on the target ski turning radius, so as to generate the recommended ski route according to the obtained target path points, includes: Constructing a search space based on the ski trail, wherein each grid node in the search space includes a state vector for representing the facing angle and skiing turn radius of the skier at the grid node, and marking the second predicted skiing trajectory as a roadblock node in the search space; Determining a grid node corresponding to the current position of the target skier as an initial node, and searching for path points in the search space according to the state vector of each grid node under the constraints of the target skiing turning radius and the ski trail radius; If the destination position of the target skier is found to be a path point, a target path point is determined based on each path point between the destination position and the initial node, so as to generate the recommended skiing route based on the obtained target path point.

10. The method according to claim 2, characterized in that Generating a path based on the first predicted skiing trajectory, the second predicted skiing trajectory, and the multimodal information to obtain a recommended skiing path and providing the path to the target skier includes: Get the pre-trained path generation model; The recommended ski route is generated according to the first predicted ski trajectory, the second predicted ski trajectory and the multimodal information by using the pre-trained route generation model.

11. The method according to any one of claims 1 to 10, characterized in that The collecting of multimodal information of the target skier includes: Get skier registration information; Obtain sensor information from the skier's wearable device; Performing feature extraction on the collected skier images to obtain image feature information; The multimodal information is determined according to the registration information, the sensor information, and the image feature information.

12. A ski control device, characterized in that: include: An information collection module is used to collect multimodal information of the skier, wherein the multimodal information includes the skier's main information and the skier's environmental interaction information during the skiing process; a trajectory prediction module, configured to track the skier and predict the skiing trajectory based on the multimodal information, thereby obtaining a first predicted skiing trajectory of the target skier and second predicted skiing trajectories of skiers surrounding the target skier; A control scheduling module is used to analyze and calculate the first predicted skiing trajectory, the second predicted skiing trajectory and the multimodal information to obtain a skiing control strategy, and to provide collision warning and / or path recommendation to the target skier according to the skiing control strategy.

13. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 11 is implemented.

14. An electronic device, characterized in that: include: processor; as well as a memory for storing executable instructions of the processor; The processor is configured to perform the method of any one of claims 1 to 11 by executing the executable instructions.

Images