Low-power-consumption navigation system of unmanned aerial vehicle

By integrating low-power GPS modules and intelligent energy management modules on drones, and optimizing cruise routes with reinforcement learning models and environmental perception data, the problem of insufficient energy consumption optimization in the existing technology is solved, and longer battery life and More efficient power grid patrol capabilities.

CN119063722APending Publication Date: 2024-12-03CHENGDU POWER SUPPLY COMPANY OF STATE GRID SICHUAN ELECTRIC POWER

Patent Information

Application Number
CN202411286099.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-13
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing low-power solution for drones mainly saves energy from the perspective of avoiding over-discharge of batteries, and the operating logic is simple, resulting in limited energy saving effects and the inability to comprehensively optimize the energy consumption of drones, especially in insufficient battery life during power grid inspection.

Method used

A low-power navigation system for drones was designed, combining low-power GPS module and intelligent energy management module to realize energy consumption by adjusting the number of synchronized satellites during positioning and optimizing cruise routes. multi-dimensional optimization reduces the overall energy consumption of drones.

Benefits of technology

Effectively reduce the energy consumption of drones, improve endurance, and ensure cruise accuracy and flexibility. It is suitable for large-scale tasks such as power grid inspection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119063722A_ABST
    Figure CN119063722A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of unmanned aerial vehicles, and discloses an unmanned aerial vehicle low-power-consumption navigation system which comprises a low-power-consumption GPS module and an intelligent energy management module which are integrated on an unmanned aerial vehicle. The unmanned aerial vehicle is also provided with an inertial measurement unit and an environment sensing system; the low-power-consumption GPS module is used for positioning the position of the unmanned aerial vehicle; a power consumption adjusting unit is arranged in the low-power-consumption GPS module; the power consumption adjusting unit adjusts the number of synchronous satellites during positioning according to the cruising stage of the unmanned aerial vehicle; the intelligent energy management module is used for acquiring an initial cruise route and an initial cruise action instruction of the unmanned aerial vehicle, and correcting the initial cruise route and the initial cruise action instruction according to basic information acquired by the environment sensing system and the inertial measurement unit; a reinforcement learning model is preset in the intelligent energy management module. The cruising behavior of the unmanned aerial vehicle can be intelligently optimized, the energy consumption of the unmanned aerial vehicle is effectively reduced, and the cruising ability of the unmanned aerial vehicle is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to a low-power consumption navigation system for unmanned aerial vehicles. Background Art

[0002] With their unique flying capabilities, flexible carrying platforms and powerful data processing capabilities, drones can quickly carry out operations in different environments and provide key information support for power grid inspections. They have become an indispensable force in the intelligent power grid management system.

[0003] Since the inspection area of ​​the power grid is often very large, this places high demands on the endurance of drones. Analyzing the existing drone energy consumption optimization solutions, most of the existing solutions simply reduce energy consumption by switching the battery low power mode or controlling the current conduction state; for example, the "UAV Low Power Circuit and UAV" with the publication number CN209545206U published in 2019 sets a unidirectional conduction component to avoid leakage current between the power protection module and the control module, and reduces the discharge power consumption of the drone when it is not in use to achieve energy saving; for example, the "UAV Battery Low Power Control Device and System" with the publication number CN117411162B published in 2024 sets the main control circuit, low power wake-up circuit and battery input control circuit, and realizes active switching of working mode or low power mode by conducting different circuits to reduce overall energy consumption. In these solutions, an active signal (whether the drone is running or not) is required as a trigger, and they are all designed for energy saving from the perspective of avoiding over-discharge of the battery, and the energy saving effect is very limited.

[0004] And there is also a publication number CN118139171A published in 2024, "A Low-Power Positioning Method and Positioning Module for UAVs", which introduces energy consumption control of the positioning unit on the UAV. When the UAV is detected to be running, the positioning data is reported at a high frequency. When the UAV is detected to be not in use, it enters PSM+ mode, the main control unit stops running, and ultra-low power standby is achieved, thereby reducing overall energy consumption. Although this solution pays attention to other components in the UAV and proposes a new direction for energy consumption control, its operating logic is still a simple judgment of whether the UAV is running or not. The energy-saving control is not comprehensive, and the energy-saving effect is still relatively limited. Summary of the invention

[0005] The present invention aims to provide a low-power navigation system for a UAV, which can intelligently optimize the cruising behavior of the UAV, effectively reduce the energy consumption of the UAV, and improve the endurance of the UAV.

[0006] The basic solution provided by the present invention is: a low-power navigation system for unmanned aerial vehicles, comprising a low-power GPS module and an intelligent energy management module for integration on the unmanned aerial vehicle; the unmanned aerial vehicle is also provided with an inertial measurement unit and an environmental perception system;

[0007] The low-power GPS module is used to locate the position of the UAV; the low-power GPS module is provided with a power consumption adjustment unit; the power consumption adjustment unit adjusts the number of synchronous satellites during positioning according to the cruising stage of the UAV;

[0008] The intelligent energy management module is used to collect the initial cruise route and initial cruise action instructions of the UAV, and to correct the initial cruise route and initial cruise action instructions according to the basic information collected by the environmental perception system and the inertial measurement unit; the intelligent energy management module is preset with a reinforcement learning model;

[0009] When correcting the initial cruise route and the initial cruise action instructions, the intelligent energy management module first constructs a three-dimensional environmental model based on the basic information, and uses the three-dimensional environmental model as the state space of the reinforcement learning model; and combines the initial cruise route and the initial cruise action instructions to predict multiple sets of operations that the UAV may take in the next period; each set of operations contains multiple cruise action instructions; each cruise action instruction is attached with an energy consumption label; the operation set is used as the action space of the reinforcement learning model; then the reward function of the reinforcement learning model is set to minimize energy consumption, and a set of operations is output as the target cruise action instructions; and the initial cruise route is corrected according to the target cruise action instructions to obtain the target cruise route.

[0010] The working principle and advantages of the present invention are:

[0011] The present invention provides a low-power navigation system for unmanned aerial vehicles, which can optimize energy consumption in multiple dimensions based on the cruising behavior of the unmanned aerial vehicle, in combination with the cruising stage of the unmanned aerial vehicle, real-time perception data, etc., and can maintain a good cruising effect. The key point is that, first, this solution designs a low-power GPS module, which is not as simple as the conventional solution to switch the energy consumption mode according to whether the unmanned aerial vehicle is running or not (this is actually not applicable in the unmanned aerial vehicle cruising process that needs to be online all the time), but can be combined with the cruising stage of the unmanned aerial vehicle, and the number of synchronous satellites during positioning can be flexibly adjusted. By carrying out planned resource allocation between positioning accuracy and positioning chip energy consumption, the overall energy consumption during the unmanned aerial vehicle cruising process can be reduced, which can ensure that the GPS positioning function of the unmanned aerial vehicle is effective and has a longer endurance.

[0012] Second, this solution designs an intelligent energy management module, which provides a new control optimization direction for the energy consumption control of drones - reducing energy consumption by intelligently optimizing the cruising behavior of drones. During the cruising process of the drone, the intelligent energy management module will combine the basic information collected by the drone in real time, synchronously analyze the three-dimensional environment space where the drone is located, determine the multiple sets of operations that the drone is suitable for taking in the next period of time, and select the cruise action instructions with the goal of minimizing energy consumption (that is, adjusting the optimal action sequence with the lowest energy consumption), and optimize the cruise route, which can maneuverably reduce unnecessary action consumption during flight, realize flexible consumption reduction in the entire cruising process, and effectively improve the endurance of the drone and maintain a good cruising effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a schematic diagram of the system structure of an embodiment of a low-power navigation system for an unmanned aerial vehicle. DETAILED DESCRIPTION

[0014] The following is a further detailed description through specific implementation methods:

[0015] The embodiment is basically as shown in the attached Figure 1 Shown: A low-power navigation system for a UAV, including a low-power GPS module and an intelligent energy management module for integration on the UAV.

[0016] The drone is also provided with an inertial measurement unit, an environmental perception system and a wireless communication module. Specifically, the drone refers to a drone used for power grid inspection. The environmental perception system includes a laser radar, a temperature and humidity sensor, a wind resistance sensor and a camera module. The components in the environmental perception system all use existing components, and all establish communication connections with the intelligent energy management module. Among them, the perception data provided by the temperature and humidity sensor and the wind resistance sensor will actually affect the flight state of the drone during flight (such as excessive wind resistance, the drone needs to maintain speed, and the energy consumed is actually more; and in rainy weather, the drone needs to maintain speed, and its difficulty is greater than in normal weather), which helps to improve the authenticity of the three-dimensional environmental model constructed subsequently, and is more convenient for objectively setting cruise action instructions, as well as objectively setting the energy consumption labels of each cruise action instruction.

[0017] The low-power GPS module is used to locate the position of the UAV; a power consumption adjustment unit is provided in the low-power GPS module; the power consumption adjustment unit adjusts the number of synchronous satellites during positioning according to the cruising stage of the UAV.

[0018] Specifically, the cruise stage of the drone includes the pre-cruise stage, the cruise stage and the cruise completion stage; and when the cruise stage is the cruise stage, the cruise stage is also divided into a high positioning index stage and a low positioning index stage according to the inspection task indicators of the drone. When determining the pre-cruise stage, the cruise stage and the cruise completion stage, the starting point and the end point of the initial cruise route set in the drone are used as the basis for division. When the inspection task indicators of the drone include indicators such as power line detection and mapping that require recording the precise size information of the target object, the corresponding area is determined to be a high positioning index stage.

[0019] In the pre-cruise stage, the cruise completion stage and the low positioning index stage, the power consumption adjustment unit reduces the number of synchronous satellites during positioning. In this embodiment, in the high positioning index stage, the power consumption adjustment unit sets the number of synchronous satellites during positioning to 4. When the number of synchronous satellites needs to be reduced, the number of synchronous satellites can be reduced to 2 or 3 according to the current power of the UAV. When the current power of the UAV is less than 30%, it is reduced to 2; in other cases, it is reduced to 3.

[0020] With this arrangement, the solution can flexibly reduce the energy consumption of the UAV while ensuring that the inspection and positioning accuracy of the UAV meets the mission requirements.

[0021] The intelligent energy management module is used to collect the initial cruise route and initial cruise action instructions of the drone, and to correct the initial cruise route and initial cruise action instructions based on the basic information collected by the environmental perception system and the inertial measurement unit; a reinforcement learning model is preset in the intelligent energy management module.

[0022] The basic information includes point cloud data, temperature and humidity data, wind resistance data, image data, acceleration data, and angular velocity data. For the basic information, preprocessing operations are first performed; the preprocessing includes noise reduction, format conversion, and screening out abnormal data. In addition, for the image data in the basic information, a convolutional neural network (CNN) is used to extract features from the image data to extract features useful for understanding the environment; and feature correction is performed in combination with the point cloud data to facilitate the subsequent three-dimensional environment model.

[0023] The cruise action instructions include hovering, moving forward, moving backward, moving left, moving right, ascending, descending, and rotating.

[0024] When correcting the initial cruise route and initial cruise action instructions, the intelligent energy management module first constructs a three-dimensional environment model based on basic information, and uses the three-dimensional environment model as the state space of the reinforcement learning model. Specifically, when constructing a three-dimensional model based on basic information, a depth estimation model based on stereo vision is first used to generate a depth map of the environment, and pose estimation is performed in combination with acceleration data, angular velocity data, etc., to correct the geometric errors in the depth map, and a complete three-dimensional environment model is constructed by combining point cloud data through a point cloud fusion algorithm.

[0025] Preferably, the state space also includes the current position, speed and remaining power information of the drone, so as to provide more dimensional environmental information, which helps to improve the decision accuracy of the subsequent reinforcement learning model.

[0026] Combined with the initial cruise route and the initial cruise action instructions, multiple sets of operations that the drone may take in the next period are predicted; each set of operations contains multiple cruise action instructions; each cruise action instruction is attached with an energy consumption label; the operation set is used as the action space of the reinforcement learning model; the reward function of the reinforcement learning model is set to minimize energy consumption, and a set of operations is output as the target cruise action instructions; and the initial cruise route is corrected according to the target cruise action instructions to obtain the target cruise route. In this embodiment, when correcting the initial cruise route, the DQN algorithm can be used to intelligently complete the correction, and the deviation between the target cruise route and the initial cruise route is less than 10%, so as to ensure that the cruise task can be completed even if the route is adjusted.

[0027] The energy consumption tag of the cruise action instruction is obtained by the following steps:

[0028] The preset prediction model is trained using a historical drone data set, which contains multiple sets of drone historical cruise route data, as well as corresponding cruise action instructions, energy consumption change curve data, and environmental perception data; when training the preset prediction model, the historical drone data set is used as input, and the cruise action instructions and their corresponding energy consumption ratios are used as output; the energy consumption label is set according to the energy consumption ratio. The preset prediction model is a BP neural network model.

[0029] When the drone completes the cruise and returns, the remaining power information of the drone is also collected. If the remaining power is greater than the expected power, the reinforcement learning model is given a positive reward; otherwise, the reinforcement learning model is given a negative reward; the expected power refers to the estimated remaining power of the drone after it completes the cruise and returns under the initial cruise route and initial cruise action instructions. With this setting, the reinforcement learning model can gradually become more accurate and reliable as the number of actual applications increases.

[0030] The estimated power is also predicted using a BP neural network model. This BP neural network model is trained using a historical flight data set. The historical flight data set contains multiple sets of historical cruise route data of the drone, as well as the corresponding cruise action instructions and the remaining power data of the drone after completing the cruise and returning. During training, the historical flight data set is used as input and the estimated remaining power is used as output. During prediction, the initial cruise route and the initial cruise action instructions are used as input, and the estimated remaining power is used as output.

[0031] Optionally, the reward function of the reinforcement learning model also includes the shortest route and the least collision. With this arrangement, the system can support more diverse inspection requirements and has a wider range of applications.

[0032] The low-power consumption navigation system for a drone provided in this embodiment can intelligently optimize the drone's cruising behavior, effectively reduce the drone's energy consumption, and improve the drone's endurance.

[0033] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field are aware of all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement the scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for the ordinary technicians in the relevant field to implement this application. It should be pointed out that for the technicians in this field, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A low-power navigation system for unmanned aerial vehicles, characterized in that: It includes a low-power GPS module and an intelligent energy management module for integration on a drone; the drone is also provided with an inertial measurement unit and an environmental perception system; The low-power GPS module is used to locate the position of the UAV; the low-power GPS module is provided with a power consumption adjustment unit; the power consumption adjustment unit adjusts the number of synchronous satellites during positioning according to the cruising stage of the UAV; The intelligent energy management module is used to collect the initial cruise route and initial cruise action instructions of the UAV, and to correct the initial cruise route and initial cruise action instructions according to the basic information collected by the environmental perception system and the inertial measurement unit; the intelligent energy management module is preset with a reinforcement learning model; When correcting the initial cruise route and the initial cruise action instructions, the intelligent energy management module first constructs a three-dimensional environmental model based on the basic information, and uses the three-dimensional environmental model as the state space of the reinforcement learning model; and combines the initial cruise route and the initial cruise action instructions to predict multiple sets of operations that the UAV may take in the next period; each set of operations contains multiple cruise action instructions; each cruise action instruction is attached with an energy consumption label; the operation set is used as the action space of the reinforcement learning model; then the reward function of the reinforcement learning model is set to minimize energy consumption, and a set of operations is output as the target cruise action instructions; and the initial cruise route is corrected according to the target cruise action instructions to obtain the target cruise route.

2. The low-power consumption navigation system for unmanned aerial vehicles according to claim 1, characterized in that: The environmental perception system includes a laser radar, a temperature and humidity sensor, a wind resistance sensor and a camera module.

3. The low-power consumption navigation system for unmanned aerial vehicles according to claim 1, characterized in that: The cruise stage of the UAV includes a pre-cruise stage, a cruise stage and a cruise completion stage; and when the cruise stage is the cruise stage, the cruise stage is also divided into a high positioning index stage and a low positioning index stage according to the inspection task index of the UAV.

4. The low-power consumption navigation system for unmanned aerial vehicles according to claim 3, characterized in that: In the pre-cruise stage, the cruise completion stage and the low positioning index stage, the power consumption adjustment unit reduces the number of synchronous satellites during positioning.

5. The low-power consumption navigation system for unmanned aerial vehicles according to claim 1, characterized in that: The cruise action instructions include hovering, moving forward, moving backward, moving left, moving right, ascending, descending, and rotating.

6. The low-power consumption navigation system for unmanned aerial vehicles according to claim 5, characterized in that: The energy consumption tag of the cruise action instruction is obtained by the following steps: A preset prediction model is trained using a historical drone data set, which contains multiple sets of drone historical cruise route data, as well as corresponding cruise action instructions, energy consumption change curve data, and environmental perception data. When training the preset prediction model, the historical drone data set is used as input, and the cruise action instructions and their corresponding energy consumption ratios are used as output; the energy consumption label is set according to the energy consumption ratio.

7. The low-power consumption navigation system for unmanned aerial vehicles according to claim 6, characterized in that: The preset prediction model is a BP neural network model.

8. The low-power consumption navigation system for unmanned aerial vehicles according to claim 1, characterized in that: The state space also includes the current position, speed and remaining power information of the drone.

9. The low-power consumption navigation system for unmanned aerial vehicles according to claim 1, characterized in that: The reward function of the reinforcement learning model also includes the shortest route and the least collision.

10. The low-power consumption navigation system for unmanned aerial vehicles according to claim 1, characterized in that: When the UAV completes the cruise and returns, the remaining power information of the UAV is also collected. If the remaining power is greater than the expected power, the reinforcement learning model is given a positive reward; otherwise, the reinforcement learning model is given a negative reward; the expected power refers to the expected remaining power of the UAV after it completes the cruise and returns under the conditions of the initial cruise route and initial cruise action instructions.

Citation Information

Patent Citations

  • UAV battery low power consumption control device and system

    CN117411162B

  • Low-power-consumption positioning method and positioning module for unmanned aerial vehicle

    CN118139171A

  • Unmanned aerial vehicle low-power-consumption circuit and unmanned aerial vehicle

    CN209545206U

Cited By

  • Unmanned aerial vehicle endurance intelligent optimization method and system based on multi-modal data fusion

    CN120973041A