Method and device for determining the position of an unmanned vehicle and an unmanned aerial vehicle

Abstract

A method for determining the position of an unmanned vehicle (UV) is provided. The method includes receiving positioning signals from a plurality of satellites in a global navigation satellite system. Furthermore, the method includes estimating the position of the UV based on (i) a three-dimensional model of the UV's environment and (ii) possible signal paths for each of the positioning signals. At least a portion of the possible signal paths includes reflections of the positioning signals by one or more objects in the UV's environment.

Description

Technical Field

[0001] The present disclosure relates to determining the position of an unmanned vehicle (UV). Specifically, examples relate to methods and apparatus for determining the position of an UV. Other examples relate to unmanned aerial vehicles (UAVs). Background Art

[0002] Global Navigation Satellite Systems (GNSS) such as the Global Positioning System (GPS) are commonly used to provide absolute, globally referenced position information for autonomous vehicles such as drones and autonomous cars. However, such systems rely on radio signals that can be blocked or reflected by large structures in the robot's environment, resulting in degraded position estimates or even complete disruption.

[0003] Therefore, there may be a need for improved position determination. Summary of the Invention

[0004] This need is met by an apparatus and a method according to the independent claims. Advantageous embodiments are set forth in the dependent claims.

[0005] According to a first aspect, the present disclosure provides a method for determining the position of a UV. The method includes receiving positioning signals from a plurality of satellites of a global positioning system (GNSS). Furthermore, the method includes estimating the position of the UV based on (i) a three-dimensional model of the UV's environment and (ii) possible signal paths for each of the positioning signals. At least a portion of the possible signal paths includes reflections of the positioning signals from one or more objects in the UV's environment.

[0006] According to a second aspect, the present disclosure provides a non-transitory machine-readable medium having stored thereon a program having program code for controlling the UV to perform the method for determining the position of the UV as described herein when the program is executed on a processor or programmable hardware of the UV.

[0007] According to a third aspect, the present disclosure provides an apparatus for determining the position of a UV. The apparatus includes receiver circuitry configured to receive positioning signals from a plurality of GNSS satellites. Furthermore, the apparatus includes processing circuitry configured to estimate the position of the UV based on (i) a three-dimensional model of the UV's environment and (ii) possible signal paths for each of the positioning signals. At least a portion of the possible signal paths for each of the positioning signals includes reflections of the positioning signal from one or more objects in the UV's environment.

[0008] According to a fourth aspect, the present disclosure provides a UAV comprising an apparatus for determining a position of the UAV as described herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Some examples of apparatus and / or methods will be described below, by way of example only, and with reference to the accompanying drawings, in which:

[0010] Figure 1 A flow chart illustrating an example of a method for determining the position of a UV;

[0011] Figure 2 An example of a UAV in an environment is shown;

[0012] Figure 3 An example of a two-dimensional representation of a three-dimensional model of the UAV's environment is shown;

[0013] Figure 4 An example of an architecture for estimating the position of UVs is shown;

[0014] Figure 5 Another example of a UAV in an environment is shown; and

[0015] Figure 6 An example of a UAV including means for determining a position of the UAV is shown. DETAILED DESCRIPTION

[0016] Various examples will now be described more fully with reference to the accompanying drawings, in which some examples are shown. In the accompanying drawings, the thickness of lines, layers and / or regions may be exaggerated for clarity.

[0017] Therefore, although other examples can have various modifications and alternative forms, some specific examples thereof are shown in the drawings and these specific examples will be described in detail later. However, this detailed description does not limit other examples to the specific forms described. Other examples may encompass all modifications, equivalents, and alternatives that fall within the scope of this disclosure. Throughout the description of the drawings, identical or similar numbers refer to identical or similar elements, and when compared to each other, identical or similar elements can be implemented identically or in modified form while providing the same or similar functions.

[0018] It should be understood that when an element is referred to as being "connected" or "coupled" to another element, these elements may be directly connected or coupled via one or more intermediate elements. Without further explicit or implicit definition, if two elements A and B are combined using "or", this will be understood to disclose all possible combinations, i.e., only A, only B, and A and B. Alternative wording for the same combination is "at least one of A and B" or "A and / or B". This also applies, mutatis mutandis, to combinations of more than two elements.

[0019] The terms used herein to describe specific examples are not intended to limit other examples. Whenever singular forms such as "a", "an", and "the" are used and the use of only a single element is neither explicitly nor implicitly defined as mandatory, other examples may also use multiple elements to implement the same function. Similarly, when a function is subsequently described as being implemented using multiple elements, other examples may use a single element or processing entity to implement the same function. It will also be understood that the terms "comprises", "comprising", "includes", and / or "including" when used specify the presence of the features, integers, steps, operations, processes, actions, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, processes, actions, elements, components, and / or any groups thereof.

[0020] Unless otherwise defined, all terms (including technical and scientific terms) are used herein in accordance with the common meaning of the art to which they belong.

[0021] Figure 1 A flow chart of a method 100 for determining the position of a UV is shown. For simplicity, the proposed technology will be described below primarily with reference to UAVs. However, it should be noted that the technology is not limited to UAVs and can also be used with other unmanned vehicles (unmanned vehicles) such as unmanned ground vehicles (UGVs).

[0022] Figure 2 An exemplary scenario including UAV 200 is shown. Figure 2 As can be seen, UAV 200 is located between two multi-story buildings 210 and 220. Further illustrated are four GNSS satellites 230-1, ..., 230-4 (such as GPS, GLONASS, Galileo, or BeiDou). Each of the four satellites 230-1, ..., 230-4 provides one or more positioning signals that allow for geospatial positioning of UAV 200. Specifically, UAV 200 can accurately measure the time it takes for a positioning signal (e.g., a radio frequency signal) to travel from the satellite to a receiver of UAV 200 (as well as other quantities such as Doppler shift).

[0023] As from Figure 2It can be seen that the geometric shapes of buildings 210 and 220 in the environment (surrounding area) of UAV 200 affect the propagation of positioning signals. For example, the positioning signals of the four satellites 230-1, ..., 230-4 may be blocked or reflected by objects in the environment of UAV 200, so that UAV 200 cannot receive some of the positioning signals or can only receive the reflected positioning signals. Figure 2 In the example shown in FIG2 , positioning signal 231-1 of satellite 230-1 is reflected by building 220, resulting in no line-of-sight propagation of positioning signal 231-1 to UAV 200. Positioning signal 231-4 of satellite 230-4 is completely blocked by building 220, resulting in no reception of positioning signal 231-4 by UAV 200. Only positioning signals 231-2 and 231-3 of satellites 230-2 and 230-3 are unaffected by the environment of UAV 200, resulting in line-of-sight propagation of positioning signals 231-2 and 231-3 to UAV 200.

[0024] If the positioning signal of one of the satellites 230-1, ..., 230-4 is blocked, the UAV 200 cannot use it for position determination. If the positioning signal of one of the satellites 230-1, ..., 230-4 is reflected, processing the positioning signal like an unreflected signal (i.e., as a sum of line-of-sight propagation) can result in an erroneous position estimate because the distance to the satellite is incorrectly estimated.

[0025] It should be noted that Figure 2 The buildings 210 and 220 shown are merely examples of objects in the environment of the UAV 200. Any large natural or man-made object (structure) between the GNSS satellites and the UAV 200 can completely block the satellite's positioning signals or reflect them. For example, such objects can be tall buildings in a city, large rocks in a mountainous area, valleys, dams, etc.

[0026] The method 100 may also allow for more accurate determination of the position of a UV, such as the UAV 200, in the presence of objects in the UV's environment. Figure 1 , the method 100 includes receiving 102 positioning signals from a plurality of satellites of a GNSS (e.g., Figure 2). In addition, method 100 includes estimating (determining) (104) the position of the UV based on (i) the three-dimensional model of the environment of the UV and (ii) possible signal paths of each of the positioning signals. At least a portion of the possible signal paths include reflections of the positioning signal by one or more objects in the environment of the UV. In other words, at least some of the possible signal paths are signal paths that reproduce reflections of the positioning signal by one or more objects in the environment of the UV. The three-dimensional model of the environment of the UV includes at least information about the position, orientation, and shape of objects in the environment of the UV.

[0027] The position of the UV is estimated based on a three-dimensional model of the UV's environment and the possible signal paths of the reconstructed reflections of the positioning signal. The information provided by the reflected positioning signal can also be used to determine the UV's position. By changing the UV's current position (or thus estimated) and taking into account the three-dimensional model of the UV's environment and the a priori known positions of the satellites, the possible signal paths for each of the positioning signals can be determined (e.g., using a probabilistic model). The corresponding most likely signal path among the possible signal paths for each of the positioning signals can be determined based on the positioning signal (or information derived from / related to the positioning signal). Therefore, the distance of the UV to the satellite whose positioning signal reached the UV via one or more reflections can also be accurately determined (or at least with greater accuracy than conventional methods). Therefore, compared to conventional methods, the reflected positioning signal can also be used for precise position determination. Therefore, all received positioning signals can be used for position determination. For example, according to the proposed technology, positioning signals 231-1, 231-2, and 231-3 can all be used to determine the position of UAV 200.

[0028] Figure 3 Shown Figure 2 A two-dimensional representation of an exemplary three-dimensional model of the depicted scene. Figure 3 As can be seen, the geometric shapes of buildings 210 and 220 existing in the real world are modeled by objects 310-1, 310-2, and 320 in the three-dimensional model of the environment of UAV 200. In addition, objects 310-1, 310-2, and 320 are semantically labeled. Objects 310-1, 310-2, and 320 are labeled as buildings.

[0029] Labeling objects in the three-dimensional model of the environment of the UAV 200 may allow for prediction of the effect of objects (structures) on the positioning signals of the GNSS satellites 230-1, ..., 230-4. Figure 1As shown, objects can be semantically tagged with information about the material category of the corresponding object (e.g., material suitable for buildings, glass, concrete, steel, stone, rock, etc.). Alternatively or additionally, objects can be semantically tagged with information about the radio frequency characteristics of the corresponding object (e.g., radio frequency reflectivity, radio frequency absorption, etc.).

[0030] exist Figure 3 In the example of FIG, the three-dimensional model of the environment of the UAV 200 includes only static objects for modeling buildings 210 and 220. In other examples, the three-dimensional model of the environment of the UV may include at least one mobile (dynamic) object for modeling dynamic (moving) objects in the real world, such as cars. For example, a car moving in the environment of the UV may reflect a positioning signal (e.g., a GPS signal) in different directions over time and further affect its Doppler shift. The semantic model of the environment of the UV including one or more dynamic objects (e.g., modeling geometry, pose, material properties, and velocity) may allow the impact of a moving car to be predicted, so that the position estimate of the UV may be further improved. In other words, some implementations may take into account the effects of dynamically moving objects.

[0031] like Figure 1 As shown, the method 100 may also include determining (106) a three-dimensional model of the environment of the UV based on sensor data of one or more sensors of the UV. For example, determining (106) the three-dimensional model of the environment of the UV and estimating (104) the position of the UV may be performed simultaneously. In some examples, the one or more sensors of the UV include at least an optical sensor (e.g., a still camera or a moving picture camera) and an inertial measurement unit (IMU). Optionally or alternatively, the one or more sensors of the UV may include other sensors, such as one or more of a magnetometer, a barometer, a radar sensor, and a lidar sensor.

[0032] exist Figure 4 An example of an architecture for simultaneously estimating (104) the position of a UV (e.g., UAV 200) to determine (106) a three-dimensional model of the UV's environment is shown in FIG.

[0033] A simultaneous localization and mapping (SLAM) algorithm 410 is used to determine a three-dimensional representation of the UV's environment based on the sensor data 401, 402 from the optical sensor and the IMU. Optionally, sensor data from other sensors (e.g., a barometer, a radar sensor, etc.) can be provided to the SLAM algorithm to determine a three-dimensional representation of the UV's environment. In addition to the three-dimensional representation of the UV's environment, the SLAM algorithm also determines a pose estimate 404 and a velocity estimate 405 for the UV.

[0034] Furthermore, image classification algorithm 420 is used to classify at least one object of the UV environment, represented by sensor data 401 from the optical sensor, into one of a plurality of predefined categories. To determine the category of the object, sensor data 401 is evaluated by image classification algorithm 420. That is, the classification of at least one object of the UV environment is based on sensor data 401 from the optical sensor. For example, unsupervised, semi-supervised, or supervised classification (e.g., using a convolutional neural network) can be used for image classification. The output of image classification algorithm 420 is semantic labels 406 for the objects in three-dimensional representation 403 of the UV environment, these semantic labels representing corresponding classification categories. As described above, the predefined categories may include radio frequency characteristics and / or material categories.

[0035] At least one object in the three-dimensional representation of the environment of the UV is semantically labeled with a classification category to obtain a three-dimensional model 430 of the environment of the UV.

[0036] The SLAM algorithm 410 is used in conjunction with a particle filter 440, which represents a probabilistic model of the possible signal paths of each positioning signal received from a GNSS satellite. The particle filter 440 receives as input the pose estimate 404 and velocity estimate 405 of the UV from the SLAM algorithm 410, a three-dimensional model 430 of the UV's environment, the position 407 of the satellite, and information 408 related to the positioning signal. The position 407 of the satellite is known a priori. The information 408 related to the positioning signal can be, for example, the positioning signal itself or a quantity derived from the positioning signal (e.g., time of arrival at the UV, propagation time from the satellite to the UV, etc.).

[0037] Particle filter 440 is used to determine the most likely signal path among the possible signal paths for each positioning signal. Particle filter 440 can be understood as a probabilistic model that includes the probability of the positioning signal being reflected in various ways or not being reflected at all. Based on the UV's attitude estimate 404 and velocity estimate 405, the particle filter can use the UV's three-dimensional model 430 of the environment, the satellite's position 407, and information 408 related to the positioning signal to determine the most likely signal path among the possible signal paths for each positioning signal. In other words, the path of the reflected positioning signal can be estimated with the help of particle filter 440. Therefore, particle filter 440 allows the posterior probability of the current attitude estimate of the UV to be maximized.

[0038] The SLAM algorithm 410 updates the pose estimate and the velocity estimate using the output of the particle filter 440. In other words, the pose estimate 404 and the velocity estimate 405 of the SLAM algorithm 410 are updated based on the corresponding most likely signal path of the possible signal paths of each of the positioning signals. The position indicated by the updated pose estimate of the SLAM algorithm 410 is determined as the position of UV.

[0039] Furthermore, the three-dimensional model 430 of UV's environment is updated based on the updated pose estimate and the updated velocity estimate of the SLAM algorithm 410 .

[0040] exist Figure 4 The data exchanged via the dashed lines in the example may include uncertain information.

[0041] Figure 4 The architecture shown may allow the paths of positioning signals from satellites (with known positions) to be used to simultaneously estimate the position of the UV and the geometry and characteristics of the UV's environment.

[0042] The proposed technology can allow information from all available (i.e., received) positioning signals to be combined via a semantically annotated global reference model of the UV's environment. As described above, the model of the UV's environment can be obtained by SLAM combined with semantic classification of image data. Using the (a priori known) position of the satellite and the three-dimensional shape, position, and orientation of obstacles in the path of the positioning signal (given by the model of the UV's environment), even signals that reach the UV's GNSS receiver via one or more reflections can provide information to increase the accuracy of the position information. Semantic information attached to the estimated geometry of the environment (given by the model of the UV's environment) can allow the prediction of the effect of objects (structures) on the positioning signal (e.g., whether the object will absorb or reflect the positioning signal).

[0043] The proposed technique can visually estimate the position, orientation, shape, and material of (large) objects / structures in a UV environment to directly predict and account for their impact on the positioning signals emitted by GNSS satellites. Therefore, the accuracy of GNSS-based position estimation in environments with large structures such as large buildings in cities or large rocks in mountainous areas can be improved.

[0044] Despite Figure 4 In the example of FIGURE 4 , a particle filter 440 is used, but it should be noted that in alternative examples, another filter such as a Kalman filter or an expectation-maximization filter may be used.

[0045] Compared to conventional methods, the proposed technique allows to extract position and velocity information from localization signals reflected by objects / structures in the UV environment. Figure 5An exemplary scenario is shown in which a UAV 500 is flying under a road bridge 510. There is no direct line of sight to any of the four GNSS satellites 520-1, ..., 520-4. Positioning signals 521-2 and 521-3 from satellites 520-2 and 520-3 are blocked by the road bridge 510. Positioning signals 521-1 and 521-4 from satellites 520-1 and 520-4 are reflected by soil 530. In an environment such as Figure 5 In the illustrated scenario, the proposed technique is able to obtain additional information from positioning signals 521-1 and 521-4 to update the estimated position of UAV 500. A semantically annotated three-dimensional model of the environment of UAV 500 allows the reflections of positioning signals 521-1 and 521-4 to be predicted, so that the position can be accurately estimated from positioning signals 521-1 and 521-4.

[0046] exist Figure 6 An example of using UVs determined by position according to the proposed technique is further shown in . Figure 6 In the figure, the UV is a UAV 600 that is depicted as a multi-rotor drone (e.g., a quadcopter or a bi-rotor aircraft) including a plurality of rotors 640-1, 640-2, .... However, it should be noted that the UV is not limited thereto. In general, the UV can be any type of UGV or UAV, such as a single-rotor aircraft, a fixed-wing UAV (e.g., an airplane or a vertical take-off and landing (VTOL) aircraft), or an autonomous vehicle.

[0047] The UAV 600 includes an apparatus 610 for determining the location of the UAV 600. The apparatus 610 includes a receiver circuit 611 coupled to a receiving antenna 630 and configured to receive positioning signals 601 from a plurality of satellites of a GNSS (e.g., GPS). In addition, the apparatus 610 includes a processing circuit 612. For example, the processing circuit 612 can be a single dedicated processor, a single shared processor, or multiple individual processors (some or all of which can be shared), digital signal processor (DSP) hardware, application specific integrated circuit (ASIC), or field programmable gate array (FPGA). The processing circuit 612 can optionally be coupled to, for example, a read-only memory (ROM), random access memory (RAM), and / or non-volatile memory for storing software. The apparatus 610 can also include other hardware—conventional and / or custom.

[0048] The processing circuit 612 receives data related to the received positioning signal 601 and optionally other data from one or more sensors of the UAV (e.g., optical sensor 620 or IMU). The processing circuit 612 processes the data according to the techniques for determining the position of the UAV 600 described above.

[0049] The UAV 600 may also include other hardware—conventional and / or custom—602 (e.g., an antenna for receiving control signals from a remote control).

[0050] The following examples relate to other embodiments:

[0051] (1) A method for determining a position of a UV. The method includes receiving positioning signals from a plurality of satellites in a global navigation satellite system. Furthermore, the method includes estimating the position of the UV based on (i) a three-dimensional model of an environment of the UV and (ii) possible signal paths for each of the positioning signals. At least a portion of the possible signal paths includes reflections of the positioning signals by one or more objects in the environment of the UV.

[0052] (2) The method according to (1), wherein objects in the three-dimensional model of the UV environment are semantically labeled with information about the radio frequency characteristics and / or material class of the corresponding objects.

[0053] (3) The method according to (1) or (2), wherein the three-dimensional model of the UV environment includes only static objects.

[0054] (4) The method according to (1) or (2), wherein the three-dimensional model of the UV environment includes at least one moving object.

[0055] (5) The method according to any one of (1) to (4), wherein the method further comprises determining a three-dimensional model of the environment of the UV based on sensor data of one or more sensors of the UV.

[0056] (6) The method of (5), wherein the one or more sensors of the UV include at least an optical sensor and an inertial measurement unit.

[0057] (7) The method according to (6), wherein determining the three-dimensional model of the environment of the UV includes: determining a three-dimensional representation of the environment of the UV using a SLAM algorithm based on sensor data of an optical sensor and an inertial measurement unit; classifying at least one object of the environment of the UV represented by the sensor data of the optical sensor into one of a plurality of predefined categories based on the sensor data of the optical sensor; and semantically labeling at least one object in the three-dimensional representation of the environment of the UV with the classified category so as to obtain the three-dimensional model of the environment of the UV.

[0058] (8) The method according to (7), wherein the predefined categories are radio frequency characteristics and / or material categories.

[0059] (9) A method according to (7) or (8), wherein the possible signal paths of each of the positioning signals are represented by a particle filter, and wherein the method includes using the particle filter to determine the corresponding most likely signal path among the possible signal paths of each of the positioning signals.

[0060] (10) The method according to (9), wherein the particle filter receives as input the pose estimate and velocity estimate of the UV from the SLAM algorithm, the three-dimensional model of the environment of the UV, the position of the satellite, and information related to the positioning signal.

[0061] (11) A method according to (9) or (10), wherein the method further includes updating the pose estimate and velocity estimate of the SLAM algorithm based on the corresponding most likely signal path among the possible signal paths of each positioning signal.

[0062] (12) The method of (11), wherein estimating the position of the UV comprises determining the position indicated by the updated pose estimate of the SLAM algorithm as the position of the UV.

[0063] (13) The method of (11) or (12), wherein the method further comprises updating the three-dimensional model of the UV environment based on the updated pose estimate and the updated velocity estimate of the SLAM algorithm.

[0064] (14) The method according to any one of (1) to (13), wherein the method is performed by UV.

[0065] (15) The method according to any one of (1) to (14), wherein UV is UAV.

[0066] (16) The method of (15), wherein the UAV is a multi-rotor drone.

[0067] (17) A non-transitory machine-readable medium having a program stored thereon, the program having program code, which, when executed on a processor or programmable hardware of a UV, controls the UV to perform the method for determining the position of the UV according to any one of (1) to (16).

[0068] (18) An apparatus for determining a position of a UV. The apparatus includes a receiver circuit configured to receive positioning signals from a plurality of satellites in a global navigation satellite system. Furthermore, the apparatus includes a processing circuit configured to estimate the position of the UV based on (i) a three-dimensional model of an environment of the UV and (ii) possible signal paths for each of the positioning signals, wherein at least a portion of the possible signal paths for each of the positioning signals includes reflections of the positioning signals by one or more objects in the environment of the UV.

[0069] (19) A UAV comprising an apparatus for determining the position of the UAV according to (18).

[0070] (20) The UAV according to (19), wherein the UAV is a multi-rotor drone.

[0071] Aspects and features mentioned and described in conjunction with one or more of the previous detailed examples and figures may also be combined with one or more of the other examples in order to replace similar features of the other examples or otherwise introduce the features into the other examples.

[0072] The description and drawings illustrate only the principles of the present disclosure. In addition, all examples listed herein are primarily and explicitly intended to be used for illustrative purposes only to help the reader understand the principles of the present disclosure and the concepts contributed by the inventors to advance the art. All statements herein reciting principles, aspects, and examples of the present disclosure, as well as specific examples of the present disclosure, are intended to encompass their equivalents.

[0073] A block diagram may, for example, show a high-level circuit diagram that implements the principles of the present disclosure. Similarly, a flow chart, a flow diagram, a state transition diagram, a pseudocode, or the like may represent various processes, operations, or steps that, for example, may be substantially represented in a non-transitory machine-readable medium (e.g., a floppy disk, DVD, Blu-ray Disc, CD, ROM, PROM, and EPROM, EEPROM, or flash memory) and thus executed by a processor or programmable hardware, whether or not such a processor or programmable hardware is explicitly shown. The methods disclosed in the specification or claims may be implemented by a device having means for performing each of the corresponding actions of the methods.

[0074] It should be understood that the disclosure of multiple actions, processes, operations, steps or functions disclosed in the specification or claims may not be interpreted as being in a specific order unless otherwise stated explicitly or implicitly, for example, for technical reasons. Therefore, the disclosure of multiple actions or functions does not limit these to a specific order unless these actions or functions are not interchangeable for technical reasons. In addition, in some examples, a single action, function, process, operation or step may include or may be divided into multiple sub-actions, multiple sub-functions, multiple sub-processes, multiple sub-operations or multiple sub-steps, respectively. Unless explicitly excluded, such sub-actions and a portion of the disclosure of the single action may be included.

[0075] In addition, the following claims are hereby incorporated into the detailed description, wherein each claim can stand on its own as a separate example. Although each claim can stand on its own as a separate example, it should be noted that although a dependent claim can refer to a specific combination with one or more other claims in a claim, other examples can also include a combination of a dependent claim with the subject matter of each other dependent claim or independent claim. Such combinations are explicitly proposed herein unless it is stated that a specific combination is not intended. In addition, even if a claim is not directly dependent on an independent claim, it is intended that the features of that claim be included in any other independent claim.

Claims

1. A method (100) for determining a position of an unmanned vehicle UV, the method comprising: receiving (102) positioning signals from a plurality of satellites in a global navigation satellite system; as well as determining (106) a three-dimensional model of the environment of the UV based on sensor data of one or more sensors of the UV, wherein the one or more sensors of the UV include at least an optical sensor and an inertial measurement unit, wherein a simultaneous localization and mapping (SLAM) algorithm is used to determine a three-dimensional representation of the environment of the UV based on the sensor data of the optical sensor and the inertial measurement unit, classifying at least one object of the environment of the UV represented by the sensor data of the optical sensor into one of a plurality of predefined categories based on the sensor data of the optical sensor, wherein the plurality of predefined categories are radio frequency characteristics and / or material categories, and semantically labeling the at least one object in the three-dimensional representation of the environment of the UV with the classified category so as to obtain the three-dimensional model of the environment of the UV; estimating (104) the position of the UV based on (i) a three-dimensional model of the environment of the UV and (ii) possible signal paths of each of the positioning signals, wherein at least a portion of the possible signal paths of each of the positioning signals includes reflections of the positioning signal by one or more objects in the environment of the UV, wherein the possible signal paths of each of the positioning signals are represented by a particle filter, and wherein the particle filter receives as input a pose estimate and a velocity estimate of the UV from the SLAM algorithm, the three-dimensional model of the environment of the UV, the positions of the satellites, and information related to the positioning signals; and A respective most probable signal path among the possible signal paths is determined for each of the positioning signals using the particle filter.

2. The method according to claim 1, wherein Objects in the three-dimensional model of the UV environment are semantically tagged with information about radio frequency characteristics and / or material categories of the respective objects.

3. The method according to claim 1, wherein The three-dimensional model of the UV environment includes only static objects.

4. The method according to claim 1, wherein The three-dimensional model of the UV environment includes at least one moving object.

5. The method according to claim 1, wherein The method also includes updating a pose estimate and a velocity estimate of the SLAM algorithm based on the corresponding most probable ones of the possible signal paths for each of the positioning signals.

6. The method according to claim 5, wherein: Estimating the position of the UV includes determining the position indicated by the updated pose estimate of the SLAM algorithm as the position of the UV.

7. The method according to claim 5, wherein: The method also includes updating the three-dimensional model of the UV's environment based on the updated pose estimate and the updated velocity estimate of the SLAM algorithm.

8. The method according to claim 1, wherein The method is performed by the UV.

9. The method according to claim 1, wherein The UV is an unmanned aerial vehicle (UAV).

10. The method according to claim 9, wherein: The UAV is a multi-rotor drone.

11. A non-transitory machine-readable medium having a program stored thereon, the program having program code, which, when executed on a processor or programmable hardware of an unmanned vehicle, is used to control the unmanned vehicle to perform the method for determining the position of the unmanned vehicle according to claim 1.

12. A device (610) for determining a position of an unmanned vehicle (UV), the device comprising: a receiver circuit (611) configured to receive positioning signals from a plurality of satellites of a global navigation satellite system; as well as The processing circuit (612) is configured to: determining (106) a three-dimensional model of the environment of the UV based on sensor data of one or more sensors of the UV, wherein the one or more sensors of the UV include at least an optical sensor and an inertial measurement unit, wherein a simultaneous localization and mapping (SLAM) algorithm is used to determine a three-dimensional representation of the environment of the UV based on the sensor data of the optical sensor and the inertial measurement unit, classifying at least one object of the environment of the UV represented by the sensor data of the optical sensor into one of a plurality of predefined categories based on the sensor data of the optical sensor, wherein the plurality of predefined categories are radio frequency characteristics and / or material categories, and semantically labeling the at least one object in the three-dimensional representation of the environment of the UV with the classified category so as to obtain the three-dimensional model of the environment of the UV; estimating the position of the UV based on (i) a three-dimensional model of the environment of the UV and (ii) possible signal paths of each of the positioning signals, wherein at least a portion of the possible signal paths of each of the positioning signals includes reflections of the positioning signal by one or more objects in the environment of the UV, wherein the possible signal paths of each of the positioning signals are represented by a particle filter, and wherein the particle filter receives as input a pose estimate and a velocity estimate of the UV from the SLAM algorithm, the three-dimensional model of the environment of the UV, the positions of the satellites, and information related to the positioning signals; and A respective most probable signal path among the possible signal paths is determined for each of the positioning signals using the particle filter.

13. An unmanned aerial vehicle (600) comprising the device (610) for determining the position of the unmanned vehicle according to claim 12.

14. The unmanned aerial vehicle according to claim 13, wherein: The unmanned aerial vehicle (600) is a multi-rotor drone.

Images