Environment-adaptive unmanned aerial vehicle computing task scheduling system and method thereof

Abstract

The invention discloses an environment-adaptive unmanned aerial vehicle calculation task scheduling system and method. The system comprises a sensing module, an environment difference detection module, a planning module, a collision detection module, a control module, an environment complexity calculation module, a system resource monitor and a frequency calculator. Through cooperative work of the modules, a closed-loop control system is constructed, and autonomous flight of the unmanned aerial vehicle in a complex environment is achieved. According to the system, under the condition of environment change, self-adaptive adjustment of map resolution can be achieved, the detection frequency of the corresponding module is optimized, and finally the flight task is completed efficiently and safely.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle scheduling, and in particular to an environment-adaptive unmanned aerial vehicle computing task scheduling system and method thereof. Background Art

[0002] With the rapid development of drone technology, a variety of drone dispatching systems have emerged on the market, which vary in function and scope of application. Generally speaking, the basic method flow of existing drone dispatching systems can be summarized as the following steps: First, the sensor module sends environmental data to the computing processing unit; second, after receiving the data, the computing processing unit processes it through a preset control algorithm to generate an assessment of the current environmental status and build a map; then, based on the established map and environmental information, the system performs path planning to determine the flight trajectory of the drone; finally, the planning results are sent to the drone's flight control system through the communication module to implement actual flight operations, continuously monitor the drone status, and update the path planning based on new sensor data.

[0003] In the UAV dispatch system, the perception frequency refers to the number of times the UAV perceives and processes information about the environment in a unit time. A high frequency can quickly respond to changes in the surrounding environment, which is especially important when avoiding obstacles. However, too high a frequency may increase the processing burden and energy consumption, and a reasonable balance needs to be made according to the application scenario; the planning frequency refers to the frequency at which the UAV updates the flight path or action plan, which directly affects the adaptability and flexibility of the system. A higher planning frequency can improve the efficiency of task execution, but it may also lead to excessive consumption of computing resources; the environmental complexity index is an important parameter to measure the density of obstacles in the UAV operating environment. It is defined by evaluating the number, size, and distribution of obstacles near the UAV flight path and their potential impact on the flight safety of the UAV. The higher the ECI value, the more complex the environment in which the UAV is located, and the higher the requirements for the obstacle avoidance, path planning, and autonomous decision-making capabilities of the UAV. Therefore, when designing a UAV dispatch system, these three parameters need to be considered comprehensively to ensure the flight safety and operation effect of the UAV.

[0004] Previously, different types of drones on the market may be equipped with different development boards such as TX2 and Raspberry Pi, and the chip computing power of different development boards varies. However, under the premise that there are obvious differences in the computing power of different development boards, most drone scheduling systems on the market use fixed perception and planning frequencies. When this statically set scheduling system is applied to different drones, it is obviously unable to adapt to the differences in computing power between development boards, and it is impossible to dynamically adjust the perception and planning frequencies during operation. When drones equipped with existing scheduling systems on the market perform autonomous flight missions, in many cases, it is precisely because they cannot dynamically adjust the perception and planning frequencies according to the actual computing power of the development board that a series of operational problems and safety hazards arise.

[0005] For example, when the computing power of the development board carried by the drone is relatively poor, if some scheduling systems ignore the computing power of the current development board in the perception and planning stages and forcibly use their fixed higher frequency for operation, the development board may not be able to withstand such high-intensity computing tasks due to insufficient computing power, causing the entire system to freeze or even crash, which greatly affects flight safety and the expected completion of the flight mission; on the contrary, when the computing power of the development board carried by the drone is sufficient, some scheduling systems use their fixed lower frequency for perception and planning operations, resulting in a waste of computing power on the development board. At the same time, the insufficiently detailed environmental information collected under this system also makes it impossible for the drone to build a complete and usable map, and thus it is impossible to plan an ideal path that can effectively avoid obstacles and reach the target point. In this way, the drone is very likely to collide with surrounding obstacles during flight, causing serious flight accidents.

[0006] In addition to the problem of using fixed perception and planning rates when running on different development boards, the current drone dispatch systems on the market also generally use fixed perception and planning frequencies when facing environments of different complexities. This static setting fails to take into account the impact of real-time environmental changes on drone operations, resulting in the inability to adjust flight strategies in a dynamic environment. Specifically, when drones are in highly complex or rapidly changing environments, such as urban high-rise areas or natural disaster sites, existing systems are often unable to dynamically adjust their perception and planning frequencies based on changes in environmental complexity. This lack of flexibility in the design makes drones slow to respond to rapidly changing environments, making it impossible to achieve efficient obstacle avoidance and path correction.

[0007] For example, when a drone is flying in an area with dense obstacles, if the system cannot increase the perception frequency according to the complexity of the environment, it may miss important obstacle information, resulting in inaccurate path planning and increased collision risk. At the same time, since the planning frequency cannot be increased, the system may not be able to update the flight path in a timely manner according to new environmental information, causing the drone to face more dangerous situations during flight. This not only significantly reduces the safety of flight, but also affects the efficiency of mission execution, causing potential economic losses and safety hazards. Therefore, the present invention aims to dynamically adjust the perception and planning frequency according to the real-time environmental complexity through an intelligent scheduling mechanism, so as to fundamentally solve the problem of insufficient adaptability of existing scheduling systems in complex environments.

[0008] In summary, the existing drone scheduling systems on the market generally adopt fixed perception and planning frequencies when equipped with development boards of different computing powers and facing environments of different complexities. This static strategy with obvious defects significantly reduces the universality and security of these scheduling systems. The present invention aims to fundamentally solve the above-mentioned key problems by dynamically adjusting the perception and planning frequencies according to the real-time environmental complexity and the computing power of the development board through an intelligent scheduling mechanism.

[0009] The Chinese invention patent application with application number 202410903819.2 discloses the "Path Planning Method for Intelligent UAV Inspection at Construction Sites". Its technical solution uses 3D reconstruction technology to build a dynamic 3D model of the construction site through images and data collected by the camera on the drone. Secondly, based on the combination algorithm of customized ant colony optimization algorithm and deep reinforcement learning, it generates strategic inspection paths and dynamically adjusts inspection tasks and paths based on real-time data feedback. Then, it integrates an event-driven dynamic scheduling system to respond to key events at the construction site, including equipment failures or the emergence of new structures, and replans the flight path of the drone. Finally, it uses machine learning methods to train a predictive obstacle avoidance model to predict potential risks and adjust the flight path in real time. The above solution aims to solve the path planning problem of drones at complex construction sites. Summary of the invention

[0010] In order to solve the technical problem that the conventional UAV scheduling system has low universality and security due to the inability to dynamically adjust the perception and planning frequency, the present invention provides an environment-adaptive UAV computing task scheduling system and method thereof. The technical solution adopted by the present invention is:

[0011] A first aspect of the present invention provides an environment-adaptive UAV computing task scheduling system, the system comprising a perception module, an environment difference detection module, a planning module, a collision detection module, a control module, an environment complexity calculation module, a system resource monitor, and a frequency calculator;

[0012] The perception module is used to detect obstacle information in the environment where the drone is located, obtain continuous perception data frames, and transmit the continuous perception data frames to the environment difference detection module, the planning module, and the environment complexity calculation module respectively;

[0013] The environmental difference detection module is used to detect the difference between the current perception data frame and the previous perception data frame according to the continuous perception data frames, obtain the environmental change difference, and transmit the environmental change difference to the planning module;

[0014] The planning module is used to adjust the resolution of the preset map according to the difference of the environmental change, and generate a flight path in the preset map according to the preset target position and the continuous perception data frames, and finally transmit the flight path to the control module;

[0015] The collision detection module is used to detect the distance between the drone and surrounding obstacles according to the continuous perception data frames, and send an emergency obstacle avoidance instruction to the control module when the distance is less than a preset safety threshold;

[0016] The control module is used to fly according to the flight path generated by the planning module and perform corresponding obstacle avoidance actions according to the emergency obstacle avoidance instructions;

[0017] The environment complexity calculation module is used to calculate the environment complexity according to the continuous perception data frames, and transmit the obtained environment complexity to the frequency calculator;

[0018] The system resource monitor is used to monitor the status information of the drone hardware and transmit the status information of the hardware to the frequency calculator;

[0019] The frequency calculator is used to calculate the perception frequency, the environment difference detection frequency and the collision detection frequency according to the environment complexity and the status information of the hardware, and transmit them to the perception module, the environment difference detection module and the collision detection module respectively.

[0020] As a preferred solution, the continuous sensing data frames include:

[0021] Obstacle location, obstacle size, obstacle speed, and obstacle distance.

[0022] As a preferred solution, the planning module includes a judgment module, a resolution adjustment module and a path generation module;

[0023] The judging module is used to judge whether the difference of the environmental change exceeds a preset threshold;

[0024] The resolution adjustment module is used to adjust the resolution of the preset map according to the judgment result of the judgment module;

[0025] The path generation module is used to generate a flight path in a preset map according to a preset target position and the continuous perception data frames.

[0026] As a preferred solution, the environment complexity calculation module includes a field of view partition module, a weight distribution module, an obstacle detection module, and a complexity calculation module;

[0027] The visual field partitioning module is used to partition the perception visual field of the perception module;

[0028] The weight allocation module is used to allocate the complexity weight of each partition view;

[0029] The obstacle detection module is used to perform obstacle detection on each partition according to the weight of the visual field complexity of each partition, and finally obtain the environmental complexity of each partition;

[0030] The complexity calculation module is used to perform weighted summation on the environmental complexity of each partition to obtain a final environmental complexity.

[0031] As a preferred solution, the hardware status information includes:

[0032] Core processor usage, memory usage.

[0033] A second aspect of the present invention provides an environment-adaptive UAV computing task scheduling method, the method is implemented based on the aforementioned environment-adaptive UAV computing task scheduling system, the method comprising:

[0034] Detecting obstacle information in the environment where the drone is located through the perception module to obtain continuous perception data frames;

[0035] The environmental difference detection module detects the difference between the current perception data frame and the previous perception data frame according to the continuous perception data frames to obtain the degree of environmental change difference;

[0036] The planning module adjusts the resolution of the preset map according to the difference in the environmental change, and generates a flight path in the preset map according to the preset target position and the continuous perception data frames;

[0037] The collision detection module detects the distance between the drone and surrounding obstacles according to the continuous perception data frames, and sends an emergency obstacle avoidance instruction to the control module when the distance is less than a preset safety threshold;

[0038] The control module flies according to the flight path generated by the planning module, and executes corresponding obstacle avoidance actions according to the emergency obstacle avoidance instructions.

[0039] As a preferred embodiment, the method further comprises:

[0040] Calculating the environmental complexity according to the continuous perception data frames by an environmental complexity calculation module;

[0041] Monitor the status information of drone hardware through the system resource monitor;

[0042] The frequency calculator calculates the perception frequency, environmental difference detection frequency and collision detection frequency according to the environmental complexity and the status information of the hardware, and enables the perception module, the environmental difference detection module and the collision detection module to perform detection according to the calculated perception frequency, environmental difference detection frequency and collision detection frequency respectively.

[0043] As a preferred solution, the method of adjusting the resolution of a preset map according to the difference of the environmental change through a planning module, and generating a flight path in the preset map according to a preset target position and the continuous perception data frames includes:

[0044] Determine by a judgment module whether the difference of the environmental change exceeds a preset threshold;

[0045] If the difference in the environmental change exceeds a preset threshold, the resolution of the preset map is adjusted by the resolution adjustment module;

[0046] If the difference in the environmental change does not exceed a preset threshold, the original map resolution is maintained;

[0047] The flight path is generated in a preset map according to the preset target position and the continuous perception data frames by the path generation module.

[0048] As a preferred solution, the method for calculating the environmental complexity according to the continuous perception data frames by the environmental complexity calculation module includes:

[0049] Partitioning the perception field of the perception module;

[0050] Assign each partition's vision complexity weight;

[0051] Obstacle detection is performed on each partition according to the weight of the complexity of the field of view of each partition, and finally the environmental complexity of each partition is obtained;

[0052] The weighted sum of the environmental complexity of each partition is performed to obtain the final environmental complexity.

[0053] The third aspect of the present invention provides a computer device, including a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of the aforementioned environment-adaptive drone computing task scheduling method are implemented.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The present invention uses an adaptive frequency adjustment strategy to monitor the CPU and memory usage in real time, and automatically adjusts the frequency according to the complexity of the current environment and the requirements of the flight mission, thereby reducing unnecessary energy consumption and extending the flight time of the UAV, enabling it to maintain efficient operation during long-term flight missions.

[0056] The present invention can accurately sense obstacles by dynamically adjusting the map resolution, so that the drone can obtain high-resolution environmental data in a timely manner and respond quickly in complex environments and sudden complex situations, effectively reducing the risk of collision and ensuring flight safety. By dynamically adjusting the map resolution, it is also possible to reasonably allocate computing resources, dynamically adjust the resolution according to the real-time status, reduce unnecessary data processing, improve the overall performance of the drone, and enhance the system's adaptability to different application scenarios.

[0057] The present invention realizes real-time monitoring of the hardware performance (such as CPU and memory occupancy rate) of the unmanned aerial vehicle through a system resource monitor. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of the structure of an environment-adaptive UAV computing task scheduling system provided in this embodiment;

[0059] Figure 2 A flow chart of an environment-adaptive UAV computing task scheduling method provided in this embodiment;

[0060] Figure 3 The drone perception map at low resolution provided for this embodiment;

[0061] Figure 4 This is a high-resolution drone perception map provided for this embodiment. DETAILED DESCRIPTION

[0062] The accompanying drawings are only used for illustrative purposes and are not to be construed as limiting the present invention;

[0063] It should be clear that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the embodiments of the present application.

[0064] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the embodiments of the present application. The singular forms of "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0065] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0066] In addition, in the description of this application, unless otherwise specified, "multiple" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0067] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0068] Example 1

[0069] Please refer to Figure 1 ,This embodiment provides an environment-adaptive UAV computing task scheduling system, the system includes a perception module, an environment difference detection module, a planning module, a collision detection module, a control module, an environment complexity calculation module, a system resource monitor and a frequency calculator;

[0070] The perception module is used to detect obstacle information in the environment where the drone is located, obtain continuous perception data frames, and transmit the continuous perception data frames to the environment difference detection module, the planning module, and the environment complexity calculation module respectively;

[0071] Specifically, the perception module includes an integrated radar and a camera.

[0072] The environmental difference detection module is used to detect the difference between the current perception data frame and the previous perception data frame according to the continuous perception data frames, obtain the environmental change difference, and transmit the environmental change difference to the planning module;

[0073] Specifically, the environment difference detection module builds a model of the environment state at the previous moment and compares it with the current perception data to determine whether there is a significant change in the environment. Once it is detected that the difference in environmental changes exceeds the threshold, the resolution adjustment module adjusts the map resolution to the highest and divides it into several small grids to enable more precise obstacle avoidance path planning.

[0074] The planning module is used to adjust the resolution of the preset map according to the difference of the environmental change, and generate a flight path in the preset map according to the preset target position and the continuous perception data frames, and finally transmit the flight path to the control module;

[0075] The collision detection module is used to detect the distance between the drone and surrounding obstacles according to the continuous perception data frames, and send an emergency obstacle avoidance instruction to the control module when the distance is less than a preset safety threshold;

[0076] The control module is used to fly according to the flight path generated by the planning module and perform corresponding obstacle avoidance actions according to the emergency obstacle avoidance instructions;

[0077] The environment complexity calculation module is used to calculate the environment complexity according to the continuous perception data frames, and transmit the obtained environment complexity to the frequency calculator;

[0078] The system resource monitor is used to monitor the status information of the drone hardware and transmit the status information of the hardware to the frequency calculator;

[0079] The frequency calculator is used to calculate the perception frequency, the environment difference detection frequency and the collision detection frequency according to the environment complexity and the state information of the hardware, and transmit them to the perception module, the environment difference detection module and the collision detection module respectively;

[0080] Specifically, the frequency calculator considers the environmental complexity and the system status information provided by the system resource monitor, dynamically calculates the perception frequency, collision detection frequency and environmental difference detection frequency through the reinforcement learning network, and then adjusts the working state of the corresponding module according to these calculation results. By intelligently adjusting these frequencies according to changes in the environment and system load, the system can optimize resource utilization and avoid resource waste or system overload while ensuring safety and performance.

[0081] In a specific embodiment, the continuous sensing data frames include:

[0082] Obstacle location, obstacle size, obstacle speed, and obstacle distance.

[0083] In a specific embodiment, the planning module includes a judgment module, a resolution adjustment module and a path generation module;

[0084] The judging module is used to judge whether the difference of the environmental change exceeds a preset threshold;

[0085] The resolution adjustment module is used to adjust the resolution of the preset map according to the judgment result of the judgment module;

[0086] The path generation module is used to generate a flight path in a preset map according to a preset target position and the continuous perception data frames.

[0087] In a specific embodiment, the environment complexity calculation module includes a field of view partition module, a weight distribution module, an obstacle detection module, and a complexity calculation module;

[0088] The visual field partitioning module is used to partition the perception visual field of the perception module;

[0089] The weight allocation module is used to allocate the complexity weight of each partition view;

[0090] The obstacle detection module is used to perform obstacle detection on each partition according to the weight of the visual field complexity of each partition, and finally obtain the environmental complexity of each partition;

[0091] The complexity calculation module is used to perform weighted summation on the environmental complexity of each partition to obtain a final environmental complexity.

[0092] In a specific embodiment, the hardware status information includes:

[0093] Core processor usage, memory usage.

[0094] Through the collaborative work of the above modules, the present invention constructs a closed-loop control system, which enables autonomous flight of UAVs in complex environments. The system can respond quickly to environmental changes and optimize resource utilization, ultimately completing flight missions efficiently and safely.

[0095] Example 2

[0096] Please refer to Figure 2 This embodiment provides an environment-adaptive UAV computing task scheduling method, which is implemented based on an environment-adaptive UAV computing task scheduling system described in Example 1, and includes:

[0097] S1: Detect obstacle information in the environment where the drone is located through the perception module to obtain continuous perception data frames;

[0098] S2: Detecting the difference between the current perception data frame and the previous perception data frame according to the continuous perception data frames through the environment difference detection module to obtain the degree of difference in environment change;

[0099] Specifically, by building a model of the environment state at the previous moment and comparing it with the current perception data, it is possible to determine whether there are significant changes in the environment. Once the difference in environmental changes exceeds the threshold, the system will adjust the map resolution to the highest and divide it into several small grids for more detailed obstacle avoidance path planning.

[0100] More specifically, during the flight of the drone, as the environment continues to change, the system can dynamically adjust the resolution of the map based on the distance between the flight trajectory and obstacles. Figure 3 and Figure 4 When the drone approaches an important obstacle or enters a complex area, the system will increase the resolution of the map to obtain more detailed environmental information. This process is controlled by the resolution adjustment module, which decides when to increase or decrease the resolution based on real-time monitoring data. In this way, the drone can effectively reduce unnecessary computing burden and resource consumption while maintaining efficient perception of the surrounding environment. After the resolution is improved, the system can accurately depict the location and characteristics of obstacles, providing effective support for subsequent path planning.

[0101] S3: adjusting the resolution of the preset map according to the difference of the environmental change through the planning module, and generating a flight path in the preset map according to the preset target position and the continuous perception data frames;

[0102] S4: Detecting the distance between the UAV and surrounding obstacles through the collision detection module according to the continuous perception data frames, and sending an emergency obstacle avoidance instruction to the control module when the distance is less than a preset safety threshold;

[0103] S5: Flying according to the flight path generated by the planning module through the control module, and executing corresponding obstacle avoidance actions according to the emergency obstacle avoidance instruction.

[0104] In a specific embodiment, the method further comprises:

[0105] Calculating the environmental complexity according to the continuous perception data frames by an environmental complexity calculation module;

[0106] Monitor the status information of drone hardware through the system resource monitor;

[0107] The frequency calculator calculates the perception frequency, environmental difference detection frequency and collision detection frequency according to the environmental complexity and the status information of the hardware, and enables the perception module, the environmental difference detection module and the collision detection module to perform detection according to the calculated perception frequency, environmental difference detection frequency and collision detection frequency respectively.

[0108] Specifically, after obtaining the environmental complexity index, the system will comprehensively evaluate this information through the policy controller. During this process, the system will use the ECI value as the basis and combine it with the current computing power of the development board to perform intelligent scheduling. The frequency calculator will dynamically adjust the perception and planning frequency according to the changes in ECI. A high ECI value means that the environment is more complex, and the system will automatically increase the perception frequency to collect environmental data more quickly; conversely, in a simple environment, the frequency can be reduced to save resources.

[0109] The present invention uses an adaptive frequency adjustment strategy to monitor the CPU and memory usage in real time and automatically adjust the frequency according to the complexity of the current environment and the requirements of the flight mission. This optimization not only reduces unnecessary energy consumption, but also extends the flight time of the drone, allowing it to maintain efficient operation during long-term flight missions. After effectively reducing resource consumption, the present invention saves operating costs for users and improves the overall work efficiency of the drone, giving it an important advantage in market competition.

[0110] In a specific embodiment, the method of adjusting the resolution of a preset map according to the difference in the environmental change by a planning module, and generating a flight path in the preset map according to a preset target position and the continuous perception data frames includes:

[0111] Determine by a judgment module whether the difference of the environmental change exceeds a preset threshold;

[0112] If the difference in the environmental change exceeds a preset threshold, the resolution of the preset map is adjusted by the resolution adjustment module;

[0113] If the difference in the environmental change does not exceed a preset threshold, the original map resolution is maintained;

[0114] The flight path is generated in a preset map according to the preset target position and the continuous perception data frames by the path generation module.

[0115] In a specific embodiment, the method of calculating the environmental complexity according to the continuous perception data frames by the environmental complexity calculation module includes:

[0116] Partitioning the perception field of the perception module;

[0117] Assign each partition's vision complexity weight;

[0118] Obstacle detection is performed on each partition according to the weight of the complexity of the field of view of each partition, and finally the environmental complexity of each partition is obtained;

[0119] The weighted sum of the environmental complexity of each partition is performed to obtain the final environmental complexity.

[0120] In summary, the innovation of the present invention in key technical points makes this scheduling system have significant advantages over existing drone scheduling systems in terms of dynamic adaptability, resource optimization, safety and user-friendliness. By introducing advanced technologies such as real-time monitoring, dynamic adjustment and intelligent decision-making, our solution not only improves the autonomous flight capability of drones in complex environments, but also lays a solid foundation for their wide application. These improvements enable the present invention to stand out among the many drone scheduling control systems today, meet the growing demand for drone applications, and provide new directions and impetus for the development of future drone technology.

[0121] Example 3

[0122] This embodiment provides a computer device, including a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor. When the computer program is executed by the processor, the steps of an environment-adaptive drone computing task scheduling method described in Example 2 are implemented.

[0123] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. An environment-adaptive UAV computing task scheduling system, characterized in that: The system includes a perception module, an environmental difference detection module, a planning module, a collision detection module, a control module, an environmental complexity calculation module, a system resource monitor, and a frequency calculator; The perception module is used to detect obstacle information in the environment where the drone is located, obtain continuous perception data frames, and transmit the continuous perception data frames to the environment difference detection module, the planning module, and the environment complexity calculation module respectively; The environmental difference detection module is used to detect the difference between the current perception data frame and the previous perception data frame according to the continuous perception data frames, obtain the environmental change difference, and transmit the environmental change difference to the planning module; The planning module is used to adjust the resolution of the preset map according to the difference of the environmental change, and generate a flight path in the preset map according to the preset target position and the continuous perception data frames, and finally transmit the flight path to the control module; The collision detection module is used to detect the distance between the drone and surrounding obstacles according to the continuous perception data frames, and send an emergency obstacle avoidance instruction to the control module when the distance is less than a preset safety threshold; The control module is used to fly according to the flight path generated by the planning module and perform corresponding obstacle avoidance actions according to the emergency obstacle avoidance instructions; The environment complexity calculation module is used to calculate the environment complexity according to the continuous perception data frames, and transmit the obtained environment complexity to the frequency calculator; The system resource monitor is used to monitor the status information of the drone hardware and transmit the status information of the hardware to the frequency calculator; The frequency calculator is used to calculate the perception frequency, the environment difference detection frequency and the collision detection frequency according to the environment complexity and the status information of the hardware, and transmit them to the perception module, the environment difference detection module and the collision detection module respectively.

2. The environment-adaptive UAV computing task scheduling system according to claim 1, characterized in that: The continuous sensing data frames include: Obstacle location, obstacle size, obstacle speed, and obstacle distance.

3. The environment-adaptive UAV computing task scheduling system according to claim 1, characterized in that: The planning module includes a judgment module, a resolution adjustment module and a path generation module; The judging module is used to judge whether the difference of the environmental change exceeds a preset threshold; The resolution adjustment module is used to adjust the resolution of the preset map according to the judgment result of the judgment module; The path generation module is used to generate a flight path in a preset map according to a preset target position and the continuous perception data frames.

4. The environment-adaptive UAV computing task scheduling system according to claim 1, characterized in that: The environment complexity calculation module includes a field of view partition module, a weight distribution module, an obstacle detection module, and a complexity calculation module; The visual field partitioning module is used to partition the perception visual field of the perception module; The weight allocation module is used to allocate the complexity weight of each partition view; The obstacle detection module is used to perform obstacle detection on each partition according to the weight of the visual field complexity of each partition, and finally obtain the environmental complexity of each partition; The complexity calculation module is used to perform weighted summation on the environmental complexity of each partition to obtain a final environmental complexity.

5. The environment-adaptive UAV computing task scheduling system according to claim 1, characterized in that: The hardware status information includes: Core processor usage, memory usage.

6. An environment-adaptive UAV computing task scheduling method, the method is implemented based on an environment-adaptive UAV computing task scheduling system according to any one of claims 1 to 5, characterized in that: The method comprises: Detecting obstacle information in the environment where the drone is located through the perception module to obtain continuous perception data frames; The environmental difference detection module detects the difference between the current perception data frame and the previous perception data frame according to the continuous perception data frames to obtain the degree of environmental change difference; The planning module adjusts the resolution of the preset map according to the difference in the environmental changes, and generates a flight path in the preset map according to the preset target position and the continuous perception data frames; The collision detection module detects the distance between the drone and surrounding obstacles according to the continuous perception data frames, and sends an emergency obstacle avoidance instruction to the control module when the distance is less than a preset safety threshold; The control module flies according to the flight path generated by the planning module, and executes corresponding obstacle avoidance actions according to the emergency obstacle avoidance instructions.

7. The method for scheduling computing tasks of unmanned aerial vehicles with environment adaptation according to claim 6, characterized in that: The method further comprises: Calculating the environmental complexity according to the continuous perception data frames by an environmental complexity calculation module; Monitor the status information of drone hardware through the system resource monitor; The frequency calculator calculates the perception frequency, environmental difference detection frequency and collision detection frequency according to the environmental complexity and the status information of the hardware, and enables the perception module, the environmental difference detection module and the collision detection module to perform detection according to the calculated perception frequency, environmental difference detection frequency and collision detection frequency respectively.

8. The method for scheduling computing tasks of unmanned aerial vehicles in an environment-adaptive manner according to claim 6, characterized in that: The method of adjusting the resolution of a preset map according to the difference of the environmental change by a planning module, and generating a flight path in the preset map according to a preset target position and the continuous perception data frames includes: Determine by a judgment module whether the difference of the environmental change exceeds a preset threshold; If the difference in the environmental change exceeds a preset threshold, the resolution of the preset map is adjusted by the resolution adjustment module; If the difference in the environmental change does not exceed a preset threshold, the original map resolution is maintained; The flight path is generated in a preset map according to the preset target position and the continuous perception data frames by the path generation module.

9. The method for scheduling computing tasks of unmanned aerial vehicles in an environment-adaptive manner according to claim 7, characterized in that: The method for calculating the environmental complexity according to the continuous perception data frames by the environmental complexity calculation module includes: Partitioning the perception field of the perception module; Assign each partition's vision complexity weight; Obstacle detection is performed on each partition according to the weight of the complexity of the field of view of each partition, and finally the environmental complexity of each partition is obtained; The weighted sum of the environmental complexity of each partition is performed to obtain the final environmental complexity.

10. A computer device, characterized in that: The method comprises a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor, wherein when the computer program is executed by the processor, the steps of an environment-adaptive unmanned aerial vehicle computing task scheduling method as described in any one of claims 6 to 9 are implemented.

Images