A method for unmanned aerial vehicle distributed cooperative predictive control for communication impaired scenarios
By using distributed model predictive control and robust multivariable observers, the problem of communication loss during multi-UAV cooperative flight was solved, achieving highly reliable and low-burden cooperative control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG UNIV OF TECH
- Filing Date
- 2024-01-31
- Publication Date
- 2026-06-02
Smart Images
Figure CN118192661B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the technical field of control or regulation systems for non-electrical variables, and specifically relates to a distributed cooperative predictive control method for UAVs in communication-impaired scenarios, within the field of automatic control technology for multi-UAV cooperative flight. Background Technology
[0002] With the continuous improvement of science and technology and the advancement of electronic technology over the past decade, unmanned aerial vehicles (UAVs) have distinguished themselves in many military and civilian fields, and are increasingly being used in geological exploration, remote search and rescue, remote sensing, and transportation. Facing increasingly complex mission scenarios, multi-UAV collaborative control has gradually become a very popular research direction.
[0003] Currently, numerous researchers have proposed advanced control algorithms for multi-UAV cooperative control, such as PID control, sliding mode control, and iterative control. However, these control algorithms struggle to account for the various complex constraints of UAV systems, which can negatively impact the overall cooperative flight. Furthermore, considering the overall multi-UAV system and its constraints, the high computational burden and communication load inherent in traditional centralized controllers can severely affect the reliability of multi-UAV cooperative flight.
[0004] On the other hand, multi-UAV collaborative control is carried out under a certain communication network. Most existing studies assume that the information interaction between UAVs takes place in an ideal environment. However, in actual engineering scenarios, the communication network may experience link failures due to limited network bandwidth, traffic congestion, hacker attacks, and other reasons. Such communication impairment factors will cause the multi-UAV collaborative flight control mission to deviate from the expected results. Summary of the Invention
[0005] In order to solve the problems existing in the current technology, this invention proposes a distributed cooperative predictive control method for UAVs in communication-impaired scenarios.
[0006] The technical concept of this invention is to provide a distributed cooperative predictive control method for UAVs in communication-impaired scenarios. This method designs a flight controller based on a UAV system model within a distributed model predictive control (DMPC) framework. In the independent controllers of each UAV, a robust multivariate observer (RMO) is designed to estimate the system state under false data injection (FDI) attacks and unknown disturbances. At the same time, a compensation strategy is designed for networked communication-impaired scenarios.
[0007] The technical solution adopted in this invention is a distributed cooperative predictive control method for unmanned aerial vehicles (UAVs) in communication-impaired scenarios. The method establishes a multi-UAV system model for cooperative flight missions, configuring a network model and corresponding transmission channel buffer models for each UAV in a communication-impaired scenario. An improved robust multivariate observer is used to estimate the state of the UAV after disturbance and the interference information in the output channels. The interference information includes, but is not limited to, measurement disturbances and estimation of spurious data injection attacks. The estimated state is used as the input to the distributed cooperative predictive controller to compensate for the corresponding interference, obtaining a control input generation and compensation strategy for communication-impaired scenarios. This multi-UAV distributed cooperative predictive controller is then used to achieve distributed cooperative predictive control of UAVs in communication-impaired scenarios.
[0008] Preferably, the multi-UAV system model for the cooperative flight mission includes M UAVs, and the i-th UAV satisfies equation (1) at time k.
[0009]
[0010] Where, x i y i u i w ix w iy and a i These include the system status, measurement output, control input, bounded system disturbances, measurement disturbances, and the spoofed data injection attacks faced by UAV i in a networked communication environment. and f i (·) represents the system function describing the UAV, f i (·): And satisfy f i (0,0)=0;C i D ix D iy and S i All are constant matrices with a preset dimension, and matrix [D] iy S i [This refers to a column full rank, i.e., the measured disturbance w] iy and fake data injection attack a i It is linearly separable.
[0011] Define constraints for the multi-UAV system model.
[0012] Preferably, the constraint satisfies,
[0013]
[0014] in, and These are closed sets and compact sets, respectively, including the origin.
[0015] Preferably, the network model in the communication impairment scenario is as follows:
[0016]
[0017]
[0018]
[0019] in, and drones Packet loss in the transmission of information from the drone's own sensor to the controller channel (SC packet loss), packet loss in the transmission of information from the drone i's own controller to the driver channel (CA packet loss), and packet loss in the interaction information between drone i and its neighboring drone j. N i Let i be the set of neighboring drones that have a directed communication topology connection with drone i; the maximum consecutive packet loss times are respectively and Any drone Its own SC packet loss, CA packet loss, and drone i and any drone Packet loss during information transmission between them is limited by the maximum continuous packet loss time. There is a reliable and time-free feedback link that can detect whether packet loss occurs at any given moment.
[0020] Preferably, the buffer model for each transmission channel includes deploying buffers in the UAV i's own sensor-to-controller channel and its own controller-to-driver channel. and Information transmission used to describe itself also includes deployment in the transmission channels between drones i and j. buffers This is used to describe the information interaction between it and its neighboring drones, where... Indicates a collection of neighboring drones cardinality and
[0021] Preferably, for time k buffer and The information storage lengths are 2N respectively. p +1 and N p ,buffer The storage length is N B ,satisfy And N B ≤N p .
[0022] Preferably, the improved robust multivariate observer is,
[0023]
[0024] Among them, L i For the gain of the observer to be designed, Let ξ be the estimated values of each vector, where ξ = η, z, x, y. C i0 =[C i D iy S i ], For H i1 The left inverse, H i1 =H i0 +N i0 C i0 ,
[0025] In this invention, the original system's equation (1) is described as the following augmented system.
[0026]
[0027] in, C i0 =[C i D iy S i ], in order to make z i The left-hand matrix of (k+1) is invertible, further let Add N to both sides of the brightness enhancement system i0 C i0 z i (k+1), we get,
[0028] H i1 z i (k+1)=G i0 z i (k)+F i0 (x i (k),u i (k))+D i0 w ix (k)+Z i0 y i (k)+N i0 y i (k+1),
[0029] Where H i1 =H i0 +N i0 Ci0 ,definition For H i1 The left inverse of the above expression reconstructs it into...
[0030]
[0031] in,
[0032] make As an augmented state, we obtain
[0033]
[0034] in
[0035] An improved robust multivariate observer is obtained.
[0036] Preferably, the variable r i (k) indicates whether UAV i needs optimization.
[0037]
[0038] Drone i only r i When (k) = 1, solve the optimization problem, r i If (k) = 0, then the optimization problem is not solved.
[0039] Preferably, the corresponding constraint optimization problem for the distributed cooperative predictive controller of multiple UAVs under communication impairment is as follows:
[0040]
[0041]
[0042]
[0043]
[0044]
[0045] in, Let Ω be the optimal control input sequence at time k. i (ε i ):={x|x T P i x≤ε i 2} represents the designed terminal domain.
[0046] Preferably, the cost function of UAV i is defined.
[0047]
[0048] in, and For nominal system The predicted state and predictive control input, Q i R i P i and R ij All are diagonal weighted matrices. and for The state information stored in it, κ(x) is the feedback control law.
[0049] In this invention, a cooperative controller is designed based on the concept of distributed MPC. The cost function of the UAV i∈M is initialized as follows:
[0050]
[0051] in, and For nominal system The predicted state and predicted control input are as follows. Synchronization terms are used to achieve synchronized flight between different drones. For the terminal cost, Q i R i P i and Q ij All are diagonal weighted matrices, where Q i and Q ij This represents a weighted matrix corresponding to different norm expressions; x j (s|k) and x j (s|kn) represents the drone The interactive status information that needs to be transmitted to drone i may not be available in real time due to random packet loss in the transmission channel between neighbors, and a certain compensation strategy is needed to replace it.
[0052] The cost function is then transformed into
[0053]
[0054] Based on the basic principles of MPC, As the optimal control input applied to UAV i, to compensate for the communication impairment caused by packet loss in each channel, when r i When (k) = 1, solve the optimization problem corresponding to the distributed cooperative predictive controller for multiple UAVs, and then... Applying to drone i; when r i When (k) = 0, then directly... Applying this to drone i, where k-τ is the nearest time without packet loss (referring to packet loss in neighbor interaction information) from time k; for any drone Repeat the above steps until the collaborative flight mission is completed.
[0055] This invention relates to a distributed cooperative predictive control method for unmanned aerial vehicles (UAVs) in communication-impaired scenarios. It establishes a multi-UAV system model for cooperative flight missions, configuring a network model and corresponding transmission channel buffer models for each UAV in the communication-impaired scenario. An improved robust multivariate observer is used to estimate the state of the UAVs after disturbance and the interference information in the output channels. The estimated state is used as the input to the distributed cooperative predictive controller to compensate for the corresponding interference, thus obtaining a control input generation and compensation strategy for communication-impaired scenarios. The generated multi-UAV distributed cooperative predictive controller is used to implement distributed cooperative predictive control of UAVs in communication-impaired scenarios.
[0056] The beneficial effects of this invention are as follows:
[0057] (1) The designed RMO effectively filters out measurement disturbances and FDI attack effects in the output channel and obtains an estimate of the actual state.
[0058] (2) By using a compensation strategy based on buffer updates and the designed cost function, the communication impairment caused by random packet loss in the transmission channel is effectively addressed.
[0059] (3) By using a distributed MPC framework and non-periodic optimization problem solving, the computational burden in cooperative flight missions is reduced;
[0060] (4) Distributed cooperative predictive control can effectively reduce communication load and computational burden, better overcome communication impairment factors, and has high reliability in unknown interference and complex network environments. Attached Figure Description
[0061] Figure 1 This is a flowchart of the method of the present invention;
[0062] Figure 2 This is a schematic diagram of the controller design of the present invention. Detailed Implementation
[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0064] like Figure 1As shown, this embodiment discloses a distributed cooperative predictive control method for UAVs in communication-impaired scenarios. Its main execution part is implemented on the UAV's onboard computer and includes the following three main stages: a) parameter setting, setting a weighted matrix, prediction time domain, and desired flight parameters for each UAV under a distributed MPC framework; b) offline configuration, setting a terminal domain, terminal weighted matrix, and feedback control law that meets the requirements for each UAV; c) online operation, where each UAV performs rolling time domain optimization until the multi-UAV cooperative flight mission ends.
[0065] The specific application steps are as follows:
[0066] 1) Establish a multi-UAV system model for cooperative flight missions:
[0067] This invention considers a cooperative flight mission involving M drones, and describes the system model of the i-th drone (hereinafter referred to as "drone i") in the following general form:
[0068]
[0069] in and These include the system state, measurement output, control input, bounded system disturbances, measurement disturbances, and FDI attacks faced by the UAV i in a networked communication environment. Additionally, f i (·): It describes the system functions of the UAV and has f i (0,0)=0,C i D ix D iy and S i All are constant matrices with appropriate dimensions.
[0070] Consider the following constraints of the aforementioned drone:
[0071]
[0072] Where X i and U i These are closed sets and compact sets, respectively, containing the origin. It should be noted that in this invention, the matrix [D] is assumed to be... iy S i [This refers to a column full rank, i.e., the measured disturbance w] iy and FDI attack a i It is linearly separable.
[0073] 2) Establish a network model for scenarios where communication is compromised:
[0074] This invention considers that drones in cooperative flight missions use wireless networks for information transmission between their own sensors, controllers, and actuators, as well as for information exchange with neighboring drones. Given the bandwidth limitations and traffic congestion in real-world wireless networks, this invention specifically describes the communication impairment scenario as packet loss during information transmission.
[0075] For any drone This invention primarily considers three forms of packet loss: packet loss in the transmission of information from the UAV i's own sensors to the controller channel (hereinafter referred to as "SC packet loss"); packet loss in the transmission of information from the UAV i's own controller to the driver channel (hereinafter referred to as "CA packet loss"); and packet loss in the interaction information between UAV i and its neighboring UAV j. N i Let i be the set of neighboring drones that have a directed communication topology connection with drone i.
[0076] To describe the three types of randomly generated packet loss phenomena mentioned above, this invention describes their processes as follows: and in Taking drone i as an example, the specific form is as follows:
[0077]
[0078] Packet loss is considered to occur in a wireless network environment. and It is not unlimited, but satisfies the following assumption 1.
[0079] Assumption 1: Any drone Its own SC packet loss, CA packet loss, and the packet loss between drone i and any drone Packet loss during information transmission is subject to the maximum consecutive packet loss time limit, which is as follows: and A reliable and latency-free feedback link exists that can detect whether packet loss occurs at any given moment.
[0080] 3) Establish buffer models for each transmission channel:
[0081] Given that all UAVs are in the same network communication environment during cooperative flight missions, this invention will use UAVs as an example. This example illustrates the specific deployment of buffers in each transmission channel.
[0082] Deploy buffers on the SC and CA channels of drone i respectively. and Used to describe its own information transmission, additionally deployed in the transmission channels of drones i and j. buffers Used to describe its information interaction with neighboring drones, where Indicates a collection of neighboring drones cardinality and
[0083] like Figure 2 As shown, for At any time, this invention defines buffer and The information storage lengths are 2N respectively. p +1 and N p ,buffer The storage length is N B ,in And there are N B ≤N p .
[0084] In the buffer and middle, col{x} is used to store the latest state sequence of UAV i under RMO estimation. i (k-2N p -1),…,x i (k-1),x i (k)}, so that in order to Estimate its future state in time; Used in Store the latest optimal control input sequence in real time. when Updates will stop at this time. Used for synchronization The control input information in is It provides input information in a timely manner to perform the corresponding state estimation; Used to keep The latest control input at that time.
[0085] In the buffer middle, Used to store the latest state sequence transmitted by drone j to drone i via a directed network. Where [k] j It is to satisfy And the closest time to the current time k, where [k] is... j ≤k, It is [k] j Time based on optimal control input sequence The obtained predicted state sequence of UAV j,
[0086] 4) Design a robust multivariate observer (RMO):
[0087] As can be seen from Equation (1), during cooperative flight missions, UAVs constantly face system disturbances, measurement disturbances, and FDI attacks. To address this, this invention proposes a novel robust multivariable observer (RMO) for UAVs. The state after disturbance and the measurement disturbance and FDI attack in the output channel are estimated. The estimated state is used as the input of the distributed cooperative predictive controller and the corresponding disturbance is effectively compensated.
[0088] make
[0089]
[0090]
[0091]
[0092]
[0093]
[0094]
[0095] C i0 =[C i D iy S i ]
[0096] The original system (1) can then be described as the following augmented system:
[0097]
[0098] To make z i The left-hand matrix of (k+1) is invertible, further let And add N to both sides of equation (4) i0 C i0 z i (k+1) yields the following equation:
[0099] H i1 z i (k+1)=G i0 z i (k)+F i0 (x i (k),u i (k))+D i0 w ix (k)+Z i0 y i (k)+N i0 y i(k+1) (5)
[0100] Where H i1 =H i0 +N i0 C i0 .
[0101] definition For H i1 The left inverse of the equation allows us to reconstruct the above expression as follows:
[0102]
[0103] in
[0104] make As an augmented state, combining the above equation, we can obtain:
[0105]
[0106] in
[0107] The proposed RMO implementation design is as follows:
[0108]
[0109] Where L i For the gain of the observer to be designed, These are the estimated values for each vector.
[0110] 5) Control input generation and compensation strategies for scenarios with impaired communication:
[0111] Considering both SC and CA packet loss, this invention proposes an efficient control input generation strategy. Under this strategy, the UAV... The control input is not obtained by solving an optimization problem at every moment. To describe this approach, the present invention uses the variable r. i (k) represents whether UAV i needs to be optimized, and its specific form is defined as follows:
[0112]
[0113] Drone i only r i When (k) = 1, the optimization problem is solved, and the buffer is used. In The obtained optimal control sequence Update, in addition The content will replace The content in; if r i If (k) = 0, then the optimization problem is not solved. The content will not be updated, while the buffer... In It will be updated by the control input sequence from the previous moment.
[0114] It should be noted that the complexity of the control algorithm proposed in this invention will depend heavily on the duration of continuous packet loss: the longer the continuous packet loss time, the more complex the judgment condition corresponding to equation (9) will be, which will affect the computational burden of the overall controller to a certain extent.
[0115] 6) Design of a distributed cooperative predictive controller for multiple UAVs:
[0116] To effectively address packet loss originating from neighbor interactions during multi-UAV cooperative flight, this invention designs a cooperative controller based on the concept of distributed MPC. For UAVs... Design the following cost function:
[0117]
[0118] in and For nominal system The predicted state and predicted control input are as follows. Synchronization terms are used to achieve synchronized flight between different drones. For the terminal cost, Q i R i P i and R ij All are diagonal weighted matrices. j (s|k) and x j (s|kn) represents the drone The interactive status information that needs to be transmitted to drone i may not be available in real time due to random packet loss in the transmission channel between neighbors, so a certain compensation strategy is needed to replace it.
[0119] To compensate for the negative impact of the aforementioned packet loss, this invention transforms the cost function (10) into:
[0120]
[0121] in
[0122]
[0123] and for The state information stored in it, κ(x) is the feedback control law.
[0124] The following is a constraint optimization problem for designing a distributed cooperative predictive controller for multiple UAVs under conditions of communication impairment:
[0125]
[0126] in Let Ω be the optimal control input sequence at time k. i (ε i ):={x|x T P i x≤ε i 2} represents the designed terminal domain.
[0127] Based on the basic principles of MPC, The optimal control input is applied to the UAV i. To compensate for the communication impairment caused by packet loss in each channel, when r... i Solve the optimization problem (13) when (k) = 1 and then... Applying to drone i; when r i When (k) = 0, the optimization problem (13) is not solved. This applies to drone i, where k-τ is the nearest time without packet loss (referring to packet loss in neighbor interaction information) from time k. For any drone... Repeat the above steps until the collaborative flight mission is completed.
[0128] The present invention also relates to a computer-readable storage medium storing a distributed cooperative predictive control program for unmanned aerial vehicles (UAVs) in communication-impaired scenarios. When executed by a processor, the program implements the aforementioned distributed cooperative predictive control method for UAVs in communication-impaired scenarios.
[0129] The present invention also relates to a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned distributed cooperative predictive control method for UAVs in communication-impaired scenarios.
[0130] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0131] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0132] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0133] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0134] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.
[0135] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.
Claims
1. A distributed cooperative predictive control method for unmanned aerial vehicles (UAVs) in communication-impaired scenarios, characterized in that: The method described establishes a multi-UAV system model for cooperative flight missions, comprising M UAVs, wherein the i-th UAV satisfies the following at time k. ， in, , , , , and These include the system status, measurement output, control input, bounded system disturbances, measurement disturbances, and the spoofed data injection attacks faced by UAV i in a networked communication environment. To describe the system functions of a drone, C i D ix D iy and S i All are constant matrices with a preset dimension, and the matrices are... To ensure full rank, define constraints for the multi-UAV system model; For each drone, configure a network model and corresponding buffer models for each transmission channel under the scenario of communication impairment. The network model under the communication impairment scenario is as follows: ， in, , and These are packet loss events: packet loss in the transmission of information from UAV i's own sensors to the controller channel; packet loss in the transmission of information from UAV i's own controller to the driver channel; and packet loss in the interaction information between UAV i and its neighboring UAV j. N i Let i be the set of neighboring drones that have a directed communication topology connection with drone i; the maximum consecutive packet loss times are respectively , and The buffer model for each transmission channel includes deploying buffers in the UAV i's own sensor-to-controller channel and its own controller-to-driver channel. and This also includes deploying in the transmission channels between drones i and j. buffers ,in, Indicates a collection of neighboring drones cardinality and ; An improved robust multivariate observer is used to estimate the state of the UAV after disturbance and the interference information in the output channel. The estimated state is used as the input of a distributed cooperative predictive controller to compensate for the corresponding interference, thereby obtaining a control input generation and compensation strategy for communication-impaired scenarios. The improved robust multivariate observer is... ， Among them, L i For the gain of the observer to be designed, These are the estimated values for each vector. , , , , , For H i1 The left reverse, , , , , , , , ; With variable r i (k) indicates whether UAV i needs optimization. ， Drone i only r i Solve the optimization problem when (k) = 1; A distributed cooperative predictive controller for multiple UAVs is generated to achieve distributed cooperative predictive control of UAVs in communication-impaired scenarios. The corresponding constraint optimization problem for designing a distributed cooperative predictive controller for multiple UAVs under conditions of communication impairment is as follows: ， in, Let k be the optimal control input sequence at time k. For the designed terminal domain.
2. The distributed cooperative predictive control method for UAVs in communication-impaired scenarios as described in claim 1, characterized in that: The constraints are satisfied. ， in, and These are closed sets and compact sets, respectively, including the origin.
3. The distributed cooperative predictive control method for UAVs in communication-impaired scenarios according to claim 1, characterized in that: For time k, , , ,buffer and The information storage lengths are 2N respectively. p +1 and N p ,buffer The storage length is N B ,satisfy , and .
4. The distributed cooperative predictive control method for UAVs in communication-impaired scenarios according to claim 1, characterized in that: Define the cost function of drone i. ， in, and For nominal system The predicted state and predictive control input, Q i R i P i and R ij All are diagonal weighted matrices. ,and for The state information stored in it This is a feedback control law.