Unmanned aerial vehicle intelligent flight path planning method based on flight risk map

Abstract

The invention relates to an unmanned aerial vehicle intelligent flight path planning method based on a flight risk map, and the method comprises the steps: collecting the ground information of a hospital region, obtaining the ground obstacle distribution data, ground channel data and crowded region data in the region, and building a ground map according to the data; building information of a hospital area is collected, building distribution data, building floor height, building spacing data and air channel data in the area are obtained, and an air risk map is established according to the data; after a transportation order is obtained, a delivery area where a destination is located is identified, a landing point position of the delivery area is obtained, and a safest shortest path is planned in combination with ground channel data and air channel data. According to the scheme, the transportation requirements of multiple buildings of a hospital are fully covered, a risk map navigation system is provided, the logistics safety of the unmanned aerial vehicle is greatly improved, and the probability of major accidents caused by out-of-control crash of the unmanned aerial vehicle is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) intelligent delivery, and in particular to a UAV intelligent flight path planning method based on a flight risk map. Background Art

[0002] Hospital goods transportation requires high accuracy, speed and safety. In order to solve this problem, in addition to traditional manual transportation, hospital goods transportation mainly adopts medical pneumatic tube logistics transmission system and rail logistics transmission system. The disadvantages of pneumatic tube logistics transmission are that the construction of cross-building transmission system is difficult, the cost is high, and the damage rate of transported goods is high. The main disadvantage of rail logistics transmission is that the construction of cross-floor transmission track is difficult and the transmission efficiency is low. The above two methods have high construction costs, especially when there is no reserved transmission channel in old buildings, the construction requires the transformation of the building.

[0003] With the popularization and promotion of drone technology, using drones to transport materials is becoming more and more popular.

[0004] To realize drone transportation in hospital areas, the reliability of the logistics system must be guaranteed as much as possible and the possibility of major accidents must be reduced. Existing transport drones cannot meet the special requirements of hospital delivery. Summary of the invention

[0005] Based on the above statements, the present invention provides a method for intelligent flight path planning of UAVs based on flight risk maps, which can reduce the risk of UAV transportation in hospital buildings.

[0006] The technical solution of the present invention to solve the above technical problems is as follows:

[0007] The intelligent flight path planning method of UAV based on the flight risk map collects ground information of the hospital area, obtains the distribution data of ground obstacles in the area and the data of crowded areas;

[0008] Collect building information in the hospital area, obtain data on building distribution, building height, building spacing and air channel data in the area, and establish a flight risk map based on ground information;

[0009] Calibrate the landing point locations of each material delivery area and the take-off point locations of the drones;

[0010] After obtaining the transport order, identify the delivery area where the destination is located and obtain the landing point location of the delivery area, set the take-off point location as the departure location and set the obtained landing point location as the end location, and plan the safest and fastest route in combination with the air channel data and flight risk map.

[0011] As a preferred solution: based on building map data and weather data, calibrate the flight risk level and safe speed of each air channel affected by weather factors, including rain, snow and wind; automatically plan at least two initial routes to be selected when planning the route, and then conduct real-time detection of the weather in the area where the channels constituting each initial route are located, and calculate the average risk level and estimated flight time of each initial route in combination with the flight risk map, and determine the recommended value of each initial route based on this, and set the initial route with the highest recommended value as the final flight route.

[0012] As a preferred solution: real-time detection of the crowd density and overall flow direction in the area below each air channel, based on which the risk correction factor of each air channel is determined, and the risk factor is used to correct the flight risk level of each air channel.

[0013] As a preferred solution: when planning the initial route, after obtaining the locations of the take-off point and the landing point, connect the take-off point and the landing point; plan the shortest path, and based on the shortest path, move the intersection point of the flight path and the ground pedestrian path, and move the intersection point in two directions until it is avoided. If it cannot be avoided, choose the shortest crossing route.

[0014] As a preferred solution: the distance of the moving intersection is positively correlated with the non-ground human flow.

[0015] As a preferred solution: the route recommendation value calculation formula D = a × N + b × M, where N represents the average risk level, M represents the expected flight time, and a and b are preset calculation coefficients.

[0016] As a preferred solution: when planning the initial route, after obtaining the positions of the take-off point and the landing point, connect the take-off point and the landing point; screen out the points where the connection line passes through and out of the building, and then translate the two adjacent points outward for a certain distance to obtain two turning path points. Connect the take-off point, each turning path point and the landing point and perform smoothing to obtain a continuous and smooth initial path.

[0017] Compared with the existing technology, the technical solution of this application has the following beneficial technical effects: This application innovatively proposes an application solution for low-altitude drones in hospital smart logistics. The solution proposes a flight risk map, which can be used to intelligently plan the flight path of drones in real time, meet the transportation needs between multiple buildings in the hospital campus, and greatly improve the safety and efficiency of logistics. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a schematic diagram of a ground map in this embodiment;

[0019] Figure 2 is a schematic diagram of an aerial map in this embodiment;

[0020] Figure 3 Schematic diagram of the weather map in this embodiment;

[0021] Figure 4 This is a schematic diagram of line planning in this embodiment. DETAILED DESCRIPTION

[0022] The intelligent flight path planning method of UAV based on flight risk map is as follows:

[0023] Collect ground information of the hospital area, obtain ground obstacle distribution data and crowded area data in the area. The ground obstacles in this embodiment include ground occlusion, ground sidewalks, avoidance points and other elements. Figure 1 As shown, ground obstacles such as multiple buildings marked in dark colors, such as Building A, Building B, Building C, etc., corridors with traffic, green belts, etc., form a ground map through ground detection points.

[0024] Collect building information of the hospital area, obtain building distribution data in the area, building height, building spacing and air channel data, and establish a flight risk map based on ground information. Figure 2 A schematic diagram of the flight risk map, where large search areas indicate areas that are not flyable, are not within the planning, or are beyond the signal range.

[0025] The drone can avoid aerial obstacles in the flight area, and intelligently plan detour routes along the high-altitude protective wall skirt design unique to hospital buildings and the natural barrier of the green isolation belt below to ensure the safety of personnel. Figure 2 The lighter gray areas are safe flight areas, and the darker gray areas are air obstacles. Figure 1 Buildings in. Figure 1 , it is found that the thicker dark curves are blocked pedestrian routes, and the thinner dark curves are unblocked pedestrian routes. When the drone is flying, it will dynamically divide its flight route according to this map to avoid crowds and obstacles.

[0026] Calibrate the landing point locations of each material delivery area and the take-off point locations of the drones;

[0027] After obtaining the transport order, identify the delivery area where the destination is located and obtain the landing point location of the delivery area, set the take-off point location as the departure location and set the obtained landing point location as the end location, and plan the safest and fastest route in combination with the air channel data and flight risk map.

[0028] like Figure 4As shown in the figure, the transportation target is from Building A to Building C. The system plans the first route, which is the shortest route, i.e. the middle route. If the route intersects with the pedestrian path on the ground twice, the system considers the risk to be high and tries to plan the second route. With the first intersection point, the system shifts downward and quickly finds the first route that staggers the intersection point, forming the lower route.

[0029] However, the lower road still requires two intersection points. After traversing the lower road, there is no route better than the shortest route, so the system chooses the upper road. After traversing the upper road, the system plans the upper road, which has the smallest angle with the shortest route and has only one intersection point, so it is the current optimal route.

[0030] In this flight planning, the weather conditions were relatively uniform and met safety requirements, so they were no longer calculated.

[0031] like Figure 3 As shown, in this embodiment, based on the building map data and weather data, the flight risk level and safe speed of each section of each air channel affected by weather factors are calibrated, including rain, snow and wind. Figure 3 The ellipses in the figure represent the flight risk levels of various areas. The darker the ellipse, the higher the risk level.

[0032] Numbers are used to represent the flight risk level, and the flight risk level is assigned "1", "2", "3", and "4" in order from low to high. There are four levels in total, representing "safe", "low risk level", "general risk level", and "high risk level" respectively.

[0033] Through modeling analysis, we can simulate the visibility, obstacle avoidance sensitivity, air velocity, turbulence distribution and intensity of each air channel under different weather conditions, and calibrate the wind risk level of each air channel under different weather conditions based on the simulated data.

[0034] When planning a route, at least two initial routes are automatically planned for selection. Then, the weather in the channel that constitutes each initial route is detected in real time. Combined with the flight risk map, the average risk level and estimated flight duration of each initial route are calculated. Based on this, the recommended value of each initial route is determined, and the initial route with the highest recommended value is set as the final flight route.

[0035] In this embodiment: the route recommendation value calculation formula D=a×N+b×M, where N represents the average risk level, M represents the expected flight time, and a and b are preset calculation coefficients.

[0036] The process of obtaining the average risk level is as follows: after the initial route is planned, the air channel of the initial route is divided into multiple sections, the weather data of each area is detected in real time, and the flight risk level corresponding to the area under the current weather is retrieved to obtain the flight risk level data of each section of the air channel. After that, these data are averaged to obtain the average risk level N of the initial route. M is obtained by accumulating the estimated flight time of each section. The calculation coefficients a and b can be defined and modified according to the data obtained from a large number of field simulations to meet the actual situation of the hospital.

[0037] In this embodiment: the density of human traffic and the overall flow direction of the area below each aerial channel are detected in real time, and the risk correction coefficient of each aerial channel is determined accordingly, and the flight risk level of each aerial channel is corrected using the risk coefficient. Specifically, the area below each aerial channel is divided into multiple sections, and a monitoring camera is installed at each section. By shooting the video image in the section and tracking and identifying the personnel in the image, the density of human traffic in the section can be detected and the overall flow direction of the personnel can be analyzed. Regarding the definition of the overall flow direction of personnel: the flight direction of the drone in the channel is used as the baseline, and the horizontal direction in the channel is evenly divided into 8 direction angle ranges according to 45 degrees. Identify and count the walking direction of each person in the section and determine which direction angle range it falls within. After a period of statistics, the number of walking personnel falling within each angle range is compared, and the angle range with the largest number of walking personnel is selected, and the center line direction of the angle range is defined as the overall flow direction of personnel. The angle between the center line and the baseline is α. The correction coefficient k of different angles α is calibrated in advance. When the angle α is determined, the corresponding correction coefficient k is adjusted, and the correction coefficient is multiplied by the retrieved flight risk level to obtain the corrected flight risk level. When calculating the average risk level, it is necessary to first correct the flight risk level of each section of the air channel. After correction, the data is brought into the recommended value calculation formula, so that the recommended values ​​of each initial route can be in line with the current actual status of the hospital, which is more scientific and reasonable.

[0038] In this embodiment: when planning the initial route, after obtaining the locations of the take-off point and the landing point, connect the take-off point and the landing point; plan the shortest path, and based on the shortest path, move the intersection point of the flight path and the ground flow, and move the intersection point in two directions until the intersection point is avoided. If it cannot be avoided, choose the shortest crossing route. It should be noted that the distance of the moving intersection point is positively correlated with the ground flow of people.

[0039] In this embodiment: the density of human traffic and the overall flow direction of the area below each aerial passage are detected in real time, and the risk correction coefficient of each aerial passage is determined accordingly, and the flight risk level of each aerial passage is corrected using the risk coefficient.

[0040] In this embodiment: In order to reduce the risk of drones flying over buildings, when planning the initial route, after obtaining the positions of the take-off point and the landing point, the take-off point and the landing point are connected; the points where the connection line enters and exits the building are screened out, and two adjacent points are translated outward for a certain distance to obtain two turning path points. The take-off point, each turning path point and the landing point are connected and smoothed to obtain a continuous and smooth initial path.

[0041] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent flight path planning method for UAV based on flight risk map, characterized by: Collect ground information of the hospital area, obtain data on the distribution of ground obstacles and crowded areas in the area; Collect building information in the hospital area, obtain data on building distribution, building height, building spacing and air channel data in the area, and establish a flight risk map based on ground information; Calibrate the landing point locations of each material delivery area and the take-off point locations of the drones; After obtaining the transport order, identify the delivery area where the destination is located and obtain the landing point location of the delivery area, set the take-off point location as the departure location and set the obtained landing point location as the end location, and plan the safest and fastest route in combination with the air channel data and flight risk map.

2. The method for intelligent flight path planning of unmanned aerial vehicle based on flight risk map according to claim 1 is characterized in that: Based on building map data and weather data, the flight risk level and safe speed of each air channel affected by weather factors are calibrated, including rain, snow and wind; when planning the route, at least two initial routes are automatically planned for selection, and then the weather of the channels constituting each initial route is detected in real time, and combined with the flight risk map, the average risk level and estimated flight time of each initial route are calculated, and the recommended value of each initial route is determined accordingly, and the initial route with the highest recommended value is set as the final flight route.

3. The method for intelligent flight path planning of unmanned aerial vehicle based on flight risk map according to claim 2 is characterized in that: The crowd density and overall flow direction of the area below each air channel are detected in real time, and the risk correction factor of each air channel is determined accordingly. The risk factor is used to correct the flight risk level of each air channel.

4. The method for intelligent flight path planning of unmanned aerial vehicle based on flight risk map according to claim 3 is characterized by: When planning the initial route, after obtaining the locations of the take-off point and the landing point, connect the take-off point and the landing point; plan the shortest path, and based on the shortest path, move the intersection point of the flight path and the ground pedestrian path, and move the intersection point in two directions until it is avoided. If it cannot be avoided, choose the shortest crossing route.

5. The method for intelligent flight path planning of unmanned aerial vehicle based on flight risk map according to claim 4 is characterized in that: The distance of the mobile intersection is positively correlated with the non-ground pedestrian flow.

6. The method for intelligent flight path planning of unmanned aerial vehicle based on flight risk map according to claim 2 is characterized by: The formula for calculating the recommended route value is D=a×N+b×M, where N represents the average risk level, M represents the expected flight time, and a and b are preset calculation coefficients.

7. The method for intelligent flight path planning of unmanned aerial vehicle based on flight risk map according to claim 2 is characterized by: When planning the initial route, after obtaining the positions of the take-off point and the landing point, connect the take-off point and the landing point; select the points where the connection line passes through and out of the building, and then translate the two adjacent points outward for a certain distance to obtain two turning path points. Connect the take-off point, each turning path point and the landing point and perform smoothing to obtain a continuous and smooth initial path.

Images