Intelligent auxiliary system for rescue helicopter

Abstract

The invention discloses an intelligent auxiliary system for a rescue helicopter, relates to the technical field of rescue intelligent assistance, solves the technical problems that dynamic obstacle risk quantification precision is insufficient and obstacle avoidance response lacks a hierarchical adaptation mechanism, and converts a path screening standard into a computable priority parameter through a three-level quantification screening mechanism. Compared with the prior art, artificial experience deviation is avoided, the accuracy of path adaptation rescue scene is improved, relative acceleration correction is introduced in collision time calculation, risk underestimation caused by the fact that speed changes are not considered is avoided, the false alarm rate is reduced through risk level division, differentiated responses are designed according to the risk levels, and the accuracy of the path adaptation rescue scene is improved. The advanced risk triggering system forcibly carries out emergency obstacle avoidance, the intermediate risk is combined with the task priority to be manually decided, and the low-level risk is only monitored by the background and does not interfere with the normal task, so that response lag of the high-risk scene is avoided, the efficiency loss of the low-risk scene is reduced, and the completion timeliness of the rescue task is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent assistance technology for rescue operations, specifically to an intelligent assistance system for rescue helicopters. Background Technology

[0002] In the field of emergency rescue, helicopters have become core rescue equipment due to their advantages of rapid response and strong terrain adaptability. With the development of intelligent technology, the path planning and obstacle avoidance systems of rescue helicopters are gradually upgrading from being manually dominated to being technology-assisted.

[0003] Existing technologies still have significant room for improvement in "real-world rescue scenarios": for example, path selection relies heavily on human experience to judge urgency and terrain suitability, lacking quantitative standards; TTC / DCA calculations for dynamic obstacles do not consider factors such as acceleration and three-dimensional height differences, resulting in a high rate of risk misjudgment; obstacle avoidance response does not combine rescue mission priority levels, which can easily lead to problems such as "excessive obstacle avoidance delaying the mission" or "insufficient response leading to risks," making it difficult to meet the precision and efficiency requirements of complex rescue scenarios. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent auxiliary system for rescue helicopters, which solves the problems of insufficient accuracy in quantifying dynamic obstacle risks and the lack of a graded adaptation mechanism for obstacle avoidance response.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent auxiliary system for rescue helicopters, comprising:

[0006] The rescue path generation module is used to acquire rescue mission and scene information, construct a three-dimensional dynamic spatial model to generate rescue paths, and select the optimal rescue path based on the urgency of the rescue mission, and transmit it to the rescue obstacle avoidance analysis module.

[0007] The rescue obstacle avoidance analysis module is used to monitor and avoid obstacles on the optimal rescue path. It identifies static or dynamic obstacles through sensors. For static obstacles, it generates static obstacle avoidance information based on their position and size. For dynamic obstacles, it analyzes whether there is an intersection between the dynamic obstacle and the helicopter's trajectory. If there is an intersection, it generates a collision warning; otherwise, it generates normal monitoring information.

[0008] For collision warning information, the risk is quantified based on collision time and minimum distance, and the collision level is determined by combining the two. Collision analysis signals are generated, including high-level, medium-level and low-level collision risk signals, and transmitted to the obstacle avoidance comprehensive analysis module.

[0009] The obstacle avoidance comprehensive analysis module is used to analyze collision analysis signals, perform obstacle avoidance processing according to different levels of collision analysis signals, generate obstacle avoidance processing information, and transmit it to the intelligent rescue information output module.

[0010] As a further aspect of the present invention, it also includes a rescue scene information collection module, which is used to collect rescue information of the rescue scene. The rescue information includes rescue tasks and scene information. The rescue tasks specifically cover key information corresponding to the number of trapped people, their location, and the degree of urgency. The scene information includes terrain, weather conditions, and the distribution of possible obstacles, and is transmitted to the rescue path generation module.

[0011] The intelligent rescue information output module is used to perform corresponding obstacle avoidance actions based on the acquired obstacle avoidance information.

[0012] As a further aspect of the present invention, the rescue path generation module selects the optimal rescue path in the following way:

[0013] Based on the acquired scene information, a corresponding three-dimensional dynamic spatial model is constructed, and based on the obtained three-dimensional dynamic spatial model and the rescue location, a corresponding rescue path is generated and labeled as i, where i = 1, 2, ..., j, and j represents the type of rescue path.

[0014] Obtain the corresponding rescue missions and use the urgency of the rescue missions as the primary filtering condition. Sort all rescue paths in ascending order of the urgency level of their corresponding rescue missions, and only retain paths with the same highest priority to proceed to the next level of filtering. If multiple rescue missions have the same urgency level, then further use the number of trapped people as a secondary filtering condition, sorting them in descending order of the number of trapped people in the corresponding rescue missions, and retaining paths with the same highest priority to proceed to the next level of filtering. If the number of trapped people is also the same, then filter according to the complexity of the terrain and landforms in the scene information, sorting them in ascending order of complexity level, and the path ranked first is the optimal rescue path.

[0015] As a further aspect of the present invention, the urgency of rescue missions is classified according to the golden rescue time window: if the trapped personnel's vital signs are critical and they need to be reached within 1 hour, the priority is level 1; if the trapped personnel are not in danger of death but the environment is deteriorating, they need to be reached within 2-4 hours, the priority is level 2; if the trapped personnel are in a stable environment and can be reached in more than 4 hours, the priority is level 3.

[0016] The analysis method for the complexity of terrain and landform is as follows: low complexity is defined as no terrain risk, low density of obstacles, and no severe weather; medium complexity is defined as having one assessment risk; and high complexity is defined as having two or more assessment risks.

[0017] As a further aspect of the present invention, the rescue obstacle avoidance analysis module monitors and avoids obstacles along the optimal rescue path in the following way:

[0018] By using multi-sensor collaborative data acquisition, the types of obstacles within the path are identified. For static obstacles, a 3D point cloud is generated by LiDAR scanning and compared with a pre-loaded digital terrain model to identify the location and size of newly added static obstacles and generate static obstacle avoidance information.

[0019] For dynamic obstacles, millimeter-wave radar is used to track moving targets and output speed and heading angle. Obstacle types are classified by combining visual cameras with the YOLOv10 algorithm. Based on the historical motion data of dynamic obstacles, Kalman filtering or LSTM neural network is used to predict the trajectory of the next 3-5 seconds. It is determined whether the trajectory intersects with the current path of the helicopter. If there is no intersection, normal monitoring information is generated and transmitted to the intelligent rescue information output module. If there is an intersection, collision warning information is generated.

[0020] As a further aspect of the present invention, the method for quantifying risk based on collision time and minimum distance is as follows:

[0021] Regarding the collision time, when the helicopter and the dynamic obstacle are moving in the same plane, the relative position vector and relative velocity vector are first calculated, and then the relative velocity along the collision direction is calculated through projection. Finally, the collision time TTC is obtained, and TTC = ,in This represents the position vector of the helicopter relative to the obstacle. Let represent the velocity vector of the helicopter relative to the obstacle, and < ,like · ≥0 means that the two objects are moving in opposite directions and there is no possibility of collision. The collision time is then corrected by considering the acceleration of the dynamic obstacle, according to the formula TTC0= The corrected collision time TTC0 is calculated, where The relative acceleration vector is taken as a positive root to ensure that time is positive;

[0022] For the minimum distance, based on the historical motion data of the obstacle, such as its position, velocity, and acceleration in the past 1-3 seconds, different models are used to predict its position in the next 5 seconds. The straight-line distance between the predicted position of the helicopter and the predicted position of the obstacle is calculated every 100ms. After 50 iterations, the minimum value is taken to obtain the minimum distance DCA.

[0023] As a further aspect of the present invention, the collision analysis signal is generated in the following manner:

[0024] If the minimum distance is less than or equal to 50% of the safe distance and the collision time is less than or equal to 5 seconds, a high-level collision risk signal is generated. If the minimum distance is less than or equal to 50% - 100% of the safe distance and the collision time is between 5 and 10 seconds, a medium-level collision risk signal is generated. If the minimum distance is greater than the safe distance and the collision time is greater than 10 seconds, a low-level collision risk signal is generated.

[0025] As a further aspect of the present invention, the method for analyzing advanced collision risk signals is to interrupt non-critical tasks, prioritize the allocation of computing power to the obstacle avoidance module, call the speed obstacle method or the emergency artificial potential field method to generate a unique optimal detour path, and directly send obstacle avoidance commands to the flight control system without waiting for pilot confirmation. At the same time, adjustments are made based on different scenarios to generate advanced obstacle avoidance analysis information.

[0026] As a further aspect of the present invention, the method for analyzing the intermediate collision risk signal is as follows: generate 2-3 alternative obstacle avoidance paths, mark the key information of each path, select an alternative path within 5 seconds, and confirm execution via the cockpit button; if no confirmation is made within the time limit, the system defaults to selecting the lowest risk path. If the rescue mission is urgent, the shortest detour path can be selected first, and the intermediate obstacle avoidance analysis information can be generated by adjusting according to different scenarios.

[0027] As a further aspect of the present invention, the method for analyzing low-level collision risk signals is as follows: the TTC / DCA value of the obstacle is updated every 2 seconds. The data is only recorded in the background and does not trigger audible and visual alarms. The AR interface only marks the location of the obstacle with a small green icon. If the obstacle's trajectory changes, the risk is calculated in real time. Once the level 2 or level 1 conditions are met, the signal is immediately upgraded and the corresponding response is triggered, and low-level obstacle avoidance analysis information is generated.

[0028] This invention provides an intelligent auxiliary system for rescue helicopters. Compared with the prior art, it has the following advantages:

[0029] This invention employs a three-level quantitative screening mechanism to transform path selection criteria into calculable priority parameters, avoiding biases from human experience and ensuring that the optimal path selection logic is traceable and the results are verifiable. This improves the accuracy of path adaptation to rescue scenarios. Collision time calculation incorporates relative acceleration correction to avoid underestimating risks due to neglecting speed changes. Minimum distance calculation is based on three-dimensional spatial distance and combined with obstacle type to adapt to safe distances. Risk level classification reduces false alarm rates. This application designs differentiated responses based on risk levels: high-risk scenarios trigger the system to force emergency obstacle avoidance; medium-risk scenarios provide 2-3 alternative paths, with decisions made manually based on task priority; and low-risk scenarios are only monitored in the background without interfering with normal tasks. This design avoids response delays in high-risk scenarios and reduces efficiency losses in low-risk scenarios, improving the timeliness of rescue mission completion. Attached Figure Description

[0030] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0032] Please see Figure 1 This application provides an intelligent auxiliary system for rescue helicopters, including: a rescue scene information acquisition module, a rescue path generation module, a rescue obstacle avoidance analysis module, an obstacle avoidance comprehensive analysis module, and an intelligent rescue information output module, and in combination with... Figure 1 It can be seen that the information between the above functional modules is transmitted in one direction only;

[0033] The rescue scene information acquisition module is used to collect rescue information from the rescue scene. The rescue information includes the rescue mission and scene information. The rescue mission specifically covers key information such as the number of trapped people, their location, and the degree of urgency. The scene information includes terrain, weather conditions, and the distribution of possible obstacles. The rescue information is transmitted to the rescue path generation module through various sensors and data acquisition devices.

[0034] The rescue path generation module is used to construct a corresponding 3D dynamic spatial domain model based on the acquired scene information. The construction of the 3D dynamic spatial domain model is an existing technology and will not be elaborated upon here. Based on the obtained 3D dynamic spatial domain model and the rescue location, a corresponding rescue path is generated. Simultaneously, the rescue path is filtered according to the rescue mission to obtain the optimal rescue path. The specific filtering method is as follows:

[0035] All rescue routes are retrieved and labeled as i, where i = 1, 2, ..., j, and j represents the type of rescue route. Next, the corresponding rescue missions are retrieved, and the urgency of the mission is used as the primary screening criterion. Specifically, the missions are categorized according to their critical rescue time window: Extremely Urgent: The trapped person's vital signs are critical, requiring arrival within 1 hour (priority level 1); Urgent: The trapped person is not in immediate danger but the environment is deteriorating, requiring arrival within 2-4 hours (priority level 2); General: The trapped person's environment is stable, and arrival can be achieved in more than 4 hours (priority level 3). All valid routes are sorted in ascending order of the urgency level of their corresponding rescue missions (level 1 > level 2 > level 3), and only routes with the same highest priority are retained for the next level of screening.

[0036] If multiple rescue missions have the same level of urgency, the number of trapped personnel is used as a secondary screening criterion, with rescue routes corresponding to missions with more trapped personnel ranked higher. The priority is as follows: Large number (10 or more people), priority A; Medium number (3-9 people), priority B; Small number (1-2 people), priority C. For the remaining routes after the first-level screening, they are sorted in descending order of priority based on the number of trapped personnel in their respective rescue missions (A > B > C). Routes with the highest priority are retained for the next level of screening. If the number of trapped personnel is also the same, the complexity of the terrain and landforms in the scene information is used for screening. This is combined with terrain, obstacle, and weather data from the scene information, and a complexity index is used for quantification. Specific evaluation items include the following:

[0037] Terrain risk: Are there cliffs, canyons, or steep slopes (slope > 30°)? Obstacle density: The number of dynamic obstacles (such as drones, falling rocks) ≥ 3 per square kilometer is considered high density. Meteorological impact: Are there heavy rain, strong winds (wind speed > 12m / s), or low visibility (< 500m)?

[0038] The complexity levels are categorized based on the assessment items: low complexity (priority I): no terrain risk, low obstacle density, and no severe weather; medium complexity (priority II): contains one assessment item risk; and high complexity (priority III): contains two or more assessment item risks. The remaining paths after the secondary screening are sorted in ascending order of terrain complexity (I > II > III). The path ranked first is the optimal rescue path, which is then transmitted to the rescue obstacle avoidance analysis module.

[0039] The rescue obstacle avoidance analysis module is used to perform obstacle avoidance analysis based on the acquired optimal rescue path. It collects data through multi-sensor collaboration, quickly identifies obstacles in the flight path, and distinguishes between static and dynamic types. For static obstacles, LiDAR scans to generate 3D point clouds, which are compared with a pre-loaded digital terrain model (DTED) to identify new static obstacles and mark their location and size. Static obstacle avoidance is then performed based on their location and size, generating static obstacle avoidance information. For dynamic obstacles, millimeter-wave radar tracks moving targets in real time, outputting their speed and heading angle. Visual cameras use the YOLOv10 algorithm to assist in classifying obstacle types. Based on the historical motion data of dynamic obstacles, Kalman filtering or LSTM neural networks are used to predict the trajectory of the helicopter in the next 3-5 seconds, determining whether it will intersect with the helicopter's current path. If there is an intersection, a collision warning is generated; otherwise, normal monitoring information is generated and transmitted to the intelligent rescue information output module.

[0040] For the generated collision warning information, the risk of the detected obstacles is quantified to determine whether obstacle avoidance needs to be initiated. The risk quantification specifically includes the collision time and minimum distance. The specific risk quantification method is as follows:

[0041] Regarding the collision time, when the helicopter and the dynamic obstacle are moving in the same plane, the relative position vector and relative velocity vector are first calculated, and then the relative velocity along the collision direction is calculated through projection. Finally, the collision time TTC is obtained, and TTC = ,in This represents the position vector of the helicopter relative to the obstacle. Let represent the velocity vector of the helicopter relative to the obstacle, and < ,like · ≥0 means that the two objects are moving in opposite directions and there is no possibility of collision. The collision time is then corrected by considering the acceleration of the dynamic obstacle, according to the formula TTC0= The corrected collision time TTC0 is calculated, where The relative acceleration vector is taken as a positive root to ensure that time is positive;

[0042] For the minimum distance, based on the historical motion data of the obstacle, such as its position, speed, and acceleration in the past 1-3 seconds, different models are used to predict its position in the next 5 seconds. Specifically, if the obstacle is moving at a constant speed or with uniform acceleration, a linear motion model is used for prediction; if the obstacle is moving irregularly, an LSTM neural network is used for prediction. The straight-line distance between the predicted position of the helicopter and the predicted position of the obstacle is calculated every 100ms, and after 50 iterations, the minimum value is taken to obtain the minimum distance DCA.

[0043] Next, a collision level analysis is performed by combining the obtained collision time and minimum distance. If the minimum distance is ≤ 50% of the safe distance and the collision time is ≤ 5s, a high-level collision risk signal is generated. If the minimum distance is ≤ 50% - 100% of the safe distance and the collision time is 5s-10s, a medium-level collision risk signal is generated. If the minimum distance is > the safe distance and the collision time is > 10s, a low-level collision risk signal is generated. At the same time, the generated collision analysis signal is transmitted to the obstacle avoidance integrated analysis module.

[0044] The obstacle avoidance comprehensive analysis module is used to acquire collision analysis signals, perform obstacle avoidance processing based on different levels of collision risk signals, generate corresponding obstacle avoidance processing information, and transmit it to the intelligent rescue information output module.

[0045] To generate advanced collision risk signals, non-critical tasks are interrupted, and computing power is prioritized for the obstacle avoidance module. The velocity obstacle method or emergency artificial potential field method is invoked to generate a unique optimal detour path. Obstacle avoidance commands, such as roll angle ≤15° and climb rate ≥3m / s, are sent directly to the flight control system without waiting for pilot confirmation. Adjustments are made based on different scenarios: for low-altitude urban areas, lateral detours are prioritized with a detour distance ≥20 meters; for mountainous / canyon areas, emergency climbs are prioritized to avoid triggering terrain obstacles during lateral detours; for areas at sea, the system prioritizes translation away from obstacles while maintaining altitude and generating advanced obstacle avoidance analysis information.

[0046] For the generated intermediate collision risk signal, 2-3 alternative obstacle avoidance paths are generated, and key information for each path is marked. For example, Path 1: Shortest detour distance (prioritizing rescue timeliness, suitable for emergency missions); Path 2: Lowest risk (no other obstacles in the detour area, suitable for general missions); Path 3: Most stable (gradient change ≤ 5° / second, suitable for medical transport missions, avoiding discomfort to the injured). Select an alternative path within 5 seconds and confirm execution via the cockpit button. If no confirmation is made within the time limit, the system defaults to the lowest risk path. If the rescue mission is urgent, such as the transport of seriously injured persons, the shortest detour path can be selected first, but it is necessary to confirm that there is sufficient remaining fuel. Adjustments are also made according to different scenarios. For severe weather, the system automatically eliminates alternative paths close to complex terrain, retaining only paths in open areas. For multi-obstacle environments, alternative paths must avoid no-fly zones and high-voltage lines, and the location of alternate landing points is marked, generating intermediate obstacle avoidance analysis information.

[0047] For the generated low-level collision risk signal, the TTC / DCA value of the obstacle is updated every 2 seconds. The data is only recorded in the background and does not trigger audible and visual alarms. The AR interface only marks the obstacle position with a small green icon. If the obstacle's movement trajectory changes, the risk is calculated in real time. Once the level 2 or level 1 conditions are met, the signal is immediately upgraded and the corresponding response is triggered, and low-level obstacle avoidance analysis information is generated.

[0048] The intelligent rescue information output module is used to perform corresponding obstacle avoidance actions based on the acquired obstacle avoidance information.

[0049] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0050] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An intelligent auxiliary system for rescue helicopters, characterized in that, include: The rescue path generation module is used to acquire rescue mission and scene information, construct a three-dimensional dynamic spatial model to generate rescue paths, and select the optimal rescue path based on the urgency of the rescue mission, and transmit it to the rescue obstacle avoidance analysis module. The rescue obstacle avoidance analysis module is used to monitor and avoid obstacles on the optimal rescue path. It identifies static or dynamic obstacles through sensors. For static obstacles, it generates static obstacle avoidance information based on their position and size. For dynamic obstacles, it analyzes whether there is an intersection between the dynamic obstacle and the helicopter's trajectory. If there is an intersection, it generates a collision warning; otherwise, it generates normal monitoring information. For collision warning information, the risk is quantified based on collision time and minimum distance, and the collision level is determined by combining the two. Collision analysis signals are generated, including high-level, medium-level and low-level collision risk signals, and transmitted to the obstacle avoidance comprehensive analysis module. The obstacle avoidance comprehensive analysis module is used to analyze collision analysis signals, perform obstacle avoidance processing according to different levels of collision analysis signals, generate obstacle avoidance processing information, and transmit it to the intelligent rescue information output module.

2. The intelligent auxiliary system for rescue helicopters according to claim 1, characterized in that, It also includes a rescue scene information collection module, which is used to collect rescue information in the rescue scene. The rescue information includes rescue mission and scene information. The rescue mission specifically covers key information such as the number of trapped people, their location, and the degree of urgency. The scene information includes terrain, weather conditions, and the distribution of possible obstacles, and is transmitted to the rescue path generation module. The intelligent rescue information output module is used to perform corresponding obstacle avoidance actions based on the acquired obstacle avoidance information.

3. The intelligent auxiliary system for rescue helicopters according to claim 1, characterized in that, The rescue route generation module selects the optimal rescue route in the following way: Based on the acquired scene information, a corresponding three-dimensional dynamic spatial model is constructed, and based on the obtained three-dimensional dynamic spatial model and the rescue location, a corresponding rescue path is generated and labeled as i, where i = 1, 2, ..., j, and j represents the type of rescue path. Obtain the corresponding rescue missions and use the urgency of the rescue missions as the primary filtering condition. Sort all rescue paths in ascending order of the urgency level of their corresponding rescue missions, and only retain paths with the same highest priority to proceed to the next level of filtering. If multiple rescue missions have the same urgency level, then further use the number of trapped people as a secondary filtering condition, sorting them in descending order of the number of trapped people in the corresponding rescue missions, and retaining paths with the same highest priority to proceed to the next level of filtering. If the number of trapped people is also the same, then filter according to the complexity of the terrain and landforms in the scene information, sorting them in ascending order of complexity level, and the path ranked first is the optimal rescue path.

4. The intelligent auxiliary system for rescue helicopters according to claim 3, characterized in that, The urgency of rescue missions is classified according to the golden rescue time window: Level 1: the trapped person's vital signs are critical and they need to be reached within 1 hour; Level 2: the trapped person is not in life danger but the environment is deteriorating and they need to be reached within 2-4 hours; Level 3: the trapped person is in a stable environment and can be reached in more than 4 hours. The analysis method for the complexity of terrain and landform is as follows: low complexity is defined as no terrain risk, low density of obstacles, and no severe weather; medium complexity is defined as having one assessment risk; and high complexity is defined as having two or more assessment risks.

5. The intelligent auxiliary system for rescue helicopters according to claim 1, characterized in that, The rescue obstacle avoidance analysis module monitors and avoids obstacles along the optimal rescue path in the following way: By using multi-sensor collaborative data acquisition, the types of obstacles within the path are identified. For static obstacles, a 3D point cloud is generated by LiDAR scanning and compared with a pre-loaded digital terrain model to identify the location and size of newly added static obstacles and generate static obstacle avoidance information. For dynamic obstacles, millimeter-wave radar is used to track moving targets and output speed and heading angle. Obstacle types are classified by combining visual cameras with the YOLOv10 algorithm. Based on the historical motion data of dynamic obstacles, Kalman filtering or LSTM neural network is used to predict the trajectory of the next 3-5 seconds. It is determined whether the trajectory intersects with the current path of the helicopter. If there is no intersection, normal monitoring information is generated and transmitted to the intelligent rescue information output module. If there is an intersection, collision warning information is generated.

6. The intelligent auxiliary system for rescue helicopters according to claim 1, characterized in that, The method for quantifying risk based on collision time and minimum distance is as follows: Regarding the collision time, when the helicopter and the dynamic obstacle are moving in the same plane, the relative position vector and relative velocity vector are first calculated, and then the relative velocity along the collision direction is calculated through projection. Finally, the collision time TTC is obtained, and TTC = ,in This represents the position vector of the helicopter relative to the obstacle. Let represent the velocity vector of the helicopter relative to the obstacle, and < ,like · ≥0 means that the two objects are moving in opposite directions and there is no possibility of collision. The collision time is then corrected by considering the acceleration of the dynamic obstacle, according to the formula TTC0= The corrected collision time TTC0 is calculated, where The relative acceleration vector is taken as a positive root to ensure that time is positive; For the minimum distance, based on the historical motion data of the obstacle, such as its position, velocity, and acceleration in the past 1-3 seconds, different models are used to predict its position in the next 5 seconds. The straight-line distance between the predicted position of the helicopter and the predicted position of the obstacle is calculated every 100ms. After 50 iterations, the minimum value is taken to obtain the minimum distance DCA.

7. The intelligent auxiliary system for rescue helicopters according to claim 1, characterized in that, The method for generating collision analysis signals is as follows: If the minimum distance is less than or equal to 50% of the safe distance and the collision time is less than or equal to 5 seconds, a high-level collision risk signal is generated. If the minimum distance is less than or equal to 50% - 100% of the safe distance and the collision time is between 5 and 10 seconds, a medium-level collision risk signal is generated. If the minimum distance is greater than the safe distance and the collision time is greater than 10 seconds, a low-level collision risk signal is generated.

8. The intelligent auxiliary system for rescue helicopters according to claim 7, characterized in that, The method for analyzing advanced collision risk signals is to interrupt non-critical tasks, prioritize computing power to the obstacle avoidance module, call the speed obstacle method or emergency artificial potential field method to generate a unique optimal detour path, and directly send obstacle avoidance commands to the flight control system without waiting for pilot confirmation. At the same time, adjustments are made based on different scenarios to generate advanced obstacle avoidance analysis information.

9. The intelligent auxiliary system for rescue helicopters according to claim 7, characterized in that, The method for analyzing intermediate collision risk signals is as follows: generate 2-3 alternative obstacle avoidance paths, mark the key information of each path, select an alternative path within 5 seconds, and confirm execution via the cockpit button; if no confirmation is made within the time limit, the system will default to selecting the lowest risk path. If the rescue mission is urgent, the shortest detour path can be selected first. At the same time, adjustments are made to generate intermediate obstacle avoidance analysis information based on different scenarios.

10. The intelligent auxiliary system for rescue helicopters according to claim 7, characterized in that, The method for analyzing low-level collision risk signals is to update the TTC / DCA value of the obstacle every 2 seconds. The data is only recorded in the background and does not trigger audible and visual alarms. The AR interface only marks the location of the obstacle with a small green icon. If the obstacle's trajectory changes, the risk is calculated in real time. Once the level 2 or level 1 conditions are met, the signal is immediately upgraded and the corresponding response is triggered, and low-level obstacle avoidance analysis information is generated.

Images