Computer system and method for providing wildfire evacuation support

Abstract

The present disclosure relates to computer systems and methods for providing wildfire evacuation support. A computer system is included in a system for providing wildfire evacuation support. The computer system is configured to obtain (402) sensor data representing a wildfire in a target area from an unmanned aerial vehicle deployed in the target area, obtain (403) a current location of a person in a dangerous situation in the target area, and obtain (404) a desired destination for the person in the target area. Based on the sensor data, the computer device identifies (405) one or more danger zones that pose a fire-related threat to the person, and determines (406) at least one evacuation path that extends from the current location to the desired destination while avoiding the one or more danger zones. The computer system can be located on the unmanned aerial vehicle or on a central computer resource in communication with the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention relates generally to providing evacuation support to personnel at the scene of a wildfire, and more particularly to using unmanned aerial vehicles for this purpose. Background Art

[0002] Wildfires are one of the most common natural disasters in places like Siberia, California, and Australia. Regions with Mediterranean climates or within coniferous forest biomes are particularly vulnerable. With the increasing impacts of climate change, wildfires are expected to become more prevalent in these and other regions around the world. Wildfires involve intense burning, producing high heat, smoke, and gases, which can damage property and people. People trapped in wildfires may have little chance of escaping, as the fire's extent and growth are often difficult to detect.

[0003] While there are systems in place to detect and monitor wildfires and to guide firefighters into them, there are no systems to safely guide personnel away from them.

[0004] Therefore, there is a general need to support the evacuation of persons trapped in wildfires. Preferably, such evacuation support should not require the use of wireless electronic devices carried by the persons, not only because wireless communication may be disrupted in or around the wildfire, but also because persons without wireless electronic devices would be unable to obtain appropriate evacuation support. Summary of the Invention

[0005] It is an object of the invention to at least partially overcome one or more limitations of the prior art.

[0006] Another object of the invention is to provide technology that can safely guide people (optionally by car) away from wildfires.

[0007] Yet another object of the invention is to provide technology that can guide anyone and everyone away from a wildfire.

[0008] One or more of these objects, as well as other objects that may appear from the description below, are achieved at least in part by a computer system for providing wildfire evacuation support.

[0009] Other objects, features, aspects and technical effects of the invention will appear from the following detailed description and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 is a perspective view of the target area including the wildfire.

[0011] Figures 2A to 2B Included are a side view and block diagram of an example unmanned aerial vehicle used in a wildfire evacuation support system.

[0012] Figure 3is a schematic diagram of an example computer system used in a wildfire evacuation support system.

[0013] Figure 4 is a flowchart of an example approach for wildfire evacuation support.

[0014] Figure 5 An example configuration space is shown, which is related to Figure 1 , and is defined for a path planning algorithm that can be used to determine evacuation paths.

[0015] Figure 6 An example of a configuration space with obstacles, nodes, and connections generated for a target area during operation of the probabilistic roadmap algorithm is shown.

[0016] 7A to 7B An example calculation of the danger zone margin is illustrated.

[0017] Figures 8A to 8C An example of data transmission in a wildfire evacuation support system including one or more unmanned aerial vehicles is shown.

[0018] Figure 9 yes Figure 3 A block diagram of an example computer system in . DETAILED DESCRIPTION

[0019] The embodiments will now be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all, embodiments are shown. Indeed, the subject matter of the present disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure may satisfy applicable legal requirements.

[0020] In addition, it should be understood that, where possible, any advantages, features, functions, devices and / or operational aspects of any embodiment described and / or envisioned herein may be included in any other embodiment described and / or envisioned herein, and / or vice versa. In addition, unless otherwise expressly stated, where possible, any term expressed in the singular herein also includes the plural form and / or vice versa. As used herein, "at least one" should refer to "one or more", and these terms are intended to be interchangeable. Therefore, the term "a and / or an" should refer to "at least one" or "one or more", even if the phrase "one or more" or "at least one" is also used herein. As used herein, unless the context requires otherwise due to the language of expression or necessary meaning, the word "comprising" or variations such as "having" or "containing" are used in an inclusive sense, that is, specifying the presence of the features described, but not excluding the presence or addition of other features in various embodiments. The term "computing" and its derivatives are used in their conventional sense and can be considered to involve performing calculations involving one or more mathematical operations to produce a result, such as by using a computer.

[0021] As used herein, the terms "plurality" and "plurality" mean that two or more items are provided, while the term "set" means that one or more items are provided. The term "and / or" includes any and all combinations of one or more of the associated listed items.

[0022] In addition, it should be understood that although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element without departing from the scope of the present invention.

[0023] For brevity and / or clarity, well-known functions or constructions may not be described in detail. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0024] Like reference numerals refer to like elements throughout.

[0025] Before describing the embodiments in more detail, some definitions will be given.

[0026] As used herein, a "wildfire" is an unexpected, unwanted, or uncontrolled fire that occurs in an area of ​​flammable vegetation. Wildfires can affect rural and / or urban areas. A wildfire may consist of one or more "confined space fires." A confined space fire is a coherent area of ​​flaming and / or smoldering fire that is physically isolated from other confined space fires (if any).

[0027] As used herein, an "unmanned aerial vehicle" (UAV) is an aircraft that is controlled by an onboard automatic control system, a ground control system, or a ground pilot. Such aircraft are also known as "uncrewed aircraft" or drones.

[0028] Will refer to Figure 1 Example implementation, Figure 1 A plurality of drones 1 are depicted operating in a geographic area 2 (the "Target Area") containing a wildfire 3. In the example shown, the wildfire 3 comprises a plurality of confined space fires F. It should be understood that the nature of the wildfire 3 may change over time in terms of extent and intensity. For example, new confined space fires F may emerge, a confined space fire may be split, or two confined space fires may merge, etc. Thus, Figure 1 The example in is a snapshot of wildfire 3 causing damage to area 2.

[0029] Anyone located within or near Wildfire 3 is at risk of injury or even death from the heat and / or gases generated by Wildfire 3 or secondary risks (e.g., explosions, falling trees, etc.). Figure 1 The current location P1 of a person in a dangerous situation due to being trapped within a wildfire 3 is depicted. As discussed in the background section, in such situations, the person often cannot fully grasp the extent of the wildfire 3 or find a safe way out, such as to reach a desired destination P2. As disclosed herein, a drone 1 can be used to gather information about the wildfire 3, and possibly other relevant information, thereby enabling a computer system to determine one or more evacuation routes for the person in danger. The computer system can also provide evacuation guidance to the person, for example, via drone 1.

[0030] The following references to individuals also apply to a group of individuals. Thus, an evacuation route can be determined for a group of individuals, and evacuation guidance can be provided to such a group of individuals collectively.

[0031] Figure 2A is a side view of an example drone 1 that can be used for this purpose, Figure 2Bis a schematic block diagram of such a drone 1. The drone 1 can have a conventional structure, but can include a sensor system customized for the mission of the drone 1 (see figure below). The drone 1 includes a control system 10, which is a circuit for controlling the overall operation of the drone 1. The drone 20 also includes a conventional propulsion system 11, which is used to generate thrust to move the drone 1 through the air. The drone also includes a positioning system 12 for determining the location of the drone 1. In some embodiments, the positioning system 12 includes a global navigation satellite system (GNSS) receiver. The GNSS receiver can be configured to communicate with a satellite navigation system (such as GPS, GLONASS, Galileo, or BeiDou). The positioning system 12 can also include an altimeter. The control system 10 is configured to receive position data from the positioning system 12, and optionally receive additional data from various conventional sensors on the drone 1, and control the propulsion system 11 to move the drone along a flight path, which may or may not be predetermined. For example, the control system 10 can include a conventional flight controller.

[0032] The control system 10 is also connected to a communication system 13 for wireless data transmission. In some embodiments, the communication system 13 includes a transceiver for short-range wireless communication, such as short-range wireless communication via one or more of Bluetooth, BLE, WiFi, LR-WPAN, UWB, ZigBee, Z-Wave, ANT, ANT+, 6LoWPAN, WirelessHART, ISA100, etc. Short-range communication can be used for data communication with other drones 1 (see Figure 8C ). Alternatively or additionally, the communication system 13 may include a transmitter or transceiver for long-range wireless communication. Such long-range communication may utilize any available proprietary or standardized communication technology, including but not limited to GSM, EDGE, HSDPA, W-CDMA, CDMA, TDMA, LTE, 5G, Wi-MAX, LORAN, etc. Alternatively or additionally, the long-range communication may involve satellite communication. Long-range communication may be used for data communication with a remote computer (see Figures 8A to 8B ).

[0033] The control system 10 is also connected to a sensor system 14, which includes one or more sensors capable of generating data representative of the wildfire 3. In some embodiments, the sensor system 14 includes one or more sensors for remotely monitoring heat, temperature, or topography beneath the drone 1 and, therefore, within at least a portion of the wildfire 3. Such sensors can be used to generate a 1D, 2D, or 3D representation ("image") of the terrain beneath the drone 1 in one or more spectral regions. Figure 2AA drone 1 is shown with a camera 14', which can be movable or fixed. The field of view of the camera 14' is indicated by a dotted line. The camera 14' can respond to radiation in at least one of the visible light and infrared spectral regions. The camera 14' can be an optical camera, a thermal imaging camera, a multispectral camera, a hyperspectral camera, a stereo camera, etc. The sensor system 14 may include multiple cameras, such as cameras configured for different spectral regions. Alternatively or additionally, the sensor system 14 may include a radar-based system, such as a SAR system (synthetic aperture radar) and / or a lidar system (light detection and ranging). Other sensors may also be included in the sensor system 14, such as an air temperature sensor, an air pressure sensor, a smoke or gas detector, a wind speed sensor, a wind direction sensor, etc.

[0034] The control system 10 is also connected to a guidance system 15, which can be used to provide cognitive guidance to personnel, for example, by using light or sound. In some embodiments, the guidance system 15 includes one or more speakers for generating audio instructions for personnel. In some embodiments, the guidance system 15 includes one or more operable emitters for projecting light onto the ground to generate a guiding light pattern.

[0035] The components of the drone 1 are powered by a power source 16 , which may include one or more of a battery, a fuel cell, an energy harvesting unit, a fuel tank, and the like.

[0036] Figure 3 is a block diagram of an example computer system 20 configured to determine and output one or more evacuation routes EP for persons believed to be in a dangerous situation related to a wildfire. The computer system 20 includes one or more input interfaces 21A, 21B for receiving input data and an output interface 22 for providing output data. The interfaces 21A, 21B, 22 can be of any type. The input interfaces 21A, 21B generally define the "input end" of the computer system 20. As shown in the figure, the computer system 20 is configured to receive sensor data SD generated by one or more drones 1 to represent an area 2 ( Figure 1 ) in a wildfire 3. The computer system 1 can also receive the position data P1 of the person (see Figure 1 ) and the personnel's desired destination location data P2 (see Figure 1 As described below, computer system 20 may alternatively derive position data P1 and / or position data P2 from sensor data SD. Furthermore, as shown, computer system 20 may receive metadata MD associated with region 2. Examples of metadata MD and its use are provided below.

[0037] In some embodiments, the computer system 20 is implemented on a computer resource remote from the drone 1 and the area 2, such as on a server or cloud computer. In other embodiments, the computer system 20 is implemented on one or more drones 1, such as as part of the control system 10. Figures 8A to 8C Describes an implementation example.

[0038] Figure 4 is a flow chart of an example method 400 for providing evacuation guidance associated with wildfires using one or more drones. The method 400 may be Figure 3 The computer system 20 in the embodiment is implemented and executed by the computer. The optional steps and processes are as follows: Figure 4 The dotted line in the figure indicates that Figure 1 The scenario in describes this approach.

[0039] In optional step 401, the metadata MD is obtained. For example, the metadata MD may represent the topography of area 2, current and future weather conditions in area 2, buildings or other facilities in area 2, roads and sidewalks in area 2, storage locations of hazardous materials, and the types of hazardous materials in the respective storage locations. The metadata MD may be obtained by computer system 20 from an external source, such as one or more remote computers. Alternatively, at least a portion of the metadata MD may be pre-stored in a memory unit accessible by computer system 20.

[0040] In step 402, sensor data SD is obtained from one or more drones 1 being deployed in the target area 2. The sensor data SD is generated by the respective drones 1 to at least represent a wildfire. In some embodiments, the sensor data SD represents the location and extent of the wildfire 3 corresponding to the confined space fire F. Such sensor data will allow the identification of the danger zone in the following step 405. In some embodiments, the sensor data SD further represents the intensity F of the one or more confined space fires. In this case, "intensity" refers to the heat released by the fire. In step 406, the fire intensity can be used to determine a safe evacuation path for personnel. The intensity of the fire can be estimated based on an infrared image of the fire. Based on the sensor system 14 ( Figure 2B ), the sensor data SD may represent additional information, such as the location of the person, the type of transportation available to the corresponding person, the topography of the target area 2, the weather conditions of the target area 2, etc.

[0041] Sensor data SD may include raw data and / or preprocessed data generated by sensor system 14. In some embodiments, sensor data SD represents real-time images generated by sensor system 14. As described above, such images may be 1D, 2D, or 3D representations of the terrain below drone 1. In this context, "real-time" means that sensor data SD is acquired via step 402 within a short period of time (e.g., less than 1-5 minutes) after the image is detected by sensor system 14. In some embodiments, sensor data SD represents an infrared image containing the temperature distribution within target area 2 or a portion thereof. Note that sensor data SD need not include raw images, but may include preprocessed images or data extracted from the images through preprocessing.

[0042] In step 403, the current position P1 of the person in the target area 2 is obtained. The current position P1 can be provided to the computer system 20 as input data separate from the sensor data, such as Figure 3 As shown. For example, the current position P1 can be provided by a wireless signal from a transponder or other electronic device carried by the person or installed in a vehicle used by the person. Such a signal can be intercepted by the drone 1 or other device, and the corresponding current position P1 can be provided as input data to the computer system 20. Alternatively, the current position P1 can be included in the sensor data SD. For example, the current position P1 can be determined through the above-mentioned preprocessing and provided as part of the sensor data SD.

[0043] In some embodiments of step 403, the computer system 20 detects the person and determines the current position P1 of the person based on the sensor data SD. An advantage of such an embodiment is that the detection of the person in distress is independent of the transponder and external systems. For example, the computer system 20 can process one or more images included in the sensor data SD, for example by using conventional image processing to detect the person in each image, and can map the position of the person in each image to a physical position in the coordinate system of the target area 2, for example in the GNSS coordinate system. The mapping can be based on the position and orientation of the drone 1 when the corresponding image was captured, and optionally the camera 14' ( Figure 2A ) direction, and possible terrain of target area 2.

[0044] In step 404, the personnel's desired destination P2 is obtained. The desired destination P2 may be predefined and included in metadata MD. Alternatively, the firefighter at target area 2 may input the desired destination P2 into an external system, which provides the destination as input data to computer system 20. In another alternative, the desired destination P2 is determined by processing one or more images captured by sensor system 14 on drone 1. It is also contemplated that computer system 20 may determine the desired destination P2 based on the configuration space defined in step 406A (below).

[0045] In step 405, one or more danger zones are identified in the target area 2 based on the sensor data SD. The corresponding danger zones are considered to pose a fire-related threat to personnel. The danger zones can be detected by the above-mentioned pre-processing and included in the sensor data SD received by the computer system 20. Alternatively, the computer system 20 can identify or detect the danger zones by processing the sensor data SD, for example, by processing one or more images included in the sensor data SD. An example of a danger zone is a fire, such as Figure 1 Any confined space fire F in the image can be detected by conventional image processing (e.g., infrared image). The location and extent of the fire can be detected in an image (e.g., infrared image). Another example of a danger zone is a building that could explode and / or spread hazardous materials due to wildfire. Such a danger zone can be indicated in the metadata MD. Alternatively, it can be detected in the image and optionally combined with information about the type of building, for example, if it contains explosives and / or hazardous materials. The latter information can be provided by the metadata MD. Examples of buildings that may pose a fire-related threat include gas stations, power plants, industrial plants, etc.

[0046] In step 406, the computer system 20 determines at least one evacuation path EP extending from the current position P1 to the desired destination P2 while avoiding the danger zone identified in step 405. It is conceivable that the computer system 20 determines a plurality of evacuation paths, optionally sorted by preference, for example according to risk to the personnel.

[0047] In step 407, the computer system 20 guides the individual from their current location P1 based on the evacuation path determined in step 406. In some embodiments, guiding the individual is accomplished using one or more drones 1. For example, the control system 20 may cause a drone 1 to activate its guidance system 15 to indicate the individual's evacuation path. Alternatively or additionally, the control system 20 may cause one or more drones 1 to fly at a low altitude to guide the individual along the evacuation path. It is also conceivable that multiple drones 1 may operate in formation to guide the individual along the evacuation path. Step 407 does not require the use of drones 1 to guide the individual. In an alternative embodiment, the computer system 20 transmits information about the evacuation path to an electronic device carried by the individual or installed in a vehicle used by the individual. The vehicle may be manually operated by the individual or may be an autonomous vehicle, such as a self-driving car, that automatically configures itself to follow the evacuation path received from the computer system 20. As used herein, "guiding the individual" also includes the vehicle in which the individual is guided.

[0048] Method 400 can be executed once to determine one or more evacuation paths from a current location P1 to a desired destination P2, so that a computer system or another system can use the evacuation paths to guide a person from the current location P1 to the desired destination P2. By determining multiple evacuation paths, if conditions in the target area change as the person is guided to the desired destination P2, switching between evacuation paths can be performed.

[0049] In some embodiments, as Figure 4 As shown by the dashed arrows in , as the person moves through the target area 2, step 406 and at least some of steps 402 to 405 are repeated to enable the evacuation path to be updated to account for changes over time. Steps 402 and 405 can be repeated to determine an evacuation path via step 406 based on changes in the hazard zone in the target area, changes in the type of transportation (for example, if a vehicle breaks down), etc. Step 403 can be repeated to determine an evacuation path for the person's new current location P1 via step 406. Thus, the method 400 will track the movement of the person and repeatedly determine an evacuation path associated with the person's current location. Step 404 can be repeated to determine an evacuation path for the updated desired destination P2 via step 406. Thus, the method 400 will adapt to changes in the wildfire 3 that require a change in destination.

[0050] In some embodiments, as Figure 4As shown by the dashed lines in , method 400 includes step 406A of representing each of the one or more hazard zones as an obstacle in a configuration space of a path planning algorithm (PPA), and step 406B of operating the PPA to determine an evacuation path. The term "path planning" refers to the computational problem of calculating a continuous path connecting a starting configuration and a target configuration while avoiding collisions with known obstacles. Path planning is also known as motion planning, navigation problem, or piano mover problem. Traditionally, PPA operates on a configuration space, which is defined as a set of configurations that avoid collisions with obstacles (denoted as "free space"), and the complement of the free space (denoted as "obstacle region"). In steps 406A to 406B, any suitable PPA can be used to determine the evacuation path, such as a grid-based search algorithm, an interval-based search algorithm, a geometric algorithm, an artificial potential field algorithm, or a sampling-based algorithm. Commonly used PPAs include rapidly exploring random trees, probabilistic roadmaps, and their variations.

[0051] Figure 5 It is for Figure 1 Schematic diagram of an example configuration space 30 defined for a target area 2 in FIG. 1 and also indicating an evacuation path EP that has been determined for the configuration space 30. The configuration space 30 corresponds to the target area 2 and includes a plurality of obstacles represented by dashed lines. Some obstacles may correspond to physical obstacles in the target area 2, such as mountains, impenetrable terrain, buildings, etc. For example, Figure 5 The rectangular obstacle 31' in corresponds to Figure 1 Other obstacles may correspond to the danger zones mentioned above. Figure 5 In the figure, the danger zone corresponding to the fire is indicated by reference numeral 32, and the danger zone corresponding to the hazardous material storage facility is indicated by reference numeral 33. The corresponding obstacles in the configuration space 30 are indicated by dotted lines and are indicated by reference numerals 32' and 33', respectively.

[0052] Although obstacles can be defined to have the same location and extent as danger zones, Figure 5 As shown, it may be advantageous to define obstacles as having a margin from the danger zone. This margin may correspond to a "safe distance" from the danger zone and will ensure the safety of people along the evacuation path.

[0053] For a danger zone 32 representing a fire, a margin may be set to ensure that personnel are not exposed to excessive heat at the boundary of the obstacle 32'. In one example, the margin may increase as the intensity of the fire increases. In another example, the margin may be set based on the spread rate and / or direction of the fire in the corresponding confined space. The intensity, spread rate, and spread direction are characteristics of the wildfire 3 and may be determined by processing one or more images captured by the sensor system 14. In another example, the margin may be set based on wind data such as wind speed and / or wind direction and, optionally, terrain. As described above, the sensor system 14 may be configured to measure wind speed / direction. Alternatively, the wind speed / direction may be included in the metadata MD. Wind speed / direction may also be used when defining obstacles in the configuration space to ensure that personnel are not exposed to unhealthy or lethal smoke and gaseous substances at the boundary of the obstacle. Figure 5 The wind direction W is schematically shown. Figure 5 As shown, the margin extends in the wind direction W by comparing the obstacle 32' with its hazard zone 32. It can also be noted that the wind speed is expected to be lower behind the mountain range represented by the large rectangular object 31, resulting in a smaller extension of the margin in the wind direction for the hazard zone 32 located at the foot of the mountain.

[0054] For hazardous areas that do not represent an actual fire, but rather a risk of explosion and / or release of hazardous substances in the event of a fire, a margin can be set to ensure that during the evacuation of personnel, personnel will not be harmed by explosions or releases of hazardous substances at the boundaries of the barrier. Figure 5 In FIG, the danger zone 33 is a gas station, and the corresponding obstacle 33 ′ has equal margins in all directions of the danger zone 33, although the margins can also be set according to wind speed and / or wind direction, terrain, etc. Figure 4 ,It should be noted that since the characteristics of Wildfire 3 may change over time, the margin between ,repeated steps 406 may be different.

[0055] Figure 6 An example of a probabilistic roadmap generated for another configuration space is shown, which includes a plurality of obstacles 31' corresponding to physical objects, a plurality of obstacles 32' corresponding to fires, and a plurality of obstacles 33' corresponding to explosion and / or emission risks. Figure 6 In the figure, obstacles 31', 32', 33' are defined as polygonal objects, but can be of any conceivable shape, e.g. Figure 5 The circle or oval in the middle. Figure 6The route map in has been calculated by the PPA in step 406B. The route map is represented as a graph of points (nodes) connected by lines. These lines correspond to sub-paths that people can traverse in the target area. As shown in the figure, all points and lines are located outside obstacles 31', 32', 33'. A weight can be assigned to each route, for example, depending on the terrain along the route, the proximity of obstacles, traffic patterns (see below), etc. In step 406B, the route map can be queried by any suitable search algorithm to determine one or more evacuation paths from the current position P1 to the desired destination P2. The search algorithm can be configured to find an evacuation path that meets one or more goals, such as minimizing the distance or transition time from the current position P1 to the desired destination P2, or optimizing personnel safety. Non-limiting examples of search algorithms for finding the shortest path include Dijkstra's algorithm, Bellman-Ford algorithm, and Floyd-Warshall algorithm.

[0056] Back to Figure 4 In step 406, the computer system 20 may also describe the mode of transportation that will be used when the person moves from the current location P1 to the desired destination P2. The mode of transportation ("traffic type") may be given by the sensor data SD. For example, the computer system 20 may determine the traffic type by image processing for object detection and object classification. Alternatively, the traffic type may be indicated by traffic parameters generated by the above-mentioned preprocessing and contained in the sensor data SD. In some embodiments, the margin may be adjusted based on a shielding parameter associated with the traffic type. The shielding parameter may indicate the degree to which the corresponding traffic type protects the person from the threat of fire. For example, the person may travel on foot (without a car) or by bicycle, motorcycle, car, bus, truck, military vehicle, helicopter, or the like. Any number of traffic types may be used. 7A to 7B The use of two traffic types is illustrated: unshielded and shielded. Figure 7A The distances D1 to D3 that can be set and accumulated to the margin D relative to the danger zone are illustrated. D1 represents a first distance given by the intensity of the fire, D2 represents a second distance given by the wind speed and the fire rate, and D3 represents a third distance for traffic types that will shield people (large shielding parameter), such as closed vehicles, here exemplified by cars or buses. Figure 7B In the Figure 7A The same as in D3, but D3 is extended because the traffic type used by people does not block people (small blocking parameters), here using people walking or cycling as an example. The same traffic type can use different blocking parameters for different types of hazard zones, for example, if the hazard zone represents a fire or explosion hazard or the risk of hazardous material release.

[0057] There are other conceivable methods of calculating the traffic type in step 406 .

[0058] In some embodiments, if personnel can enter the vehicle, step 406 takes the size of the vehicle into account when determining the evacuation path, such as by adjusting margins based on size and / or excluding terrain that is incompatible with the vehicle (see below).

[0059] In some embodiments, step 406 involves determining an evacuation path based on the estimated speed of the traffic type. The estimated speed can be predefined for each traffic type. For example, if an evacuation path is to be determined in a configuration space based on the transition time from the current location P1 to the desired destination P2, the estimated speed can be applied. Furthermore, given the predicted fire spread, some potential evacuation paths may only be open within a short time window, and the estimated speed can be used to determine whether the potential evacuation path is feasible.

[0060] The estimated speed may also be used to prioritize when providing directions to different persons in the target area 2. For example, persons with a low estimated speed may be prioritized over persons with a high estimated speed, or vice versa.

[0061] The estimated speed may also affect the repetition rate of step 406 (and other steps) in method 400. For a person traveling at a higher speed, the need to update the evacuation path may be less significant because the travel time from the current location P1 to the desired destination P2 may be smaller. Therefore, the repetition rate may decrease as the estimated speed increases.

[0062] In some embodiments, step 406 involves excluding terrain that is not traversable by the type of traffic. For example, a person walking may have other options for getting from the current location P1 to the desired destination P2 than a person riding a bicycle or driving a car. For example, the non-traversable terrain can be represented as a physical obstacle in the configuration space ( Figures 5 and 6 31 '). Therefore, the physical barriers may vary depending on the type of traffic. Alternatively or additionally, as described above with reference to Figure 6 As described above, the probabilistic roadmap algorithm can assign weights to subpaths in the configuration space based on traffic type. By assigning weights to represent the suitability of the terrain along the corresponding subpath for the traffic type, this method can effectively exclude terrain that is inaccessible to the traffic type and even prioritize fast routes for each traffic type. Those skilled in the art will recognize that traffic types can be integrated into other PPAs accordingly by using weights or corresponding adjustment factors.

[0063] If no transportation type is identified by method 400 , computer system 20 may be configured to assume that all persons utilize one transportation type, such as walking.

[0064] Figures 8A to 8CDifferent implementations of a wildfire evacuation support system comprising one or more drones 1 are depicted.

[0065] Figure 8A A first centralized implementation is shown, in which the computer system 20 is implemented on a remote device, such as a server or a cloud computer. The sensor data SD are generated by the respective drone 1 and transmitted to the computer system 20 via wireless communication. Figure 3 It can be noted that the current position P1 and / or the desired destination P2 may be included in or given by the sensor data SD, while the computer system 20 may obtain the metadata MD from other places, for example, from an external computer device (not shown). It is also conceivable that the computer system 20 obtains the current position P1 and / or the desired destination P2 from an external computer device. Figure 5 As shown, the computer system 20 is configured to determine one or more evacuation paths EP by using the method 400 and return the evacuation paths EP to the corresponding drone 1. Thus, the drone 1 can be used to guide people along the evacuation path EP, optionally together with other drones (not shown).

[0066] Figure 8B Shown with Figure 8A A second, different centralized implementation involves computer system 20 sending the evacuation path EP to another drone 1, rather than to the drone 1 that provided the sensor data SD. In one example, the guidance system may include drones dedicated to data collection and drones dedicated to providing evacuation guidance. In another example, computer system 20 may proactively identify the drone 1 within a group of drones to receive the evacuation path EP, based on, for example, the current drone location, drone thermal tolerance, drone functionality, and so forth.

[0067] It is also conceivable that the computer system 2 merges the sensor data SD from a plurality of drones and determines an evacuation route based on the merged information.

[0068] In a variant example, the computer system 20 returns the evacuation path EP to the drone 1, such as Figure 8A As shown, drone 1 then consults with other drones via drone-to-drone communication to determine the drone that guides the personnel along the evacuation path EP.

[0069] In a third centralized implementation not shown, the computer system 20 is located on a master drone among a plurality of drones deployed in a target area. The master drone can be configured to obtain sensor data SD from its sensor system and receive sensor data SD from other drones via short-range wireless communication. The master drone can be configured to determine an evacuation path EP based on the available sensor data SD. The master drone can also be configured to select one or more drones that guide personnel along the evacuation path EP and communicate the evacuation path EP to the selected drones. Figure 3 The current position P1 and / or the desired destination P2 may be given by sensor data SD or received from a remote computer (not shown), for example, via wireless communication. Metadata MD may also be obtained from the remote computer or pre-stored in the internal memory of the master drone.

[0070] Figure 8C A distributed implementation is shown, where the computer system 20 is located on a plurality of drones 1 deployed in the target area. Each drone 1 generates its own sensor data SD and determines an evacuation path EP based at least in part on the sensor data SD. Figure 8C As shown, the drones 1 can be configured to exchange sensor data SD, for example, via wireless short-range communication, and determine an evacuation path EP based on all available sensor data SD. Alternatively, each drone 1 can be configured to determine its own evacuation path EP using only its own sensor data SD, thereby exchanging its evacuation path EP with other drones. Each drone can also be configured to negotiate with other drones to decide on a drone that guides personnel along the evacuation path EP. If there are multiple different personnel who need to be guided, the drones can also negotiate to select a drone that guides the corresponding personnel. Figure 3 The current position P1 and / or the desired destination P2 may be given by sensor data SD or received from a remote computer (not shown), for example, via wireless communication. Metadata MD may also be obtained from the remote computer or pre-stored in the internal memory of the corresponding drone 1.

[0071] In some embodiments of a distributed implementation, drones 1 collaboratively create a heat map of target area 2 based on images captured by one or more cameras on each drone. The heat map may represent one or more of temperature, topography, people, buildings, roads, vehicles, and the like. The heat map is continuously shared and updated as the drones traverse target area 2. Thus, all drones have access to the same heat map. The corresponding drones 1 match the heat map with a geographic map of target area 2 (e.g., included in metadata MD). The geographic map may represent roads, topography, buildings, storage facilities for hazardous materials, and the like. As described above, the corresponding drones operate PPA to determine one or more evacuation routes based on one or more of the following factors: 1) wildfire spread rate and direction, 2) traffic patterns, 3) estimated vehicle speed, such as maximum speed, 4) shielding parameters, 5) available routes, including routes not typically traveled by vehicles, such as sidewalks and bike paths, and 6) safe distances from fire and other risks.

[0072] After determining an evacuation path in a distributed or centralized implementation, evacuation guidance can be provided by a single drone per person or group of people, or by a fleet of drones creating a visual and / or auditory path for safe evacuation. During an evacuation, the drones can use their sensor systems to proactively detect unexpected hazards and guide people around them. These hazards might include fallen trees or other obstacles, or approaching vehicles that could be hidden from view.

[0073] exist Figure 8C In a variant of , each drone 1 automatically uses its own sensor data SD to determine its own evacuation path EP and provides guidance accordingly.

[0074] It is important to note that short-range wireless communications between drones are unlikely to be affected by wildfires. If a disruption occurs, drones may simply increase their altitude to reduce the impact of the wildfire. For long-range wireless communications, even though cellular communications may be disrupted in or around a wildfire, drones can use other communication technologies, or drones can intermittently move away from the wildfire and communicate with remote computers within range of an operational cellular network.

[0075] Computer system 20 may be implemented by one or more software-controlled computer resources. Figure 9Such a computer resource 90 is schematically depicted and includes a processing system 91, a computer memory 92, and a communication interface 93 for inputting and / or outputting data. The computer resource 90 may or may not be a single device. The processing system 91 may, for example, include one or more CPUs ("Central Processing Units"), DSPs ("Digital Signal Processors"), microprocessors, microcontrollers, ASICs ("Application Specific Integrated Circuits"), a combination of discrete analog and / or digital components, or some other programmable logic device, such as an FPGA ("Field Programmable Gate Array"). A control program 92A comprising computer instructions is stored in the memory 92 and executed by the processing system 91 to perform any of the methods, operations, functions, or steps described above. Figure 9 As shown, the memory 92 may also store control data 92B for use by the processing system 92. The control program 92A may be provided to the computer resource 90' ​​on a computer-readable medium 100, which may be a tangible (non-transitory) product (e.g., magnetic media, optical disk, read-only memory, flash memory, etc.) or a propagated signal.

[0076] While the inventive subject matter has been described in connection with what are presently considered to be the most practical embodiments, it is to be understood that the inventive subject matter is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

[0077] Furthermore, while operations are depicted in a particular order in the drawings, this should not be understood as requiring that such operations be performed in the particular order shown, or in sequential order, or that all illustrated operations be performed to achieve desirable results. In certain circumstances, parallel processing may be advantageous.

[0078] In the following, clauses are cited to summarize some aspects and embodiments as disclosed in the foregoing.

[0079] C1. A computer system for providing wildfire evacuation support, the computer system comprising: an input end 83 for receiving sensor data SD generated by at least one unmanned aerial vehicle 1 when it is deployed in a target area 2 containing a wildfire 3, the at least one unmanned aerial vehicle 1 being configured to generate the sensor data SD representing the wildfire 3 in the target area by using one or more sensors 15, at least one processor 81, the at least one processor being connected to the input end 83 and being configured to: obtain a current position P1 of a person in a dangerous situation in the target area 2; obtain a desired destination P2 of the person in the target area 2; based on the sensor data SD, identify one or more danger zones 32, 33 in the target area 2, the one or more danger zones 32, 33 posing a fire-related threat to the person; and determine at least one evacuation path EP extending from the current position P1 to the desired destination P2 while avoiding the one or more danger zones 32, 33.

[0080] C2. The computer system of C1, wherein the sensor data SD represents the location and extent of one or more confined space fires F in the wildfire 3, and the intensity of one or more confined space fires F.

[0081] C3. The computer system according to C1 or C2, the computer system being configured to obtain the current position P1 of the person by detecting the person based on the sensor data SD and determining the current position P1 of the person.

[0082] C4. A computer system according to any of the preceding clauses, wherein the sensor data SD represents a real-time image.

[0083] C5. A computer system according to C4, wherein the real-time image comprises an infrared image.

[0084] C6. The computer system according to claim 1, wherein the computer system is configured to enable the at least one UAV 1 to guide the personnel based on the at least one evacuation route EP.

[0085] C7. A computer system according to any of the preceding clauses, wherein the computer system is further configured to: represent each of the one or more danger zones 32, 33 as an obstacle 32', 33' in the configuration space 30 of the path planning algorithm, and run the path planning algorithm to determine the at least one evacuation path EP.

[0086] C8. The computer system of C7, wherein the corresponding hazard zone 32, 33 of the one or more hazard zones corresponds to a confined space fire 32 in the wildfire 3, or a storage facility 33 for hazardous materials.

[0087] C9. A computer system according to C7 or C8, wherein the computer system is configured to define the obstacles 32', 33' by a margin D from the corresponding danger zone 32, 33, wherein the margin D represents a safe distance between the personnel and the corresponding danger zone 32, 33, wherein the computer system 20 is configured to set the margin D based on wind data or at least one of the characteristics of the wildfire 3.

[0088] C10. The computer system of C9, configured to determine the characteristics of the wildfire 3 based on the sensor data SD.

[0089] C11. A computer system according to C9 or C10, wherein the characteristic includes at least one of a rate of spread, a fire intensity, or a direction of spread.

[0090] C12. The computer system according to any one of C9 to C11, further configured to: determine a traffic pattern of the person based on the sensor data SD, and determine the margin D based on a shielding parameter related to the traffic pattern.

[0091] C13. The computer system according to any one of C1 to C11, further configured to: determine a traffic pattern of the personnel based on the sensor data SD, and determine the at least one evacuation path EP based on the traffic pattern.

[0092] C14. The computer system according to C13, wherein the computer system is configured to determine the at least one evacuation path EP based on an estimated speed of the traffic pattern.

[0093] C15. The computer system according to C13 or C14, wherein the computer system is configured to determine at least one evacuation path EP while excluding terrain in the target area 2 that cannot be traversed by the traffic mode.

[0094] C16. A computer system according to any of the preceding clauses, wherein the computer system is located on the unmanned aerial vehicle 1.

[0095] C17. A computer system according to any of the preceding clauses, wherein the computer system is configured to repeatedly obtain the current position P1 of the person and optionally repeatedly obtain the desired destination P2, identify the one or more danger zones 32, 33 in the target area 2, and determine the at least one evacuation path EP, so that the at least one evacuation path is updated over time as the person passes through the target area 2.

[0096] C18. A system for providing wildfire evacuation support, the system comprising a plurality of unmanned aerial vehicles 1 and a computer system according to any of the preceding clauses, wherein the computer system is located on at least one of the plurality of unmanned aerial vehicles 1, and wherein the plurality of unmanned aerial vehicles 1 are configured to exchange the sensor data SD between the plurality of unmanned aerial vehicles via wireless transmission.

[0097] C19. A computer-implemented method for providing wildfire evacuation support, the method comprising: obtaining 402 sensor data generated by one or more sensors on at least one unmanned aerial vehicle operating in a target area containing a wildfire, the sensor data representing the wildfire in the target area; obtaining 403 a current location of a person in a dangerous situation in the target area containing the wildfire; obtaining 404 a desired destination in the target area; identifying 405 one or more hazard zones in the target area based at least in part on the sensor data, the one or more hazard zones posing a fire-related threat to the person; and determining 406 at least one evacuation path extending from the current location to the desired destination while avoiding the one or more hazard zones.

[0098] C20. A computer-readable medium comprising computer instructions, which, when executed by at least one processor 81, cause the at least one processor 81 to perform the method according to C19.

[0099] C21. An unmanned aerial vehicle for deployment in a target area 2 containing a wildfire 3, the unmanned aerial vehicle comprising one or more sensors 15 and configured to: generate sensor data SD representing the wildfire 3 in the target area 2 by using the one or more sensors 15; obtain a current position P1 of a person in a dangerous situation in the target area 2; obtain a desired destination P2 of the person in the target area 2; identify one or more danger zones 32, 33 in the target area 2 based on the sensor data SD, the one or more danger zones 32, 33 posing a fire-related threat to the person; and determine at least one evacuation path EP extending from the current position P1 to the desired destination P2 while avoiding the one or more danger zones 32, 33.

[0100] C22. An unmanned aerial vehicle according to C21, wherein the unmanned aerial vehicle is included in a plurality of unmanned aerial vehicles for deployment in the target area 2, wherein the unmanned aerial vehicle is also configured to receive other sensor data representing a wildfire 3 in the target area 2 from one or more other unmanned aerial vehicles among the plurality of unmanned aerial vehicles, wherein the unmanned aerial vehicle is configured to identify the one or more danger zones based on the sensor data and the other sensor data.

[0101] C23. The UAV according to C22, further configured to: select at least one UAV from the plurality of UAVs, and enable the at least one UAV to guide the personnel based on the at least one evacuation path EP.

Claims

1. A computer system for providing wildfire evacuation support, the computer system comprising: an input (83) for receiving sensor data (SD) generated by at least one unmanned aerial vehicle (1) when deployed in a target area (2) containing a wildfire (3), the at least one unmanned aerial vehicle (1) being configured to generate the sensor data (SD) representative of the wildfire (3) in the target area by using one or more sensors (15), at least one processor (81) connected to the input (83) and configured to: Obtaining the current position (P1) of the person in danger in the target area (2), obtaining the desired destination (P2) of the person in the target area (2), identifying one or more danger zones in the target area (2) based on the sensor data (SD), the one or more danger zones posing a fire-related threat to the personnel, and determining at least one evacuation path (EP) extending from said current position (P1) to said desired destination (P2) while avoiding said one or more danger zones, Wherein, the computer system is further configured to: Representing each of the one or more danger zones as an obstacle in a configuration space of a path planning algorithm, the obstacle being defined as having a margin (D) representing a safe distance between the person and the corresponding danger zone; determining a traffic pattern of the person based on the sensor data (SD), and determining the margin (D) based on a shielding parameter associated with the traffic pattern, wherein the shielding parameter represents the extent to which the respective traffic pattern protects the person from fire-related threats; and The path planning algorithm is run to determine the at least one evacuation path.

2. The computer system according to claim 1, wherein: The sensor data (SD) represents the location and extent of one or more confined space fires (F) in the wildfire (3), and the intensity of the one or more confined space fires (F).

3. The computer system according to claim 1, wherein the computer system is configured to obtain the current position (P1) of the person by detecting the person based on the sensor data (SD) and determining the current position (P1) of the person.

4. The computer system according to claim 1, wherein: The sensor data (SD) represents a real-time image.

5. The computer system according to claim 4, wherein: The real-time image includes an infrared image.

6. The computer system according to claim 1, wherein the computer system is configured to cause the at least one UAV (1) to guide the personnel based on the at least one evacuation path (EP).

7. The computer system according to claim 1, wherein: A respective hazard zone among the one or more hazard zones corresponds to a confined space fire (32) in the wildfire (3), or a storage facility (33) for hazardous materials.

8. The computer system according to claim 1, wherein: The computer system (20) is configured to set the margin (D) based on at least one of wind data or a characteristic of the wildfire (3).

9. The computer system of claim 8, configured to determine the characteristics of the wildfire (3) based on the sensor data (SD).

10. The computer system according to claim 8, wherein: The characteristic includes at least one of rate of spread, fire intensity, or direction of spread.

11. The computer system according to claim 1, further configured to determine a traffic pattern of the persons based on the sensor data (SD), and to determine the at least one evacuation path (EP) based on the traffic pattern.

12. The computer system of claim 11, configured to determine the at least one evacuation path (EP) based on an estimated speed of the traffic pattern.

13. The computer system of claim 11, wherein the computer system is configured to determine the at least one evacuation path (EP) while excluding terrain in the target area (2) that cannot be traversed by the traffic pattern.

14. The computer system of claim 1, wherein the computer system is located on an unmanned aerial vehicle (1).

15. The computer system of claim 1 , wherein the computer system is configured to repeatedly obtain the current position (P1) of the person, identify the one or more danger zones in the target area (2), and determine the at least one evacuation path (EP), such that the at least one evacuation path is updated over time as the person passes through the target area (2).

16. The computer system of claim 1, configured to repeatedly obtain the desired destination (P2).

17. A system for providing wildfire evacuation support, the system comprising a plurality of unmanned aerial vehicles (1) and a computer system according to claim 1, wherein: The computer system is located on at least one of the plurality of UAVs (1), and wherein the plurality of UAVs (1) are configured to exchange the sensor data (SD) between the plurality of UAVs via wireless transmission.

18. A computer-implemented method for providing wildfire evacuation support, the method comprising: obtaining sensor data generated by one or more sensors on at least one unmanned aerial vehicle operating in a target area containing a wildfire, the sensor data representing the wildfire in the target area; Obtain the current location of persons in hazardous situations within a target area containing a wildfire; obtaining a desired destination in the target area; identifying one or more hazard zones in the target area based at least in part on the sensor data, the one or more hazard zones posing a fire-related threat to the personnel; Representing each of the one or more danger zones as an obstacle in a configuration space of a path planning algorithm, the obstacle being defined as having a margin (D) representing a safe distance between the person and the corresponding danger zone; determining a traffic pattern of the person based on the sensor data (SD), and determining the margin (D) based on a shielding parameter associated with the traffic pattern, wherein the shielding parameter represents the extent to which the corresponding traffic pattern protects the person from fire-related threats; as well as The path planning algorithm is executed to determine at least one evacuation path extending from the current location to the desired destination while avoiding the one or more hazard zones.

19. A computer-readable medium comprising computer instructions which, when executed by at least one processor (81), cause the at least one processor (81) to perform the method according to claim 18.

20. An unmanned aerial vehicle for deployment in a target area (2) containing a wildfire (3), the unmanned aerial vehicle comprising one or more sensors (15) and configured to: generating sensor data (SD) representative of the wildfire (3) in the target area (2) using the one or more sensors (15); Obtaining a current position (P1) of a person in a dangerous situation in the target area (2); obtaining a desired destination (P2) of the person in the target area (2); identifying one or more danger zones in the target area (2) based on the sensor data (SD), the one or more danger zones posing a fire-related threat to the personnel; Representing each of the one or more danger zones as an obstacle in a configuration space of a path planning algorithm, the obstacle being defined as having a margin (D) representing a safe distance between the person and the corresponding danger zone; A traffic pattern of the person is determined based on the sensor data (SD), and the margin (D) is determined based on a shielding parameter associated with the traffic pattern, wherein The shielding parameter represents the degree to which the corresponding traffic pattern protects the personnel from fire-related threats; as well as The path planning algorithm is run to determine at least one evacuation path (EP) extending from the current location (P1) to the desired destination (P2) while avoiding the one or more danger zones.

21. The UAV according to claim 20, being included in a plurality of UAVs for deployment in the target area (2), wherein: The UAV is also configured to receive additional sensor data representing a wildfire (3) in the target area (2) from one or more other UAVs among the plurality of UAVs, wherein the UAV is configured to identify the one or more hazard zones based on the sensor data and the additional sensor data.

22. The UAV according to claim 21, further configured to: select at least one UAV from the plurality of UAVs, and enable the at least one UAV to guide the personnel based on the at least one evacuation path (EP).

Images