SYSTEM AND METHOD FOR TESTING A SURFACE OF AN AIR OR SPACE VEHICLE

Description

Technical area

[0001] The present description relates to a system and a method for inspecting a surface of an aircraft or spacecraft. Technical background

[0002] In the manufacture of fuselage structures, for example for aircraft or spacecraft, a high surface quality is desired. If the fuselage has riveted joints, for example, it is common practice to inspect the quality of the riveted joints visually and / or tactilely and to measure them selectively using measuring tools. To do this, individual riveted joints are illuminated from the side, and a shadow image is used to check whether a rivet is seated too deep or too high in a rivet hole or is otherwise conspicuous. It is also common practice to run a finger over a rivet head and / or use a dial indicator to detect protruding or overly recessed edges.

[0003] In addition, surface scanning devices are known that use light projection and image capture techniques to examine surface sections for specific features. These devices are typically handheld and feature a foam frame that is placed on the surface section to be inspected, followed by an optical scan of the area enclosed by the foam frame. If points with shape deviations are detected, information about them is stored. This is done primarily in the form of image information, in which the respective conspicuous points are marked accordingly.

[0004] CN 211 055 414 U discloses an autonomous vehicle for three-dimensional measurement of an aircraft comprising a multi-part gripper arm with a triangulation sensor.

[0005] For the inspection of longer rows of rivets on the fuselage of an aircraft or spacecraft, which may include hundreds or more rivet joints, such an inspection is time-consuming. Furthermore, the surface structure can also be checked for other anomalies, such as dents or scratches. Description

[0006] The task can be considered to be to propose a system and method that is as simple as possible and with which an inspection of a surface of an aircraft or spacecraft can be carried out automatically as quickly and reliably as possible, while protecting the integrity of the aircraft or spacecraft and the surrounding assembly environment from damage.

[0007] This object is achieved by the system having the features of independent claim 1. Advantageous embodiments and further developments emerge from the subclaims and the following description.

[0008] A system for testing a surface structure of an aircraft or spacecraft is proposed, comprising a ground vehicle with a drive, a manipulator arranged on the ground vehicle with a plurality of movement members arranged in an articulated manner relative to one another, an inspection device arranged on the manipulator and movable by the manipulator, a plurality of proximity sensors, and a control unit, wherein the proximity sensors are distributed at least on the manipulator and are divided therein into a plurality of zones and are designed to detect a distance of the proximity sensor to an object in a respective detection area, wherein the drive and the proximity sensors are coupled to the control unit, wherein the inspection device is designed to determine a property of the surface structure and to identify points with a shape deviation by comparison with a target property, wherein the system is designed toto successively inspect the surface structure of the aircraft or spacecraft by successively moving the inspection device along the surface structure, and wherein the control unit is designed to record distances detected by the proximity sensors and to control the drive and the manipulator taking into account the detected distances such that the inspection device is successively placed at inspection positions at a predetermined distance from the surface structure and, at the same time, a sufficient distance of the ground vehicle and the manipulator from the surface structure and any objects surrounding the surface structure is maintained.

[0009] The ground vehicle is designed to move the manipulator and the inspection device mounted on it along a floor at an assembly or construction site, along the surface to be inspected. It is conceivable that the ground vehicle could be parked in another room and driven from this room to the assembly or construction site. The ground vehicle could therefore preferably be designed to cover longer distances, which could encompass several times the size of the surface to be inspected. The ground vehicle could be designed to be as compact as possible and, preferably, self-sufficient. The ground vehicle could, for example, be designed as a standardized, driverless transport vehicle (UGV, "unmanned ground vehicle"), which can be understood as a mobile robot.These types of transport systems are familiar in industrial settings and are used, among other things, to move pallets or other objects automatically along floor lines or with the assistance of other navigation aids. The drive is preferably electric, and the ground vehicle can have an electrical energy storage unit that can be connected to the drive and other components.

[0010] The manipulator can have multiple movement elements, which are, for example, rod-shaped and connected to one another in an articulated manner. The joints are preferably actively rotatable in order to be able to move the movement elements relative to one another. The inspection device is provided as the end effector on the manipulator. The manipulator has a first end arranged on the ground vehicle. An opposite second end has the end effector or the inspection device. For example, the manipulator can have one, two, three, four, or more movement elements that form a kinematic chain and are each connected via a joint to one of the other movement elements or to the ground vehicle. It is preferred to design the movement elements or the joints so that they can be moved independently of one another, with the control unit being connected to all actuators in order to control them and to move or align the inspection device as desired.

[0011] The inspection device can be implemented in different ways. Preferably, as explained above, it can be designed as an optical scanning device to inspect certain features non-contact by capturing and evaluating an image of a projected pattern of visible light. Such an inspection device is known and is typically used for inspecting aircraft surface structures. However, mechanical scanning devices or scanning devices based on invisible light are also conceivable.

[0012] The proximity sensors are designed to detect a distance to an object, preferably without contact. Such a proximity sensor could, for example, be implemented as an ultrasonic sensor that emits ultrasonic signals and receives their reflections. Inductive proximity sensors could also be used for metallic objects. Capacitive proximity sensors or optical sensors are also conceivable, the latter being feasible using laser or camera systems. Each proximity sensor has a design-specific detection range and detection direction.

[0013] A core of the invention lies in enabling improved motion control and thus faster movement of the inspection device at a construction or assembly site. When the ground vehicle and / or the manipulator approach an object, not only is a stop signal generated by a proximity sensor that detects the object in question. Instead, the entirety of the movements of the individual movement elements and the ground vehicle is controlled in such a way that the inspection device can continue to move even when approaching the object. This is achieved by maintaining the individual distances between the individual components of the ground vehicle and the manipulator at a predetermined (minimum) value and only restricting the mobility or speed of individual components as needed.By dividing the system into zones, multiple areas of the manipulator and ground vehicle can be individually monitored to plan local movement restrictions. This enables intelligent, predictive motion control, which allows for rapid sequencing of inspection positions, even when obstacles are not known in advance, and accelerates surface structure testing. The system can be used during manufacturing, maintenance, and ongoing operations, for example, between two consecutive flights.

[0014] In an advantageous embodiment, the control unit is designed to move those elements of the manipulator that are closer to an object or the surface structure than other elements at a lower speed than the other elements. Consequently, an adaptive movement of the ground vehicle and the manipulator is carried out in order to carry out the testing of the surface structure smoothly. Moving past obstacles or other objects therefore takes place in such a way that areas on the ground vehicle or manipulator with the highest collision probability, i.e. with the shortest distance to an obstacle or another object, are moved the slowest. The control unit can be controlled in such a way that movement sequences of the ground vehicle or the manipulator are planned in such a way that further approach to an object or obstacle is prevented.

[0015] In an advantageous embodiment, the zones comprise an end effector zone that is adjacent to the inspection device or in whose area the inspection device is arranged, wherein the control unit is configured to support the placement of the inspection device at one of the inspection positions based on the distances detected by the proximity sensors in the end effector zone. A subset of the detected distances is used to precisely position the inspection device on the surface structure to be inspected. Dedicated proximity sensors intended exclusively for positioning the inspection device on the surface structure are not necessary. However, it is also possible for the inspection device itself to be equipped with distance sensors that allow the inspection device to be positioned on the surface structure.The end effector zone could be located at or adjacent to a joint where the inspection device is located. It is conceivable that the detection ranges of the proximity sensors arranged in the end effector zone are at least partially adapted to the orientation of the inspection device.

[0016] In an advantageous embodiment, the control unit is designed to plan a movement path of the inspection device to reach the inspection positions, with the distances detected by the proximity sensors being weighted differently. The planned movement path could, in particular, be made dependent on the shorter detected distances, for example, by applying greater weighting. This can support maintaining a minimum distance between all system components and an obstacle or other object.

[0017] In an advantageous embodiment, the control unit is designed to use at least one machine learning model to plan the movement path for the ground vehicle and / or the manipulator, wherein the movement path is derived at least partially on the basis of a value function which comprises the detected distances and describes collision probabilities between the ground vehicle or the manipulator and the surface structure and other objects, and wherein the movement path comprises a sequence of movement sequences of the ground vehicle and / or the manipulator. The sequence of movement sequences could in particular relate to the entirety of all partial movements of all actuators of the manipulator or the drive of the ground vehicle which serve to place the inspection device at the inspection positions. The machine learning enables the system toThe aim of the system is to continuously improve the efficiency of the system and to increase the speed of testing. The control unit can feed the measured distances into the machine learning algorithm for analysis. Since machine learning algorithms are probabilistic in nature, the control unit can determine at least one statistical feature from the measured distances and calculate at least one status indicator from the at least one statistical feature, which can then be used to assess the probabilities of various undesirable states, i.e. a collision with the surface to be tested, equipment in the assembly or spacecraft.The construction site, the system itself, another similar system (if several are used), a human, or the like, are used, which are presented to the machine learning algorithm during training. The detection of an undesirable condition can then be interpreted as exceeding a previously defined threshold probability.

[0018] In an advantageous embodiment, the control unit is designed to sequentially control the proximity sensors to avoid mutual interference. This can, for example, involve the sequential control of proximity sensors whose detection ranges partially overlap. By avoiding mutual interference, the accuracy of distance detection is significantly increased. For example, if several proximity sensors are distributed along the moving elements of the manipulator, each with a radially outward-facing detection range spanning a plane perpendicular to the respective moving element, it might be advisable to simultaneously control only proximity sensors that are opposite one another rather than adjacent ones.

[0019] In an advantageous embodiment, the manipulator is a robot arm connected to the ground vehicle via a base joint, wherein the movement members are elongated arm members, and wherein the manipulator comprises an end effector joint on which the inspection device is arranged. The base joint and / or the end effector joint are each rotatable about at least two axes and allow a high degree of mobility of the inspection device in order to place the inspection device at a desired inspection position depending on the position of the ground vehicle and simultaneously align it in the desired manner.

[0020] In an advantageous embodiment, the manipulator comprises a network of electrical connecting cables that runs at least partially along a surface of the manipulator. The proximity sensors are arranged on the surface of the manipulator and can be connected to the network. The proximity sensors can thus be easily replaced, removed, or retrofitted. Depending on the type and design of the manipulator, it may be advisable to make certain modifications to enable improved motion planning. The network could extend over a significant portion of the manipulator and allows for very flexible equipping of the manipulator with proximity sensors.

[0021] In an advantageous embodiment, the connecting lines are designed to provide an electrical supply voltage for the proximity sensors, with the proximity sensors being designed to communicate wirelessly with the control unit. The network can therefore be very simple in design and limited to the distribution of two voltage poles on the surface. Individual proximity sensors could simply be glued to the surface of the manipulator to be supplied with voltage there.

[0022] In an advantageous embodiment, the system is designed to mark an identified point with a shape deviation using a removable marker. The marker could comprise a removable ink that is used to circle the shape deviation. A sticker, optionally printed with a barcode or QR code and allowing the provision of additional information, could be affixed to or next to the shape deviation. Attaching an RFID tag is also conceivable.

[0023] The system is preferably designed to generate data representing the surface to be scanned and / or at least detected anomalies in the surface structure. For this purpose, the system is capable of precisely determining the position of the respective anomaly, for example, using a positioning system installed at the assembly or construction site that enables position determination, such as a marking grid on the floor or surface structure, a laser system, or similar.

[0024] The invention further relates to a method for testing a surface structure of an aircraft or spacecraft, comprising the steps of providing at least one ground vehicle with a drive, a manipulator arranged on the ground vehicle with a plurality of movement members arranged in an articulated manner relative to one another, an inspection device arranged on the manipulator and movable by the manipulator, and a plurality of proximity sensors which are distributed at least on the manipulator and are divided therein into a plurality of zones and are designed to detect a distance of the proximity sensor to an object in a respective detection range;and successively moving the inspection device along the surface structure by controlling it with a control unit, and identifying points with a shape deviation by determining a property of the surface structure and comparing it with a target property, wherein the controlling comprises recording distances detected by the proximity sensors and controlling the drive and the manipulator taking into account the detected distances, so that the inspection device is placed one after the other at inspection positions at a predetermined distance from the surface structure and at the same time a sufficient distance of the ground vehicle and the manipulator from the surface structure and any objects surrounding the surface structure is maintained.

[0025] In an advantageous embodiment, the control unit controls those elements of the manipulator that are located closer to an object or the surface structure in such a way that they move at a lower speed than other elements.

[0026] In an advantageous embodiment, the control unit plans a movement path of the inspection device to reach the inspection positions, wherein the detected distances of the proximity sensors are weighted differently.

[0027] In an advantageous embodiment, the control unit uses at least one machine learning model to plan the movement path for the ground vehicle and / or the manipulator, wherein the movement path is derived at least partially based on a value function that includes the detected distances and describes collision probabilities therefrom, and wherein the movement path comprises a sequence of movement sequences of the ground vehicle and / or the manipulator.

[0028] The system and method according to the invention are fundamentally designed to inspect the surface structure of multiple sections simultaneously. For this purpose, multiple systems, i.e., multiple ground vehicles with a manipulator and inspection device mounted thereon, can be used. The respective control unit is also capable of preventing a collision between one system and another thanks to the large number of proximity sensors.

[0029] It is fundamentally conceivable that the control units of the multiple systems can be linked together in order to mutually transmit the positions of the individual ground vehicles and manipulators in order to proactively prevent a collision.

[0030] At the same time, it is conceivable that distances recorded by one of the systems are also transmitted to control units of other systems with a corresponding position information of the system recording the distances, so that a proactive planning of movement sequences is enabled.

[0031] In an advantageous embodiment of the method, a plurality of ground vehicles with a manipulator and inspection device arranged thereon, or at least one ground vehicle and at least one other device carrying a manipulator and an inspection device and having a control unit, could jointly inspect a plurality of sections of the surface structure, wherein the control units can be coupled to one another in order to communicate position information of the respective ground vehicle or the other device and / or the respective manipulator and / or the respective inspection device, and / or in order to communicate respectively detected distances with position information of the respective ground vehicle or the other device and / or the respective manipulator and / or the respective inspection device.The "other device" could also be understood as a vehicle that is not necessarily ground-bound and can be moved along the surface structure, for example, on a cable, a rail arranged above the ground, or via buoyancy means. Short description of the characters

[0032] The following examples are described in more detail with reference to the accompanying drawings. The illustrations are schematic and not to scale. Like reference numerals refer to like or similar elements. They show: Fig. 1 is a schematic representation of a system according to an embodiment for scanning a surface of an aircraft. Fig. 2 is a schematic representation of a sequence of inspection procedures on the aircraft. Fig. 3 is a schematic view of the system. Fig. 4 is a schematic view of proximity sensors distributed over the circumference of an arm member. Fig. 5 is a schematic representation of a network of connecting cables. Detailed description of implementation examples

[0033] Fig. 1 shows a system 2 for testing a surface structure 4 of an aircraft 6. The surface structure 4 comprises, for example, rows of rivets on a fuselage surface, for example between, above or below windows 8. The system 2 Fig. 1 The area shown is merely exemplary for illustration purposes and is not to be understood as a limitation. The aircraft 6 is arranged at an assembly or construction site 10, where a floor 12 is provided that extends along the surface structure 4. Located there is a ground vehicle 14 belonging to the system 2, on which a manipulator 16 is arranged.

[0034] The ground vehicle 14 is exemplified as a standardized, driverless transport vehicle (UGV, "unmanned ground vehicle") and can preferably move autonomously on the ground 12. For mechanical or optical guidance along a lane, a ground rail 18 is provided, extending linearly along the surface structure 4 in or on the ground 12. A railing 20 is provided to demarcate the ground 12 at the assembly or construction site 10 from the aircraft 6. This railing could have an irregular design depending on the equipment of the assembly or construction site 10. For example, individual doors for entering the fuselage of the aircraft 6 could be provided.

[0035] The ground vehicle 14 has a drive 22 and can move along the ground 12. The surface structure 4 of the aircraft 6 is inspected using an inspection device 24 arranged on the manipulator 16. In this case, the inspection device 24 is a device that detects shape deviations in the surface structure 4. This is achieved by projecting a pattern of visible light onto the surface structure at a predetermined distance from the surface structure 4, and capturing and evaluating a resulting image 26. Such inspection devices 26 are known for scanning surface structures, but are typically used handheld.

[0036] As in Fig. 1 As is clear, only small sections of the surface structure 4 are examined by the inspection device 24, so that inspection positions on the surface structure 4 must be approached one after the other in order to be able to capture the relevant area of ​​the surface structure 4 as a whole. For this purpose, the ground vehicle 14 has a control unit 28 which is designed to record distances detected by a plurality of proximity sensors 30 and to control the drive 22 and the manipulator 16, taking the detected distances into account, in such a way that the inspection device 24 is placed one after the other at inspection positions at a predetermined distance from the surface structure 4 and, at the same time, a sufficient distance between the ground vehicle 14 and the manipulator 16 and the surface structure 4 and objects 20 surrounding the surface structure 4 is maintained.

[0037] A plurality of proximity sensors 30 are provided, which are distributed over the manipulator 16 and divided into a plurality of zones 32a, 32b, 32c, and 32d. The manipulator 16 has, for example, two rod-shaped arm members 34, which are connected via a joint 36. The manipulator 16 is connected to the ground vehicle 14 via a base joint 38. An end effector joint 40 is arranged at a free end of the manipulator 16 and carries the inspection device 24. The proximity sensors 30 are arranged in circumferential rows on the arm members 34 and each have a detection range that is directed radially outwards. This will be explained further below with reference to Fig. 3 und 4 explained. The control unit 28 is coupled to the proximity sensors 30, the drive 28, and the manipulator 16.

[0038] Fig. 2 shows the surface structure 4 of the aircraft 6 in a schematic partial view, with the system 2 indicated. Individual inspection positions 42 are shown here, which are successively traversed by the system 2 in order to successively inspect an interesting area of ​​the surface structure 4 by locally scanning the surface structure 4. Only three superimposed rows of inspection positions 42 are shown here as an example, which are successively traversed one after the other to carry out a local inspection there.

[0039] This may mean, for example, that the system 2 is moved along the entire area of ​​the surface structure 4 three times, each time maintaining the same height of the inspection device 24. Initially, the system 2 starts at a first starting position 44a and moves to a first end position 44b. The inspection device 24 is then lowered slightly to move from a second starting position 44c located below the first end position 44b to a second end position 44d located below the first starting position 44a. The inspection device 24 is then lowered again to move from a third starting position 44e located below the second end position 44d to a third end position 44f located below the second starting position 44c. Of course, other sequences are possible, and the inspection positions 42 could also be moved column by column.

[0040] Throughout the entire movement of system 2 on floor 12, proximity sensors 30 sense their surroundings and transmit the detected distances to control unit 28. This can be done in the form of measured variables that directly indicate the distance. However, it is also conceivable for control unit 28 to be equipped with evaluation electronics, filters, and the like to convert abstract signals from proximity sensors 30 into corresponding distance values. By weighting the distances detected by all proximity sensors 30, control unit 28 can control drive 22 and manipulator 16 in order to maintain a minimum distance from both surface structure 4 and surrounding objects 20.

[0041] This can prevent damage to the inspection device 24 and the manipulator 16 as well as to the objects 20 and the surface structure 4.

[0042] The control unit 28 is designed, by way of example, to use at least one model for machine learning to determine the movement path for the ground vehicle 14 and / or the manipulator 16. The movement path can be derived at least partially on the basis of a value function that includes the detected distances and describes collision probabilities therefrom.

[0043] Fig. 3 shows system 2 again, wherein a plurality of individual proximity sensors 30 are provided. These could, for example, also protrude onto or above the inspection device 24 and form an end effector zone 46 there. This could assist the control unit 28 in precisely positioning the inspection device 24 at the desired distance from the surface structure 4.

[0044] Fig. 4 shows a very schematic sectional view of an arm link 34. Here, for example, six proximity sensors 30 are distributed in a common plane over the circumference of the arm link, each with a detection area 48. This detection area is, for example, conically directed radially outward. Due to the dense arrangement of the proximity sensors 30, it is advisable to control them sequentially so that they do not influence each other.

[0045] Based on Fig. 4 Three pairs I, II, and III of proximity sensors 30 are shown as examples, each consisting of two diametrically opposed proximity sensors 30. These three pairs I, II, and III are controlled sequentially, for example, in order to detect two distances one after the other. This ensures that the detection areas 48 do not overlap.

[0046] Fig. 5Finally, FIG. 1 shows a section of a surface of an arm member 34, on which a network 50 of electrical connecting lines 52 and 54 is provided. Here, first connecting lines 52 are connected, for example, to a negative voltage pole, while second connecting lines 54 are each connected to a positive voltage pole. These follow one another multiple times and in an alternating sequence, thereby forming the network 50. Individual proximity sensors 30 can each be arranged on a pair of two connecting lines 52 and 54 in order to be supplied with a voltage.

[0047] It is conceivable that the proximity sensors 30 then communicate wirelessly with the control unit 28. This can be done directly or with the aid of a router 56. This allows the manipulator 16 and, if necessary, also the ground vehicle 14 to be equipped with proximity sensors 30 particularly easily and flexibly. Their placement can be modified if necessary. Particularly when using a machine learning algorithm, the system 2 is capable of detecting even relocated proximity sensors 30 and adapting the control system accordingly.

[0048] Additionally, it should be noted that "comprising" or "having" does not exclude other elements or steps, and "a" or "an" does not exclude a plurality. Furthermore, it should be noted that features or steps described with reference to one of the above embodiments may also be used in combination with other features or steps of other embodiments described above. Reference signs in the claims are not to be considered as limitations. List of reference symbols

[0049] 2System 4Surface structure 6Aircraft 8Window 10Assembly or construction site 12Ground 14Ground vehicle 16Manipulator 18Ground rail 20Railing / object 22Drive 24Inspection device 26Image 28Control unit 30Proximity sensor 32Zone 34Arm link 36Joint 38Base joint 40End effector joint 42Inspection position 44Start position / end position 46End effector zone 48Detection area 50Network 52Connecting cable 54Connecting cable 56Router I, II, IIIPairs of proximity sensors

Claims

1. System (2) for inspecting a surface structure (4) of an aircraft or spacecraft, comprising: a ground vehicle (14) with a drive (22), a manipulator (16) arranged on the ground vehicle (14) with several articulately arranged movement members (34), an inspection device (24) arranged on the manipulator (16) and movable by the manipulator (16), several proximity sensors (30), and a control unit (28), wherein the proximity sensors (30) are distributed at least on the manipulator (16) and are divided there into several zones (32, 46) and are designed to detect a distance of the proximity sensor (30) to an object in a respective detection area (48), wherein the drive (22) and the proximity sensors (30) are coupled to the control unit (28), wherein the inspection device (24) is configured to determine a property of the surface structure (4) and to identify points with a shape deviation by comparison with a target property, wherein the system (2) is designed to successively inspect the surface structure (4) of the aircraft or spacecraft by successively moving the inspection device (24) along the surface structure (4), and wherein the control unit (28) is designed to record distances detected by the proximity sensors (30), and to control the drive (22) and the manipulator (16) taking into account the detected distances such that the inspection device (24) is successively placed at inspection positions (42) at a predetermined distance from the surface structure (4) and at the same time a sufficient distance of the ground vehicle (14) and the manipulator (16) from the surface structure (4) and any objects surrounding the surface structure (4) is maintained.

2. System (2) according to claim 1, wherein the control unit (28) is designed to move those elements of the manipulator (16) that are closer to an object or the surface structure (4) than other elements, at a lower speed than the other elements.

3. System (2) according to claim 1 or 2, wherein the zones (32, 46) comprise an end effector zone (46) that adjoins the inspection device (24) or in whose area the inspection device (24) is arranged, wherein the control unit (28) is designed to support the placement of the inspection device (24) at one of the inspection positions (42) from detected distances of the proximity sensors (30) in the end effector zone (46).

4. System (2) according to one of the preceding claims, wherein the control unit (28) is designed to plan a movement path of the inspection device (24) for reaching the inspection positions (42), wherein the detected distances of the proximity sensors (30) are weighted differently.

5. System (2) according to claim 4, wherein the control unit (28) is designed to use at least one machine learning model to plan the movement path for the ground vehicle (14) and / or the manipulator (16), wherein the movement path is derived at least partially based on a value function that includes the detected distances and describes collision probabilities between the ground vehicle (14) or the manipulator (16) and the surface structure (4) and other objects, and wherein the movement path comprises a sequence of movements of the ground vehicle (14) and / or the manipulator (16).

6. System (2) according to one of the preceding claims, wherein the control unit (28) is designed to sequentially activate the proximity sensors (30) to avoid mutual interference.

7. System (2) according to one of the preceding claims, wherein the manipulator (16) is a robotic arm that is connected to the ground vehicle (14) via a base joint (38), wherein the movement members (34) are elongated arm members (34), and wherein the manipulator (16) comprises an end effector joint (40) on which the inspection device (24) is arranged.

8. System (2) according to one of the preceding claims, wherein the manipulator (16) comprises a network (50) of electrical connection lines (52, 54) that runs at least partially on a surface of the manipulator (16), and wherein the proximity sensors (30) are arranged on the surface of the manipulator (16) and can be connected to the network (50).

9. System (2) according to claim 8, wherein the connection lines (52, 54) are designed to provide an electrical supply voltage for the proximity sensors (30), and wherein the proximity sensors (30) are designed to communicate wirelessly with the control unit (28).

10. System (2) according to one of the preceding claims, wherein the system (2) is designed to mark an identified point with a shape deviation by means of a removable marking.

11. Method for inspecting a surface structure (4) of an aircraft or spacecraft, comprising the steps of: providing at least one ground vehicle (14) with a drive (22), a manipulator (16) arranged on the ground vehicle (14) with several articulately arranged movement members, an inspection device (24) arranged on the manipulator (16) and movable by the manipulator (16), and several proximity sensors (30), which are distributed at least on the manipulator (16) and are divided there into several zones (32, 46) and are designed to detect a distance of the proximity sensor (30) to an object in a respective detection area, and successively moving the inspection device (24) along the surface structure (4) by controlling with a control unit (28), and identifying points with a shape deviation by determining a property of the surface structure (4) and comparing it with a target property, wherein the controlling comprises recording distances detected by the proximity sensors (30) and controlling the drive (22) and the manipulator (16) taking into account the detected distances, so that the inspection device (24) is successively placed at inspection positions at a predetermined distance from the surface structure (4) and at the same time a sufficient distance of the ground vehicle (14) and the manipulator (16) from the surface structure (4) and any objects surrounding the surface structure (4) is maintained.

12. Method according to claim 11, wherein the control unit (28) controls those elements of the manipulator (16) that are closer to an object or the surface structure (4) such that they move at a lower speed than other elements.

13. Method according to claim 11 or 12, wherein the control unit (28) plans a movement path of the inspection device (24) for reaching the inspection positions, wherein the detected distances of the proximity sensors (30) are weighted differently.

14. Method according to claim 13, wherein the control unit (28) uses at least one machine learning model to plan the movement path for the ground vehicle (14) and / or the manipulator (16), wherein the movement path is derived at least partially based on a value function that includes the detected distances and describes collision probabilities, and wherein the movement path comprises a sequence of movements of the ground vehicle (14) and / or the manipulator (16).

15. Method according to one of claims 11 to 14, wherein several ground vehicles (14) with a manipulator (16) and inspection device (24) arranged thereon, or at least one ground vehicle (14) and at least one other device carrying a manipulator (16) and an inspection device (24) and having a control unit (28), jointly inspect several sections of the surface structure (4), and wherein the control units (28) can be coupled to each other to communicate position data of the respective ground vehicle (14) or the other device and / or the respective manipulator (16) and / or the respective inspection device (24), and / or to communicate respectively detected distances with position data of the respective ground vehicle (14) or the other device and / or the respective manipulator (16) and / or the respective inspection device (24).