Multi-mode parking inspection system and method based on unmanned aerial vehicle path planning

Abstract

The invention discloses a multi-mode parking inspection system and method based on unmanned aerial vehicle path planning, and belongs to the technical field of intelligent traffic and unmanned aerial vehicle application, the multi-mode parking inspection system comprises an unmanned aerial vehicle, an unmanned trolley and a control center, and the bottom of the unmanned aerial vehicle is provided with a lock catch mechanism; a butt joint platform is arranged on the top of the unmanned trolley, and the butt joint platform is provided with a hook lug matched with the lock catch mechanism; the unmanned aerial vehicle and the unmanned trolley are in separable hard connection through cooperation of the clamping hooks and the hook lugs, and three working modes are formed. The control center is provided with a dynamic path planning module, and generates an optimal inspection path based on environmental data; and the control center sends an inspection instruction to the unmanned aerial vehicle and the unmanned trolley according to the generated optimal inspection path. Through mutual combination of the unmanned aerial vehicle and the unmanned trolley, multi-mode switching is realized, the optimal inspection path is generated based on the environment data, the instruction is sent to the equipment, and the regional inspection task is completed, so that the technical problems of low efficiency, insufficient endurance and the like of the traditional inspection equipment are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent transportation and unmanned aerial vehicle (UAV) applications, and particularly relates to a multimodal parking inspection system and method based on UAV path planning. Background Art

[0002] With the accelerated pace of urbanization, the number of vehicles in cities has increased dramatically, and on-street parking resources are becoming increasingly scarce, highlighting the contradiction between parking supply and demand. Efficient management of parking resources has become a crucial component of optimizing urban transportation. However, manual inspection methods are costly, inefficient, and subject to time and space constraints. Therefore, how to automate parking inspections using unmanned equipment to monitor parking status in real time and improve management efficiency has become a key issue that the industry urgently needs to address. Unmanned inspections not only reduce reliance on manpower but also cover complex environments, such as multi-story parking lots or densely populated neighborhoods, providing technical support for the development of smart cities.

[0003] In the prior art, relevant patent documents have proposed the use of drones or unmanned vehicles in parking inspections. For example, Chinese patent document with publication number CN113593057 A discloses a method for managing on-street parking spaces based on drone inspections. The method includes: inputting multiple parking space images acquired by drones into a target vehicle detection network model to obtain parking space images containing target vehicles and the location of target vehicles, and then calculating parking fees based on entry and exit times, thereby achieving vehicle record management without manual monitoring. Another Chinese patent document with publication number CN113436361A provides a roadside parking space management system based on an unmanned inspection vehicle. The system includes a vehicle-machine operation module, a communication module, a charging module, a camera module, a positioning module, a safety recorder module, an ultrasonic module, and a parking vehicle management module. By integrating multiple modules, it achieves all-weather and fully automatic parking space monitoring without environmental restrictions. However, the above technologies have significant flaws in practical applications: when traditional drones stop for inspection, they are often blocked by low-altitude interference sources such as trees and billboards, which can easily obstruct the drone's inspection path and affect the continuity of the inspection. In addition, accessories such as high-resolution cameras and lidars further shorten the drone's flight time, requiring frequent return flights for recharging, which significantly reduces operational efficiency. On the other hand, unmanned vehicle parking inspections rely on preset maps and fixed paths, lack high-level perspective support, and cannot dynamically optimize paths based on real-time environmental changes, such as traffic dynamics or obstacle displacement, resulting in mission interruptions or repeated inspections. In addition, unmanned vehicles have limited obstacle-crossing capabilities, making it difficult to quickly navigate complex scenarios such as high slopes, resulting in low inspection efficiency.

[0004] In summary, existing technologies face challenges such as low inspection efficiency and insufficient battery life in complex parking scenarios. Therefore, there is an urgent need for a multimodal parking inspection system and method based on dynamic path planning. This system integrates the advantages of drones and unmanned vehicles, improves overall performance through multimodal collaborative operation, and achieves efficient and accurate parking space management. Summary of the Invention

[0005] In response to the problems in related technologies, the present invention proposes a multimodal parking inspection system and method based on drone path planning. By combining drones and unmanned vehicles, multimodal switching is achieved, and the optimal inspection path is generated based on environmental data and instructions are sent to the equipment to complete the regional inspection task, so as to overcome the technical problems faced by existing related technologies in complex parking scenarios, such as low inspection efficiency and insufficient battery life.

[0006] The technical solution of the present invention is implemented as follows: a multimodal parking inspection system based on drone path planning, comprising a drone, an unmanned vehicle, and a control center. The bottom of the drone is provided with a locking mechanism; the locking mechanism includes a hook and a drive unit that drives the hook to rotate around its rotation axis; a first charging module is provided next to the locking mechanism for docking with the unmanned vehicle for charging; the drone is also equipped with an RTK positioning module, a visual sensor, and a laser radar; wherein the RTK positioning module is used for spatial positioning; the visual sensor is used for inspection and evidence collection; and the laser radar is used for real-time detection of terrain height changes.

[0007] The unmanned vehicle is equipped with a battery pack, which is connected to a second charging module; the second charging module is configured to be compatible with the first charging module and to charge and supply energy to the first charging module; a docking platform is provided on the top of the unmanned vehicle, which is provided with a hook ear that matches the locking mechanism; the drone and the unmanned vehicle are detachably connected through the cooperation of the hook and the hook ear, forming three working modes:

[0008] Mode 1: The drone and the unmanned vehicle are separated and perform flight inspections independently.

[0009] Second mode: After landing, the drone is fixed to the docking platform of the unmanned vehicle, and the unmanned vehicle carries it for charging and performs ground inspections.

[0010] The third mode: the drone carries out vertical inspection by hanging an unmanned vehicle;

[0011] The control center is equipped with a dynamic path planning module to generate an optimal inspection path based on environmental data; the environmental data includes the tree coverage rate, traffic density and terrain height change rate of the inspection area; the control center sends inspection instructions to drones and unmanned vehicles based on the generated optimal inspection path.

[0012] The present invention combines drones and unmanned vehicles to achieve multi-modal switching, generate optimal inspection paths based on environmental data, and send instructions to the equipment to complete regional inspection tasks, thereby solving technical problems such as low efficiency and insufficient battery life of traditional inspection equipment in complex parking scenarios.

[0013] As a further improvement to the above solution, the dynamic path planning module executes a hierarchical optimization algorithm, including:

[0014] A path decomposition unit, wherein the path decomposition unit is used to divide the inspection area into a set of flight path segments {P i} and the ground path segment set {Q j};

[0015] A cost calculation unit is used to calculate a composite cost factor for each path segment:

[0016]

[0017] Where D is the length of the path segment in meters; V is the theoretical speed of the vehicle, and the flight segment takes V air , the ground segment takes V ground , unit is meter / second; η is the environmental correction factor, which is determined by tree cover in the flight segment and by traffic density in the ground segment; E pred Estimated energy consumption for the path segment, in watt-hours; C bat The real-time remaining power of the unmanned vehicle, in watt-hours;

[0018] A global optimization unit is used to generate a path sequence that minimizes the total cost:

[0019]

[0020] As a further improvement to the above scheme, the value of the environmental correction coefficient η is as follows:

[0021] (1) Flight path segment:

[0022]

[0023] (2) Ground path segment:

[0024]

[0025] As a further improvement of the above solution, the UAV's laser radar detects the terrain height change in real time and obtains the terrain height change rate K h =ΔH / ΔL; where ΔH is the height change between two consecutive point cloud frames; ΔL is the horizontal projection distance change of the corresponding point cloud;

[0026] When three consecutive frames of point cloud data meet the |K h |≥0.4, and the height difference ΔH≥2.5m, the control center issues a vertical inspection command, and the UAV and the unmanned vehicle switch to the third mode.

[0027] As a further improvement to the above solution, the locking mechanism includes a first base plate, on which are provided two mirror-symmetrical hooks; a rotating shaft is provided at the middle end of the hook, and the hook swings around the rotating shaft; a curved hook portion is provided at the front end of the hook for engaging with the hook ear; a flat plate portion is provided at the rear end of the hook, and an elastic member is provided on the outer side of the flat plate portion; the elastic member applies force on the flat plate portion to drive the hook to swing to a locked position; the hook is engaged with the hook ear in the locked position;

[0028] The driving unit includes a motor and a driving cam pivotally connected to the output shaft of the motor; the driving cam is arranged between the two hooks, and the outer peripheral contour of the driving cam abuts against the flat plate portion; the driving cam rotates, driving the two hooks to swing to the unlocked position and separate from the hook ears.

[0029] As a further improvement of the above solution, the docking platform includes a second substrate, the hook ears are vertically arranged on the second substrate, and the number of the hook ears is adapted to the hook.

[0030] As a further improvement of the above solution, the first substrate is further provided with a plurality of first positioning blocks, each of which is provided with a V-shaped groove; the second substrate is provided with second positioning blocks corresponding to the first positioning blocks, each of which is provided with a docking protrusion;

[0031] When the UAV and the unmanned vehicle switch to the third mode, the V-shaped groove of the first positioning block is engaged with the docking protrusion of the second positioning block.

[0032] As a further improvement to the above solution, the first charging module includes two slots, each of which is provided with a charging terminal, and the charging terminal is connected to the power storage pack of the drone; the second charging module includes two electrodes, and the electrodes are adapted to the slots;

[0033] When the drone and the unmanned vehicle switch to the third mode, the two electrodes are inserted into the two corresponding slots respectively, and the charging terminals are used to charge the drone's power storage pack.

[0034] A multimodal parking inspection method based on drone path planning is applied to the multimodal parking inspection system based on drone path planning as described above, comprising the following steps:

[0035] S1. Environmental Data Collection and Path Planning: The control center obtains real-time information about tree cover, vehicle density, and terrain height change rates detected by LiDAR within the inspection area. The dynamic path planning module implements a hierarchical optimization algorithm.

[0036] S2. Multimodal collaborative inspection: The control center sends instructions to the drone and unmanned vehicle based on the hierarchical optimization algorithm to perform the following mode switching:

[0037] (a) When the path segment is open airspace and the tree coverage rate is less than 30%, the first mode is activated: the locking mechanism is unlocked and the UAV flies independently for inspection;

[0038] (b) When the route segment is in an area with high traffic density or the UAV battery level is below a threshold, the second mode is activated: the UAV lands on the docking platform, the hook is locked with the hook ear, and the UAV is carried by the UAV and docked with the second charging module through the first charging module for charging;

[0039] (c) When the terrain height change rate |K is detected h When |≥0.4 and the continuous height difference ΔH≥2.5m, the third mode is activated: the drone hangs the unmanned vehicle to perform vertical obstacle inspection;

[0040] S3. Task closed-loop execution:

[0041] Repeat steps S1-S2 until the entire area is inspected, and transmit the berth data collected by the visual sensor back to the control center.

[0042] As a further improvement to the above solution, after the third mode of step S2(c) is executed, the following steps may be further performed:

[0043] When the drone and the unmanned vehicle cross the obstacle terrain, the control center triggers the mode switch based on the remaining path segment attributes:

[0044] If the remaining path segment is a flight path segment and the drone battery level is greater than 40%, the locking mechanism is controlled to unlock and switch to the first mode to continue the inspection;

[0045] If the remaining path segment is a ground path segment or the drone's battery level is ≤40%, the drone is controlled to land on the docking platform and switched to the second mode, where it is carried and charged by the unmanned vehicle.

[0046] Beneficial effects:

[0047] (1) Breaking through the endurance and obstruction limitations of drone inspections: The second mode is achieved through the locking mechanism and the docking platform, that is, the drone lands on the car to charge. The drone lands on the unmanned car in low-power or high-traffic areas, and the car's battery pack replenishes energy in real time through the charging module; eliminating the frequent return problems caused by high-load equipment of traditional drones and improving endurance efficiency.

[0048] (2) Overcoming the rigidity of unmanned vehicle paths and lack of perspective: The control center dynamically plans paths based on traffic density, tree coverage, and terrain height change rate. In areas with high traffic volume, it automatically switches to the second mode of unmanned vehicle ground inspection to avoid airspace interference; in open areas, it activates the first mode of drone high-position perspective to identify parking space changes in real time; and solves the problem that traditional unmanned vehicles rely on preset maps and cannot dynamically respond to environmental changes, thus avoiding repeated inspections.

[0049] (3) Achieve full coverage of complex terrain: Through the third mode, that is, the drone is hard-connected to the car and then suspended to fly, directly crossing the steep slope. The lidar detects the terrain height change rate in real time and triggers the suspension command; overcome the inspection interruption defect caused by the insufficient obstacle crossing ability of the unmanned car, and effectively improve the coverage rate of complex scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a framework diagram of the multi-modal parking inspection system of the present invention;

[0051] Figure 2 A third modal perspective view of the combination of the drone and the unmanned vehicle of the present invention;

[0052] Figure 3 A side view of the third mode of the drone and unmanned vehicle combination of the present invention;

[0053] Figure 4 A perspective view of the drone of the present invention;

[0054] Figure 5 for Figure 4 A local enlarged view of point a;

[0055] Figure 6 A perspective view of the unmanned vehicle of the present invention;

[0056] Figure 7 for Figure 6 A local enlarged view of point b;

[0057] Figure 8 This is a working principle diagram of the locking mechanism and the docking platform of the present invention;

[0058] Figure 9 This is a front view of the combination of the locking mechanism and the docking platform of the present invention;

[0059] Reference numerals:

[0060] 1. Drones;

[0061] 11. Locking mechanism; 11a. First substrate;

[0062] 111. Hook; 1111. Rotating axis; 1112. Arc-shaped hook portion; 1113. Flat plate portion;

[0063] 112. Driving unit; 1121. Driving cam;

[0064] 113. Elastic member;

[0065] 12. First charging module; 121. Slot; 122. Charging terminal;

[0066] 13. First positioning block; 131. V-shaped groove;

[0067] 2. Unmanned vehicle;

[0068] 21. Docking platform; 211. Hook; 21a. Second substrate;

[0069] 22. Second charging module; 221. Electrode;

[0070] 23. Second positioning block; 231. Docking protrusion. DETAILED DESCRIPTION

[0071] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0072] Example:

[0073] like Figures 1-9 As shown, this embodiment provides a multimodal parking inspection system based on drone path planning, including a drone 1, an unmanned vehicle 2 and a control center. The bottom of the drone 1 is provided with a locking mechanism 11; the locking mechanism 11 includes a hook 111 and a drive unit 112 that drives the hook 111 to rotate around its rotation axis 1111; a first charging module 12 is provided next to the locking mechanism 11 for docking with the unmanned vehicle 2 for charging; the drone 1 is also equipped with an RTK (English full name: Real-Time Kinematic) positioning module, a visual sensor and a laser radar; wherein the RTK positioning module is used for spatial positioning; the visual sensor is used for inspection and evidence collection; and the laser radar is used to detect terrain height changes in real time;

[0074] In this embodiment, the unmanned vehicle 2 adopts a crawler-type walking mechanism. In other embodiments, a wheeled walking mechanism may also be adopted. The unmanned vehicle 2 is equipped with a battery pack, specifically, a ternary lithium battery with a capacity of 200Wh to ensure the cruising range. The battery pack is connected to a second charging module 22; the second charging module 22 is configured to be compatible with the first charging module 12, and to charge and supply energy for the first charging module 12; a docking platform 21 is provided on the top of the unmanned vehicle 2, and the docking platform 21 is provided with a hook ear 211 that matches the locking mechanism 11; the drone 1 and the unmanned vehicle 2 are connected in a detachable hard connection through the cooperation of the hook 111 and the hook ear 211, forming three working modes:

[0075] First mode: UAV 1 is separated from unmanned vehicle 2 and performs flight inspection independently;

[0076] Second mode: After landing, the drone 1 is fixed on the docking platform 21 of the unmanned vehicle 2, and is carried by the unmanned vehicle 2 for charging and performing ground walking inspections;

[0077] The third mode: UAV 1 hangs unmanned vehicle 2 to perform vertical inspection;

[0078] The control center is configured with a dynamic path planning module to generate an optimal inspection path based on environmental data; the environmental data includes the tree coverage rate, traffic density and terrain height change rate of the inspection area; the control center sends inspection instructions to the drone 1 and the unmanned vehicle 2 according to the generated optimal inspection path.

[0079] In this embodiment, the dynamic path planning module executes a hierarchical optimization algorithm, including:

[0080] A path decomposition unit, wherein the path decomposition unit is used to divide the inspection area into a set of flight path segments {P i} and the ground path segment set {Q j};

[0081] A cost calculation unit is used to calculate a composite cost factor for each path segment:

[0082]

[0083] Where D is the length of the path segment in meters; V is the theoretical speed of the vehicle, and the flight segment takes V air , the ground segment takes V ground , unit is meter / second; η is the environmental correction factor, which is determined by tree cover in the flight segment and by traffic density in the ground segment; E pred The estimated energy consumption for the path segment is in watt-hours; specifically, the flight segment E pred =0.2D+5ΔH, ground segment E pred=0.1D, unit: Wh; ΔH is the cumulative climb height.

[0084] C bat is the real-time remaining power of the unmanned vehicle 2, in watt-hours. This formula quantifies the inspection efficiency by weighing the path length, speed, environmental impact, and energy consumption. The coefficient 0.8 is an empirical weight factor used to adjust the contribution of energy consumption to the total cost.

[0085] A global optimization unit is used to generate a path sequence that minimizes the total cost:

[0086]

[0087] The optimization goal ensures that the overall inspection path is optimal in terms of time and energy consumption, where m and n are the number of flight segments and ground segments, respectively.

[0088] In this embodiment, the value of the environmental correction coefficient η is determined as follows:

[0089] (1) Flight path segment:

[0090]

[0091] (2) Ground path segment:

[0092]

[0093] It should be noted that the η value is set based on actual test data. A high value indicates high environmental complexity and the cost weight needs to be increased to give priority to other modes or paths.

[0094] In this embodiment, the laser radar of the UAV 1 detects the change of terrain height in real time and obtains the terrain height change rate K h =ΔH / ΔL; where ΔH is the height change between two consecutive frames of point cloud; ΔL is the horizontal projection distance change of the corresponding point cloud; when three consecutive frames of point cloud data meet |K h |≥0.4, and the height difference ΔH≥2.5m, the control center issues a vertical inspection command, and UAV 1 and unmanned vehicle 2 switch to the third mode. This threshold is set based on typical urban obstacles, such as steep slopes or steps, to ensure timely response to sudden changes in terrain. The point cloud data is obtained from LiDAR scans, with a frame interval of 0.1 seconds.

[0095] In this embodiment, the locking mechanism 11 includes a first base plate 11a, on which two mirror-symmetrical hooks 111 are provided; a rotating shaft 1111 is provided at the middle end of the hook 111, and the hook 111 swings around the rotating shaft 1111; an arc-shaped hook portion 1112 is provided at the front end of the hook 111 for cooperating with the hook ear 211; a flat plate portion 1113 is provided at the rear end of the hook 111, and an elastic member 113 is provided on the outer side of the flat plate portion 1113; the elastic member 113 applies force to the hook 111. The flat plate portion 1113 drives the hooks 111 to swing to the locked position; in the locked position, the hooks 111 engage with the hook ears 211. The drive unit 112 includes a motor and a drive cam 1121 pivotally connected to the motor's output shaft. The drive cam 1121 is positioned between the two hooks 111, with its outer contour abutting the flat plate portion 1113. Rotation of the drive cam 1121 drives the two hooks 111 to swing to the unlocked position, separating them from the hook ears 211. The elastic member 113 is a torsion spring that ensures a stable locking state. The drive cam 1121 rotates within an angle of 0-90 degrees, enabling rapid locking and unlocking.

[0096] In this embodiment, the docking platform 21 includes a second base plate 21a, and the hook ears 211 are perpendicularly arranged on the second base plate 21a. The number of the hook ears 211 matches the number of the hooks 111. The hook ears 211 are inverted J-shaped metal parts that match the curved hook portion 1112 of the hook 111 to ensure the strength of the hard connection.

[0097] In this embodiment, the first substrate 11a is further provided with three first positioning blocks 13, each with a V-shaped groove 131. The second substrate 21a is provided with second positioning blocks 23 corresponding to the first positioning blocks 13, each with a docking protrusion 231. When the drone 1 and the unmanned vehicle 2 switch to the third mode, the V-shaped grooves 131 of the first positioning blocks 13 mate with the docking protrusions 231 of the second positioning blocks 23. This design provides auxiliary positioning and reduces shaking during hanging. The tolerance between the V-shaped grooves 131 and the docking protrusions 231 is ±0.1mm.

[0098] In this embodiment, the first charging module 12 includes two slots 121, and the slots 121 are correspondingly provided with charging terminals 122, and the charging terminals 122 are connected to the power storage group of the drone 1; the second charging module 22 includes two electrodes 221, and the electrodes 221 are adapted to each other with the slots 121; when the drone 1 and the unmanned vehicle 2 switch to the third mode, the two electrodes 221 are respectively inserted into the two slots 121, and the power storage group of the drone 1 is charged and supplied with energy through the charging terminals 122; the battery pack is connected to the second charging module 22; the second charging module 22 is configured to adapt to each other with the first charging module 12, and to charge and supply energy to the first charging module 12.

[0099] This embodiment further provides a multimodal parking inspection method based on drone path planning, which is applied to the multimodal parking inspection system based on drone path planning as described above, and includes the following steps:

[0100] S1. Environmental Data Collection and Path Planning: The control center obtains real-time information about tree cover, vehicle density, and terrain height change rates detected by LiDAR within the inspection area. The dynamic path planning module implements a hierarchical optimization algorithm.

[0101] S2. Multimodal collaborative inspection: The control center sends instructions to UAV 1 and UAV 2 based on the hierarchical optimization algorithm to perform the following mode switching:

[0102] (a) When the path segment is open airspace and the tree coverage rate is less than 30%, the first mode is activated: the locking mechanism 11 is controlled to unlock, and the UAV 1 performs an independent flight inspection;

[0103] (b) When the path segment is an area with high traffic density or the battery level of the drone 1 is lower than a threshold, the threshold in this embodiment is preset to 20%, and the second mode is activated: the drone 1 lands on the docking platform 21, the hook 111 is locked with the hook ear 211, and the drone 1 is carried and moved by the unmanned vehicle 2 and docked and charged with the second charging module 22 through the first charging module 12; by switching to the second mode, the endurance and obstruction limitations of the drone 1 inspection are overcome. This embodiment achieves the second mode through the locking mechanism 11 and the docking platform 21, that is, the drone 1 lands on the vehicle for charging. The drone 1 lands on the unmanned vehicle 2 in a low-battery or high-traffic area, and the vehicle's battery pack recharges energy in real time through the charging module; this eliminates the frequent return trips caused by high-load equipment in traditional drones 1, and improves endurance efficiency.

[0104] (c) When the terrain height change rate |K is detected h When |≥0.4 and the continuous height difference ΔH≥2.5m, the third mode is activated: UAV 1 hangs unmanned vehicle 2 to perform vertical obstacle inspection;

[0105] S3. Task closed-loop execution:

[0106] Repeat steps S1-S2 until the entire area is inspected, and transmit the berth data collected by the visual sensor back to the control center.

[0107] This embodiment overcomes the defects of the rigid path and lack of perspective of the unmanned vehicle 2. The control center dynamically plans the path based on traffic density, tree coverage, and terrain height change rate. In areas with high traffic, it automatically switches to the second mode of ground inspection of the unmanned vehicle 2 to avoid airspace interference; in open areas, the first mode of the high-position perspective of the drone 1 is activated to identify berth changes in real time. This solves the problem that the traditional unmanned vehicle 2 relies on preset maps and cannot dynamically respond to environmental changes, thus avoiding repeated inspections.

[0108] Specifically, after the third mode of step S2(c) is executed, the following steps are further performed:

[0109] When UAV 1 and UAV 2 cross the obstacle terrain, the control center triggers the mode switch based on the remaining path segment attributes:

[0110] If the remaining path segment is a flight path segment and the battery level of the drone 1 is greater than 40%, the locking mechanism 11 is controlled to unlock and switch to the first mode to continue the inspection;

[0111] If the remaining path segment is a ground segment or the battery level of UAV 1 is ≤40%, UAV 1 is controlled to land on docking platform 21 and switch to the second mode, where it is carried and charged by unmanned vehicle 2. The mode switching logic is based on real-time data and optimization algorithms to ensure the system adapts to environmental changes. The battery threshold is calibrated experimentally to balance inspection continuity and safety redundancy. In the third mode, UAV 1 is hard-connected to the vehicle and suspended, allowing it to directly cross steep slopes. The lidar detects the rate of change in terrain height in real time and triggers the suspension command. This overcomes the inspection interruption caused by the insufficient obstacle-crossing capability of unmanned vehicle 2, effectively improving coverage in complex scenarios.

[0112] Based on the disclosure and teachings of the above description, those skilled in the art may also make changes and modifications to the above embodiments. Therefore, the present invention is not limited to the specific embodiments disclosed and described above, and any modifications and variations of the invention should also fall within the scope of protection of the claims of the present invention. In addition, although certain specific terms are used in this description, these terms are for convenience of description only and do not constitute any limitation to the present invention.

Claims

1. A multimodal parking inspection system based on drone path planning, comprising a drone, an unmanned vehicle, and a control center, characterized by: The UAV is equipped with a locking mechanism at the bottom; the locking mechanism includes a hook and a drive unit that drives the hook to rotate around its rotation axis; a first charging module is provided next to the locking mechanism for docking with the unmanned vehicle for charging; the UAV is also equipped with an RTK positioning module, a visual sensor, and a laser radar; the RTK positioning module is used for spatial positioning; the visual sensor is used for inspection and evidence collection; and the laser radar is used to detect changes in terrain height in real time; The unmanned vehicle is equipped with a battery pack, which is connected to a second charging module; the second charging module is configured to be compatible with the first charging module and to charge and supply energy to the first charging module; a docking platform is provided on the top of the unmanned vehicle, which is provided with a hook ear that matches the locking mechanism; the drone and the unmanned vehicle are detachably connected through the cooperation of the hook and the hook ear, forming three working modes: First mode: The drone and the unmanned vehicle are separated and perform flight inspections independently; Second mode: After landing, the drone is fixed to the docking platform of the unmanned vehicle, and the unmanned vehicle carries it for charging and performs ground inspections. The third mode: the drone carries out vertical inspection by hanging an unmanned vehicle; The control center is equipped with a dynamic path planning module to generate an optimal inspection path based on environmental data; the environmental data includes the tree coverage rate, traffic density and terrain height change rate of the inspection area; the control center sends inspection instructions to drones and unmanned vehicles based on the generated optimal inspection path.

2. The multimodal parking inspection system based on drone path planning according to claim 1 is characterized in that: The dynamic path planning module executes a hierarchical optimization algorithm, including: A path decomposition unit, wherein the path decomposition unit is used to divide the inspection area into a set of flight path segments {P i } and the ground path segment set {Q j }; A cost calculation unit is used to calculate a composite cost factor for each path segment: Where D is the length of the path segment in meters; V is the theoretical speed of the vehicle, and the flight segment is V air , the ground segment takes V ground , unit is meter / second; η is the environmental correction factor, which is determined by tree cover in the flight segment and by traffic density in the ground segment; E pred Estimated energy consumption for the path segment, in watt-hours; C bat The real-time remaining power of the unmanned vehicle, in watt-hours; A global optimization unit is used to generate a path sequence that minimizes the total cost:

3. The multimodal parking inspection system based on drone path planning according to claim 2 is characterized in that: The value of the environmental correction coefficient η is as follows: (1) Flight path segment: (2) Ground path segment:

4. The multimodal parking inspection system based on drone path planning according to claim 3 is characterized in that: The UAV's laser radar detects the terrain height change in real time and obtains the terrain height change rate K h =ΔH / ΔL; where ΔH is the height change between two consecutive point cloud frames; ΔL is the horizontal projection distance change of the corresponding point cloud; When three consecutive frames of point cloud data meet the |K h |≥0.4, and the height difference ΔH≥2.5m, the control center issues a vertical inspection command, and the UAV and the unmanned vehicle switch to the third mode.

5. The multimodal parking inspection system based on drone path planning according to claim 1 is characterized in that: The locking mechanism includes a first base plate, on which are provided two mirror-symmetrical hooks; a rotating shaft is provided at the middle end of the hook, and the hook swings around the rotating shaft; a curved hook portion is provided at the front end of the hook for engaging with the hook ear; a flat portion is provided at the rear end of the hook, and an elastic member is provided on the outer side of the flat portion; the elastic member applies force on the flat portion to drive the hook to swing to a locked position; the hook is engaged with the hook ear in the locked position; The driving unit includes a motor and a driving cam pivotally connected to the output shaft of the motor; the driving cam is arranged between the two hooks, and the outer peripheral contour of the driving cam abuts against the flat plate portion; the driving cam rotates, driving the two hooks to swing to the unlocked position and separate from the hook ears.

6. The multimodal parking inspection system based on drone path planning according to claim 5 is characterized in that: The docking platform includes a second base plate, and the hook ears are vertically arranged on the second base plate. The number of the hook ears matches that of the hooks.

7. The multimodal parking inspection system based on drone path planning according to claim 6 is characterized in that: The first substrate is further provided with a plurality of first positioning blocks, each of which is provided with a V-shaped groove; the second substrate is provided with second positioning blocks corresponding to the first positioning blocks, each of which is provided with a docking protrusion; When the UAV and the unmanned vehicle switch to the third mode, the V-shaped groove of the first positioning block is engaged with the docking protrusion of the second positioning block.

8. The multimodal parking inspection system based on drone path planning according to claim 7 is characterized in that: The first charging module includes two slots, each of which is provided with a charging terminal, and the charging terminal is connected to the power storage pack of the drone; the second charging module includes two electrodes, and the electrodes are adapted to the slots; When the drone and the unmanned vehicle switch to the third mode, the two electrodes are inserted into the two corresponding slots respectively, and the charging terminals are used to charge the drone's power storage pack.

9. A multimodal parking inspection method based on drone path planning, applied to a multimodal parking inspection system based on drone path planning as claimed in claim 8, characterized in that: The following steps are involved: S1. Environmental Data Collection and Path Planning: The control center obtains real-time information about tree cover, vehicle density, and terrain height change rates detected by LiDAR within the inspection area. The dynamic path planning module implements a hierarchical optimization algorithm. S2. Multimodal collaborative inspection: The control center sends instructions to the drone and unmanned vehicle based on the hierarchical optimization algorithm to perform the following mode switching: (a) When the path segment is open airspace and the tree coverage rate is less than 30%, the first mode is activated: the locking mechanism is unlocked and the UAV flies independently for inspection; (b) When the route segment is in an area with high traffic density or the UAV battery level is below a threshold, the second mode is activated: the UAV lands on the docking platform, the hook is locked with the hook ear, and the UAV is carried by the UAV and docked with the second charging module through the first charging module for charging; (c) When the terrain height change rate |Kh| is detected to be ≥0.4 and the continuous height difference ΔH is ≥2.5m, the third mode is activated: the drone hangs the unmanned vehicle to perform vertical obstacle inspection; S3. Task closed-loop execution: Repeat steps S1-S2 until the entire area is inspected, and transmit the berth data collected by the visual sensor back to the control center.

10. The multimodal parking inspection method based on drone path planning according to claim 9, characterized in that: After the third mode of step S2(c) is executed, the following steps are further performed: When the drone and the unmanned vehicle cross the obstacle terrain, the control center triggers the mode switch based on the remaining path segment attributes: If the remaining path segment is a flight path segment and the drone battery level is greater than 40%, the locking mechanism is controlled to unlock and switch to the first mode to continue the inspection; If the remaining path segment is a ground path segment or the drone's battery level is ≤40%, the drone is controlled to land on the docking platform and switched to the second mode, where it is carried and charged by the unmanned vehicle.

Images