CN115542945A - Unmanned aerial vehicle group target three-dimensional continuous monitoring method capable of adaptively adjusting range - Google Patents
Unmanned aerial vehicle group target three-dimensional continuous monitoring method capable of adaptively adjusting range Download PDFInfo
- Publication number
- CN115542945A CN115542945A CN202211331938.2A CN202211331938A CN115542945A CN 115542945 A CN115542945 A CN 115542945A CN 202211331938 A CN202211331938 A CN 202211331938A CN 115542945 A CN115542945 A CN 115542945A
- Authority
- CN
- China
- Prior art keywords
- target
- unmanned aerial
- aerial vehicle
- visual
- height
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 50
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000001514 detection method Methods 0.000 claims abstract description 54
- 230000000007 visual effect Effects 0.000 claims abstract description 47
- 230000002787 reinforcement Effects 0.000 claims abstract description 14
- 238000005265 energy consumption Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims abstract 3
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 8
- 230000004438 eyesight Effects 0.000 claims description 5
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 238000004134 energy conservation Methods 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 241001465754 Metazoa Species 0.000 description 5
- 230000033001 locomotion Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000011179 visual inspection Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003042 antagnostic effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000016776 visual perception Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
- G05D1/104—Simultaneous control of position or course in three dimensions specially adapted for aircraft involving a plurality of aircrafts, e.g. formation flying
Abstract
The invention relates to a three-dimensional continuous monitoring method for an unmanned aerial vehicle group target with a self-adaptive range adjustment function, and belongs to the technical field of unmanned aerial vehicle target monitoring. Aiming at the problem of three-dimensional continuous monitoring of a single unmanned aerial vehicle group target based on visual detection, the method utilizes an authorized graph structure to model the group target, carries out flight decision by searching a central position in the horizontal direction, and utilizes offline data set to integrate an offline reinforcement learning algorithm in the height direction to realize self-adaptive flight height decision under the condition of considering the energy consumption of the unmanned aerial vehicle and the accuracy of a visual system. The method can adaptively adjust the detection range of the unmanned aerial vehicle, and effectively improves the monitoring performance of a single unmanned aerial vehicle.
Description
Technical Field
The invention relates to a three-dimensional continuous monitoring method for an unmanned aerial vehicle group target with a self-adaptive adjustment range, in particular to a method for self-adaptively adjusting the detection range of an unmanned aerial vehicle and improving the target monitoring performance by utilizing a graph structure modeling and an offline reinforcement learning algorithm, and belongs to the technical field of unmanned aerial vehicle target monitoring.
Background
With the development of the unmanned aerial vehicle technology, the unmanned aerial vehicle has wide application prospects in a plurality of fields such as logistics, security protection, military and the like. In recent years, along with the rapid development of artificial intelligence technology, a new research idea is provided for the target monitoring problem of the unmanned aerial vehicle. At present, the target monitoring technology of the unmanned aerial vehicle becomes one of the hottest research directions in the crossing field of the unmanned aerial vehicle and artificial intelligence.
The continuous monitoring of the target by the unmanned aerial vehicle is an important research direction in the intelligent control of the unmanned aerial vehicle. And (4) continuously tracking the task, wherein the monitored target is required to be always kept in the visual field range of the unmanned aerial vehicle, so that the information such as the position of the target is acquired at any time, and an accurate sensing result is provided for subsequent tasks. The unmanned aerial vehicle continuous monitoring task is very widely applied, in the field of wild animal protection, the unmanned aerial vehicle is used for continuously monitoring animal populations, researching the migration rule of animals and protecting endangered wild animals from being stolen; under the confrontation environment, the unmanned aerial vehicle is utilized to monitor various maneuvering targets of the other side, and therefore the unmanned aerial vehicle is beneficial to making proper decisions by the other side. Therefore, the continuous monitoring task of the unmanned aerial vehicle is widely existed in multiple aspects of society, and has good social benefits.
Currently, some research progress has been made on target monitoring methods for unmanned aerial vehicles, but some problems still exist. First, the prior art fails to fully explore the maximum monitoring capability of a single drone, i.e., how to adaptively adjust the detection range of a drone. The monitoring performance of a single unmanned aerial vehicle is the basis and premise of cooperation of multiple unmanned aerial vehicles, and the existing group target monitoring method mainly utilizes multiple unmanned aerial vehicles to cooperatively monitor one target or only allocates one target to each unmanned aerial vehicle. Secondly, the prior art ignores the group characteristics of the target, such as the fleet, the animal population, and the like, and the motions of the targets often have the same rules and trends. In addition, the prior art fails to adequately incorporate the characteristics of the sensor. For example, in a countermeasure environment, in order to prevent the position exposure of the unmanned aerial vehicle from being attacked, the airborne radar system of the unmanned aerial vehicle is usually silent, and at this time, the visual sensor needs to be fully utilized, that is, the unmanned aerial vehicle can realize the maximum monitoring decision task in a three-dimensional space by adaptively adjusting various parameters.
Disclosure of Invention
The invention aims to overcome the defects and shortcomings of the prior art, and aims to fully solve the problems that monitoring performance is not high due to the fact that the detection range cannot be adjusted in a self-adaptive mode when a single unmanned aerial vehicle monitors multiple targets, and the problems that the characteristics of group targets are ignored and the relation between the characteristics and the height of a visual detection system of the unmanned aerial vehicle in an antagonistic environment is ignored. In order to effectively improve the monitoring performance of a single unmanned aerial vehicle, the three-dimensional continuous monitoring method for the unmanned aerial vehicle cluster target with the self-adaptive range adjustment is creatively provided.
First, the concept will be explained.
1. Continuous monitoring
After the unmanned aerial vehicle detects the target, the target is tracked, all the targets are always kept in the visual field range of the unmanned aerial vehicle, so that the movement trend of the target can be mastered, and accurate target information can be provided for subsequent tasks.
2. Adaptive range
The monitoring range of the unmanned aerial vehicle is adjusted in a self-adaptive mode, so that the number of covered targets is changed. The visual perception system carried by the unmanned aerial vehicle is formed by splicing image information from a plurality of cameras, so that the monitoring range can be adjusted according to internal parameters such as the focal length of the cameras. When the adjustment of the internal parameters reaches the maximum threshold, the monitoring range needs to be changed by adjusting the distance between the camera and the target, thereby changing the monitoring capability.
3. Group object
When a plurality of targets have similar movement trends, such as when the targets are far away from the unmanned aerial vehicle, the plurality of targets can be regarded as a group, and the movement in the whole group can be represented by the target with the highest centrality degree in the group.
The invention is realized by adopting the following technical scheme.
A three-dimensional continuous monitoring method for a self-adaptive range-adjusting unmanned aerial vehicle group target comprises the following steps:
step 1: the airborne sensing system of the unmanned aerial vehicle is set to be silent by a radar, only the visual sensors are used for target detection, the target detection area of the unmanned aerial vehicle is obtained by splicing the multiple visual sensors, and the group target information is obtained.
The radar silence is set to prevent the target detection result from being interfered by the other party through means such as electronic interference or the like under the countermeasure environment, or to strike and destroy the unmanned aerial vehicle.
The multi-vision sensor splicing is realized by splicing images from a plurality of cameras, so that a larger visual field range is obtained, the detection requirement of a group target is met, and the detection capability of a single unmanned aerial vehicle is improved.
Step 2: modeling the acquired group target information into a weighted full-link graph, and calculating the centrality degree of each target in the weighted graph to obtain the central position of the group target. The position is used as the flight position of the unmanned aerial vehicle in the horizontal direction or the central position of a visual detection system of the unmanned aerial vehicle.
Specifically, step 2 may include the steps of:
step 2.1: a rights full connectivity graph is established using the group target information.
The target State acquired by the visual sensor is position information State = [ x ] using a camera (unmanned aerial vehicle) as a reference point i ,y i ]Wherein x is i 、y i The abscissa and ordinate of the ith target position are shown.
Establishing a weighted full-connection graph G by using a finite set V (G) taking all detected targets as vertexes and a connecting line E (G) between the vertexes target = (V, E). And (3) establishing a tie matrix A by taking the shortest distance between the target and the target as a weight:
wherein, a ij Representing the distance between object i and object j.
Step 2.2: the center position in the cluster target is found.
Adjacency through a rights graphAs seen in the matrix, the ith row element of the matrix represents the shortest distance between the ith object and each of the other objects, and therefore, the reciprocal C of the sum of each row element in the adjacent matrix is used i The degree of position center in the entire target group representing the target:
wherein n is the total number of targets, a ij Representing the shortest distance between the target value i and the target value j. C i The larger the value, the higher the degree of centrality of the target. And selecting the target with the highest centrality degree as the central position of the group target.
Step 2.3: and carrying out the track decision of the unmanned aerial vehicle in the horizontal direction.
The unmanned aerial vehicle is in the horizontal direction, regards the central point of crowd's target as the central point of unmanned aerial vehicle's flight position or visual detecting system, on this basis, regards the position of observing central point target as observation information:
Z＝[x′ i ,y′ i ]
wherein Z is an observation vector, x ', from a visual inspection system' i 、y′ i Respectively, the abscissa and ordinate of the center position target. And predicting the position of the target by adopting an extended Kalman filtering method to obtain the smooth flight path of the unmanned aerial vehicle at the horizontal position.
And step 3: and carrying out unmanned aerial vehicle self-adaptive flying height decision.
And finding the maximum value of the row of the central target in the adjacency matrix based on the adjacency matrix and the position of the central target of the weighted graph established by the group targets. By adopting the relation between the value and the radius and the flight height of the visual detection system and adopting an off-line reinforcement learning method, the flight height capable of maximizing the monitoring performance of the unmanned aerial vehicle is obtained in a self-adaptive manner.
The relation between the radius of the visual inspection system and the flying height refers to that when the change of parameters affecting the visual field range, such as the focal length of a camera in the visual inspection system, reaches a threshold value, the distance between the camera and an object can be adjusted in a self-adaptive mode to change the visual field range.
Specifically, step 3 may include the steps of:
step 3.1: an offline dataset of the flight mission is collected.
Target state and height data of manned aircrafts with the same vision detection system are collected to serve as an offline data set when the manned aircrafts execute similar tasks, and the offline data set is used for inputting training of an offline reinforcement learning method.
Step 3.2: and setting off-line reinforcement learning elements in the flight altitude decision.
Wherein, the state includes the farthest target distance with the central target in the group target, the flying height of the unmanned aerial vehicle and the detection radius of the visual detection system, and specifically as follows:
State＝[d max ,h uav ,r reg ]
wherein State represents the State of the unmanned aerial vehicle, d max Represents the distance, h, of the farthest target from the central target uav Indicating the flight altitude of the drone, r reg Representing the detection range radius of the drone.
Reward sets up the clear degree, the height rate of change that Reward includes unmanned aerial vehicle's field of vision utilization ratio, field of vision, specifically as follows:
Reward＝w r (r reg -d max /r reg ) -1 +w h (h t,uav -h min ) -1 +w t (h t,uav -h t-1,uav ) -1
wherein r is reg Is the detection radius of the visual detection system, d max Is the maximum distance from the central target (r) reg -d max /r reg ) -1 The visual field utilization rate of the unmanned aerial vehicle is shown, and the purpose is to monitor the target with the visual sense as clear as possible; w is a r Is a scale factor that controls the field of view utilization in the total reward; h is t,uav Is the real-time flight height h of the unmanned plane at time t min Is the minimum flying height of the drone, (h) uav -h min ) -1 Is shown without a personConsidering the energy consumption and the visual resolution, the height of the unmanned aerial vehicle is as low as possible on the premise of meeting the monitoring requirements of all targets; w is a h Is a scale factor controlling the minimum altitude of the drone in the total reward; h is t-1,uav Is the flight height of the unmanned plane at the moment t-1, (h) t,uav -h t-1,uav ) -1 The method is characterized in that the frequent change of the height of the unmanned aerial vehicle is avoided due to the consideration of energy conservation and smooth flight path; w is a t Is a scaling factor that controls the change in altitude of the drone in the total reward.
Step 3.3: and carrying out height decision of the unmanned aerial vehicle.
And training by using an offline reinforcement learning algorithm based on the offline data set and the set state and reward to obtain a training model.
After model deployment is carried out, the optimal flying height is obtained in a self-adaptive mode during each decision.
And 4, step 4: exception handling of the monitoring task continues.
When the unmanned aerial vehicle is in the maximum value of the detection range of the visual detection system, if the maximum value of the row of the central target in the adjacency matrix exceeds the detection radius at the moment, the target is considered to exceed the monitoring capability of the current unmanned aerial vehicle, and other unmanned aerial vehicles are requested to cooperate or a command center is reported to process.
The maximum value of the detection range refers to the detection range of the visual detection system when the unmanned aerial vehicle reaches the maximum flying height, namely, the target detection precision is influenced when the unmanned aerial vehicle flies beyond the height.
From step 1 to step 4, based on the group target detected by the visual detection system, the weighted full-connected graph modeling is performed, the horizontal direction decision is realized according to the central position estimation of the graph, the height direction decision is realized according to the offline reinforcement learning training, and the three-dimensional continuous monitoring on the group target is realized.
Advantageous effects
Compared with the prior art, the method of the invention has the following advantages:
the method fully considers the problem of three-dimensional continuous monitoring of the single unmanned aerial vehicle group target based on visual detection in the confrontation environment, utilizes the weighted graph structure to model the group target, carries out flight decision by searching the central position in the horizontal direction, and utilizes offline data set to combine an offline reinforcement learning algorithm in the height direction to realize self-adaptive flight height decision under the condition of considering the energy consumption and the accuracy of a visual system of the unmanned aerial vehicle. Therefore, the detection range of the unmanned aerial vehicle can be adjusted in a self-adaptive mode, and the monitoring performance of a single unmanned aerial vehicle is improved.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Detailed Description
The method of the present invention will be described in further detail below with reference to the accompanying drawings and examples.
As shown in fig. 1, a method for continuously monitoring a target of an adaptive range unmanned aerial vehicle cluster in three dimensions includes the following steps:
step A: detecting system setting;
specifically, in this embodiment, the method is the same as the step 1 of the invention;
and B: modeling group target information;
specifically, in this embodiment, the object is used as a vertex, the distance between the objects is used as a weight, and a weighted full-connection graph is established, which is specifically the same as the step 2.1 of the content of the invention;
and C: calculating a center position in the group target;
specifically, in this embodiment, the step is the same as the step 2.2 of the invention;
step D: generating unmanned aerial vehicle track decision in the horizontal direction;
specifically, in this embodiment, the method is the same as the step 2.3 of the invention;
step E: acquiring a flight offline data set:
specifically, in this embodiment, the same as the step 3.1 of the invention is performed;
step F: self-adaptive decision making of the flight height of the unmanned aerial vehicle;
specifically, in this embodiment, the steps are the same as the steps 3.2 and 3.3 of the invention;
step G: continuously monitoring exception handling in the task;
specifically, the present embodiment is the same as the inventive content step 4.
Examples
The method is characterized in that the method takes the steps that in a confrontation environment, one unmanned aerial vehicle carrying a plurality of cameras continuously monitors 8 moving targets, all the moving targets have the same moving trend and can be treated as group targets, and the detection range of a visual detection system is adjusted according to the distance between the targets and the unmanned aerial vehicle as an embodiment. The present embodiment will describe, by way of specific examples, a three-dimensional continuous monitoring method for a self-adaptive range of an unmanned aerial vehicle fleet target according to the present invention, and specific operation steps of the method in the following are described in detail.
A three-dimensional continuous monitoring method for a self-adaptive range of an unmanned aerial vehicle group target, as shown in fig. 1, includes the following steps:
step A: detecting system setting;
specifically, in the embodiment, a plurality of cameras carried by the unmanned aerial vehicle are spliced to obtain a circular detection view with the radius of R;
and B: modeling group target information;
specifically, in this embodiment, a weighted graph is established with 8 targets as vertices and distances between the targets as weights, and an 8 × 8 adjacency matrix is obtained;
step C: calculating a center position in the group target;
specifically, in this embodiment, based on the adjacency matrix, a vector of 1 × 8 is obtained by calculating the centrality degree of the target, and the target having the greatest centrality is selected as the central position;
step D: generating unmanned aerial vehicle track decision in the horizontal direction;
specifically, in the embodiment, the horizontal trajectory is obtained by tracking the central position target by using an extended kalman filter algorithm.
Step E: acquiring a flight offline data set:
in the embodiment, the same visual detection system is utilized, the pilot completes the target monitoring task and records the flight data;
step F: self-adaptive decision making of the flight height of the unmanned aerial vehicle;
specifically, in the embodiment, a target state and rewards are set, and the flight height of the unmanned aerial vehicle is obtained by adopting an offline reinforcement learning algorithm;
step G: processing abnormal targets in the continuous monitoring task;
specifically, in this embodiment, when the farthest target exceeds the maximum range of the target detection system, the command system is reported or the peer drone is informed to perform cooperative monitoring.
Claims (3)
1. A three-dimensional continuous monitoring method for a target of an unmanned aerial vehicle cluster with a self-adaptive range adjustment function is characterized by comprising the following steps:
step 1: setting an airborne sensing system of the unmanned aerial vehicle as radar silence, detecting a target only by using a visual sensor, splicing multiple visual sensors to obtain a target detection area of the unmanned aerial vehicle, and acquiring group target information;
step 2: modeling the acquired group target information into an authorized full-connection graph, and calculating the centrality degree of each target in the authorized graph to obtain the central position of the group target; taking the position as the flying position of the unmanned aerial vehicle in the horizontal direction or the central position of a visual detection system of the unmanned aerial vehicle;
and step 3: carrying out unmanned aerial vehicle self-adaptive flying height decision;
based on the adjacency matrix and the position of the central target of the weighted graph established by the group targets, finding the maximum value of the row of the central target in the adjacency matrix; by adopting the relation between the value and the radius and the flight height of the visual detection system and adopting an off-line reinforcement learning method, the flight height capable of maximizing the monitoring performance of the unmanned aerial vehicle is obtained in a self-adaptive manner;
the relation between the radius of the visual detection system and the flying height refers to that when the change of parameters affecting the visual field range, such as the focal length of a camera in the visual detection system, reaches a threshold value, the distance between the camera and an object can be adjusted in a self-adaptive mode to change the visual field range;
and 4, step 4: continuously monitoring the exception handling of the tasks;
when the unmanned aerial vehicle is in the maximum value of the detection range of the visual detection system, if the maximum value of the row of the central target in the adjacent matrix exceeds the detection radius at the moment, the target is considered to exceed the monitoring capability of the current unmanned aerial vehicle, and other unmanned aerial vehicles are requested to cooperate or a command center is reported to process;
the maximum value of the detection range refers to the detection range of the visual detection system when the unmanned aerial vehicle reaches the maximum flight height, namely, the target detection precision is influenced when the unmanned aerial vehicle flies beyond the maximum flight height;
based on the group targets detected by the visual detection system, the weighted full-connected graph modeling is carried out, the horizontal direction decision is realized according to the central position estimation of the graph, the height direction decision is realized according to the offline reinforcement learning training, and the three-dimensional continuous monitoring on the group targets is realized.
2. The method for continuously monitoring the unmanned aerial vehicle group target in a three-dimensional manner with the adaptive range adjustment function as claimed in claim 1, wherein the step 2 comprises the following steps:
step 2.1: establishing a authorized full connection graph by using the group target information;
the target State acquired by the visual sensor is position information State = [ x ] using the camera as a reference point i ,y i ]Wherein x is i 、y i The abscissa and the ordinate of the ith target position are expressed;
establishing a full connection graph G with weight values by using a finite set V (G) taking all detected targets as vertexes and a connecting line E (G) between the vertexes target = (V, E); and (3) establishing a tie matrix A by taking the shortest distance between the target and the target as a weight:
wherein, a ij Represents the distance between target i and target j;
step 2.2: solving the central position in the group target;
the ith row element of the matrix represents the ith target and the respective sumShortest distance of other objects, using reciprocal C of sum of elements of each row in adjacency matrix i The degree of position center in the entire target group representing the target:
wherein n is the total number of targets, a ij Represents the shortest distance between the target value i and the target value j; c i The larger the value, the higher the degree of centrality of the target; selecting a target with the highest centrality degree as the central position of the group target;
step 2.3: carrying out track decision of the unmanned aerial vehicle in the horizontal direction;
the unmanned aerial vehicle is in the horizontal direction, regard as the central point of unmanned aerial vehicle's flight position or visual detecting system with the central point of crowd's target, regard as observation information with the position of observing central point target:
Z＝[x′ i ,y′ i ]
wherein Z is an observation vector, x ', from a visual detection system' i 、y′ i Respectively the horizontal and vertical coordinates of the target at the central position; and predicting the position of the target by adopting an extended Kalman filtering method to obtain the smooth flight path of the unmanned aerial vehicle at the horizontal position.
3. The method for three-dimensional continuous monitoring of the drone swarm target with the adaptively adjustable range as claimed in claim 1, wherein step 3 comprises the steps of:
step 3.1: collecting an offline data set of a flight mission;
the method comprises the steps that target state and altitude data of manned aircrafts carrying the same visual detection system and performing similar tasks are collected to serve as an offline data set and used for inputting training of an offline reinforcement learning method;
step 3.2: setting off-line reinforcement learning elements in the flight height decision;
wherein, the state includes the farthest target distance with the central target in the group target, the flying height of the unmanned aerial vehicle and the detection radius of the visual detection system, and specifically as follows:
State＝[d max ,h uav ,r reg ]
wherein State represents the State of the UAV, d max Represents the distance, h, of the farthest target from the central target uav Indicating the flight altitude of the drone, r reg Representing the detection range radius of the drone;
reward sets up the clear degree, the height rate of change that Reward includes unmanned aerial vehicle's field of vision utilization ratio, field of vision, specifically as follows:
Reward＝w r (r reg -d max /r reg ) -1 +w h (h t,uav -h min ) -1 +w t (h t,uav -h t-1,uav ) -1
wherein r is reg Is the detection radius of the visual detection system, d max Is the maximum distance from the central target, (r) reg -d max /r reg ) -1 The visual field utilization rate of the unmanned aerial vehicle is shown, and the purpose is to monitor the target with the visual sense as clear as possible; w is a r Is a scale factor that controls the field of view utilization in the total reward; h is t,uav Is the real-time flight height h of the unmanned plane at time t min Is the minimum flying height of the drone, (h) uav -h min ) -1 The method is characterized in that the height of the unmanned aerial vehicle is as low as possible on the premise that all targets are monitored and controlled due to the consideration of energy consumption and visual resolution of the unmanned aerial vehicle; w is a h Is a scale factor that controls the minimum altitude of the drone in the total reward; h is t-1,uav Is the flight height of the unmanned plane at the moment t-1, (h) t,uav -h t-1,uav ) -1 The method is characterized in that the frequent change of the height of the unmanned aerial vehicle is avoided due to the consideration of energy conservation and smooth flight path; w is a t Is a scale factor for controlling the altitude change of the unmanned aerial vehicle in the total reward;
step 3.3: making a height decision of the unmanned aerial vehicle;
training by using an offline reinforcement learning algorithm based on the offline data set and the set state and reward to obtain a training model;
after model deployment is carried out, the optimal flying height is obtained in a self-adaptive mode during each decision.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211331938.2A CN115542945A (en) | 2022-10-28 | 2022-10-28 | Unmanned aerial vehicle group target three-dimensional continuous monitoring method capable of adaptively adjusting range |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211331938.2A CN115542945A (en) | 2022-10-28 | 2022-10-28 | Unmanned aerial vehicle group target three-dimensional continuous monitoring method capable of adaptively adjusting range |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115542945A true CN115542945A (en) | 2022-12-30 |
Family
ID=84718105
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211331938.2A Pending CN115542945A (en) | 2022-10-28 | 2022-10-28 | Unmanned aerial vehicle group target three-dimensional continuous monitoring method capable of adaptively adjusting range |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115542945A (en) |
-
2022
- 2022-10-28 CN CN202211331938.2A patent/CN115542945A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111932588B (en) | Tracking method of airborne unmanned aerial vehicle multi-target tracking system based on deep learning | |
| Khan et al. | Cooperative robots to observe moving targets | |
| CN108983823B (en) | Plant protection unmanned aerial vehicle cluster cooperative control method | |
| WO2020103110A1 (en) | Image boundary acquisition method and device based on point cloud map and aircraft | |
| CN112580537B (en) | Deep reinforcement learning method for multi-unmanned aerial vehicle system to continuously cover specific area | |
| CN115661204B (en) | Collaborative searching and tracking positioning method for moving target by unmanned aerial vehicle cluster | |
| CN110619276B (en) | Anomaly and violence detection system and method based on unmanned aerial vehicle mobile monitoring | |
| CN110799983A (en) | Map generation method, map generation equipment, aircraft and storage medium | |
| CN107783103A (en) | A kind of flying bird tracking intelligent method of lightweight with self-learning function | |
| CN110348560A (en) | A method of based on the trajectory predictions for improving glowworm swarm algorithm optimized particle filter | |
| Wang et al. | Obstacle avoidance of UAV based on neural networks and interfered fluid dynamical system | |
| Stasinchuk et al. | A multi-UAV system for detection and elimination of multiple targets | |
| CN115113651A (en) | Unmanned robot bureaucratic cooperative coverage optimization method based on ellipse fitting | |
| CN111273701A (en) | Visual control system and control method for holder | |
| Li et al. | Vg-swarm: A vision-based gene regulation network for uavs swarm behavior emergence | |
| CN112987765B (en) | Precise autonomous take-off and landing method of unmanned aerial vehicle/boat simulating attention distribution of prey birds | |
| Sun et al. | Cooperative strategy for pursuit-evasion problem in the presence of static and dynamic obstacles | |
| CN115629600B (en) | Multi-machine collaborative trapping method based on buffer Wino diagram in complex dynamic security environment | |
| CN115542945A (en) | Unmanned aerial vehicle group target three-dimensional continuous monitoring method capable of adaptively adjusting range | |
| Liu et al. | SMART: Vision-based method of cooperative surveillance and tracking by multiple UAVs in the urban environment | |
| CN110879607A (en) | Offshore wind power blade detection method based on multi-unmanned aerial vehicle formation cooperative detection | |
| CN114463301A (en) | Intelligent accompanying flight method for unmanned aerial vehicle video target in multi-shielding environment | |
| Kainth et al. | Chasing the Intruder: A Reinforcement Learning Approach for Tracking Unidentified Drones | |
| CN116225022A (en) | Pigeon-based thinking optimization unmanned aerial vehicle and unmanned ship heterogeneous collaborative capturing method | |
| CN117250859B (en) | Multi-aircraft collaborative search algorithm under communication constraint |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |