Underwater robot formation control method based on dynamic coding and decoding

Abstract

The invention discloses an underwater robot formation control method based on dynamic coding and decoding. The method comprises the following steps: S100, constructing a multi-underwater robot system based on a leader robot and follower robot architecture; s200, bit rate allocation is carried out based on a hybrid IVYA-FA optimization algorithm; s300, a coding and decoding strategy based on error information dynamic quantization; s400, according to a finite time domain formation control strategy, a finite time domain formation controller of the follower underwater robot is designed in combination with the bit rate distribution mechanism in the step S200 and the dynamic coding and decoding strategy in the step S300, and the gain of the finite time domain formation controller is obtained by solving a proposed linear matrix inequality condition through an LMI toolbox; and S500, realizing formation control of the multi-underwater robot system. Compared with the prior art, the method has the advantages that the optimal utilization of limited channel resources by the robot can be effectively improved, the saturation error problem caused by the quantitative saturation phenomenon is reduced, and the stability and robustness of the system are improved.

Description

Technical Field

[0001] The invention relates to underwater robot formation control technology, in particular to an underwater robot formation control method based on dynamic coding and decoding. Background Art

[0002] The underwater multi-autonomous robot system is composed of many autonomous underwater robots with sensing, computing and execution capabilities. It exchanges information and makes distributed decisions through optical, acoustic and other communication modules, ultimately achieving efficient underwater group collaboration and group intelligence. However, in underwater environments such as the ocean, the propagation loss of seawater, signal propagation delay and multipath effects limit the bandwidth and transmission rate of underwater communication. At the same time, the existence of complex factors such as noise, water flow and seabed topography in the underwater environment further aggravates the difficulty of communication. The limited underwater communication caused by these complex factors seriously limits the efficient collaboration of underwater multi-autonomous system.

[0003] Existing research on formation control of underwater multi-autonomous robot systems under limited network resources usually starts from the perspective of control theory, considering communication phenomena such as packet loss and delay in the transmission process between nodes, and designs various types of control strategies to achieve the coordination of underwater multi-autonomous robot systems. However, existing methods lack consideration of the degree of bandwidth limitation of underwater networks, ignore the coupling relationship between the bandwidth of underwater communication networks and cooperative control protocols, and thus cannot effectively utilize the limited bandwidth resources in underwater communication networks to ensure the accuracy of communication data of underwater multi-autonomous robot systems and the coordination of systems.

[0004] In the formation coordination process of underwater multi-autonomous robot systems, the system information and state / control signals are often transmitted through digital communication channels. These information are usually quantized before transmission (i.e., real-valued signals are mapped to piecewise constant signals with values ​​in a finite set), and then the codec is used to achieve dual-end data sharing. However, when the system quantizes the transmission signal, when the input value exceeds the processing range of the quantizer, the excess value will be truncated, resulting in loss of transmission signal accuracy, which seriously affects the stability of the system. In the existing formation control methods for underwater multi-autonomous robot systems, the quantization strategy adopted is still to select a sufficiently large quantization range to weaken the impact of quantization saturation on the system. This quantization method has high requirements on the number of data quantization levels and cannot be applied to bandwidth-limited underwater communication networks.

[0005] Therefore, it is urgent to study an underwater robot formation control method based on dynamic encoding and decoding to achieve efficient collaboration and stability of underwater multi-autonomous robot systems in an environment with limited network bandwidth. Summary of the invention

[0006] In view of the above-mentioned defects, the purpose of the present invention is to propose a control strategy for the formation of multiple underwater robot systems in an underwater limited bandwidth environment. Compared with the existing methods, it can effectively improve the optimal utilization of limited channel resources by robots and reduce the saturation error problem caused by the quantization saturation phenomenon, thereby improving the stability and robustness of the system.

[0007] To achieve this object, the present invention adopts the following technical solutions:

[0008] A method for controlling an underwater robot formation based on dynamic coding and decoding comprises the following steps:

[0009] S100, constructing a multi-underwater robot system based on a leader robot and a follower robot architecture, wherein a leader interacts with multiple follower robots through an underwater communication network, the leader sends information, the followers receive information from the leader, and the followers send and receive information to each other;

[0010] S200, based on the hybrid IVYA-FA optimization algorithm, bit rate allocation is performed, and according to the characteristics of the underwater digital communication network, the underwater network bandwidth is characterized in the form of bit rate; according to the communication requirements of different underwater robots, the bit rate of each underwater robot is allocated based on the distribution type MAC protocol, and the bit rate allocation mechanism of the hybrid IVYA-FA optimization algorithm is used for allocation;

[0011] S300, based on the coding and decoding strategy of dynamic quantization of error information, the follower underwater robot estimates its own state information through the attitude sensor, integrates it into an estimated state vector, and quantizes the estimated state vector; the dynamic quantizer realizes the dynamic quantization of the underwater robot transmission data by dynamically adjusting parameters; the processor synthesizes the estimated state vector and the dynamic adjustment parameters into a data packet, and sends it to the communication network through the underwater acoustic sensor;

[0012] S400, finite time domain formation control strategy, combining the bit rate allocation mechanism in step S200 and the dynamic encoding and decoding strategy in step S300 to design a finite time domain formation controller of the follower underwater robot, and solving the proposed linear matrix inequality conditions through the LMI toolbox to obtain the finite time domain formation controller gain;

[0013] S500, realize formation control of multiple underwater robot systems, design the controller of each follower underwater robot based on the finite time domain formation control gain solved in step S400, calculate the control input of the underwater robot at the next movement moment according to the system state information obtained in the dynamic encoding and decoding process of step S300, and transmit it to the actuator of the underwater robot to realize formation control of multiple underwater robot systems in a finite time domain.

[0014] Preferably, the bit rate allocation mechanism of the hybrid IVYA-FA optimization algorithm in step S200 is as follows:

[0015] S210, initialize system parameters, consider the system as a multi-underwater robot system including a leader and N followers; the system state variable of the leader is x l , the follower's system state variable is x i , the initial value of the follower's system state is x i0 , calculate the sum of the initial values ​​of the state variables of the follower robot s 0 , set the quantization limit value d of the follower robot i , set the minimum dynamic adjustment parameters of the follower robot

[0016] S220, construct an optimization objective function, and construct the following optimization objective function according to the system preset parameters:

[0017]

[0018] Among them, f p (R) is the penalty function, n x Represents the number of elements in the state variable vector, R = [R 1 ,R 2 ,…,R N ];

[0019] S230, perform optimal bit rate allocation, perform optimal bit rate allocation of the follower underwater robot by using the hybrid IVYA-FA optimization algorithm, and calculate the optimal bit rate allocation result R best , set the maximum bit rate of the corresponding follower underwater robot.

[0020] Preferably, the hybrid IVYA-FA optimization algorithm comprises the following steps:

[0021] S231, Initialization parameters, Initialization parameters of ivy population: X min , X max 、T max and N pop ; Let T = 1;

[0022] Initialize firefly population parameters: light absorption intensity γ, maximum attraction θ max , random term coefficient a, maximum number of iterations T max and the firefly population size N pop ;

[0023] Calculating the ivy population The initial value of X is calculated by: i =Xmin +rand(1,D)⊙(X max -X min );

[0024] Randomly initialize N pop The position of the firefly Y i (i=1,2,…,N pop );

[0025] S232. Calculate the growth direction of the ivy, calculate the growth direction ΔGv of the ivy i , calculation formula:

[0026] S233, calculate the objective function value of all individuals in the initial population, according to the fitness function F(X i ) Calculate the objective function values ​​of all individuals in the initial population and sort them from best to worst according to their fitness values, and finally set X best =X 1 ;

[0027] S234. Calculate the luminous intensity of the firefly population; calculate N pop The luminous intensity of a firefly is L i (i=1,2,…,N pop );

[0028] S235, mixed population iterative optimization, when T is less than or equal to T max When , execute the ivy iterative optimization function, firefly iterative optimization function and mixed population optimization result and construct the optimal allocation set function until T is greater than Tmax;

[0029] S2351, Ivy search iteration, i starts from 1, and the value of i increases by 1 after each execution of the Ivy search function until the value of i is greater than N pop When , execute the firefly population optimization iteration;

[0030] S2352, Ivy League search, select the best neighbor X li , sort the members of population X from high to low according to their fitness values ​​to form a new set: The best neighbor is selected as follows:

[0031]

[0032] Calculate β = (2 + rand) / 2 and ΔGv i =rand 2 ·(N(1,D)⊙ΔGv i );

[0033] If F(X i )<β·F(Xbest ), then execute formula (1). If F(X i )<β·F(X best ), then execute formula (2);

[0034]

[0035] calculate The fitness value of

[0036] Will Add to set X as the i-th element new and set its growth direction to

[0037] S2353, Firefly population optimization iteration, j starts from 1, and the value of j increases by 1 after each execution of the firefly population optimization until the value of j is greater than N pop ;

[0038] S2354, Firefly population optimization, if L is satisfied i <L j , execute the ivy optimization iteration, otherwise, do nothing; i starts from 1, and the value of i increases by 1 after each ivy optimization iteration until the value of i is greater than N pop ;

[0039] Execution (3)

[0040]

[0041] Among them, r ij is the Cartesian distance between two fireflies;

[0042] S2355, Mix the population optimization results and build the optimal allocation set, merge the two populations X new and Y new , expressed as M = {X new ,Y new};

[0043] Calculate the fitness value of each member in M ​​according to the fitness function, and sort them from high to low to form a new set M s , then, order

[0044] S236, output the optimal bit rate allocation result, the mixed population iterates and loops until T is greater than Tmax, jumps out of the loop and outputs the optimal bit rate allocation result according to the last optimization result of S5, and obtains the final

[0045] Preferably, in step S300, the estimated state vector is The dynamic adjustment parameter is μ i , the selection rules of dynamic adjustment parameters are as follows:

[0046]

[0047] in, The setting rules meet the following requirements: Usually choose Since the decoding end also needs to obtain the dynamic adjustment parameter μ in real time i Therefore, the encoder needs to i Perform binary encoding; dynamically adjust parameter μ i It is expressed by the following formula:

[0048]

[0049] Among them, α i and β i Expressed as the encoder pair dynamically adjusts the parameter μ i (k) the number of bits used in binary encoding;

[0050] Because μ i The value of (k) is not unique, so we set the following i (k) value selection rules:

[0051]

[0052] Considering the limited communication bandwidth, in and The value is set according to the number of bits allocated in actual situation.

[0053] Further, in step S400,

[0054] Design the finite-time formation controller of the follower underwater robot as follows:

[0055]

[0056] Among them, K i is the control gain of the ith follower underwater robot, a ij is the connection weight between follower underwater robot i and follower underwater robot j; b i is the connection weight between the leader and the follower underwater robot;

[0057] In addition, the communication topology of the follower layer subsystem of the multi-underwater robot system is composed of a symmetric and irreducible matrix Description, as follows:

[0058]

[0059] If there is a coupling relationship between follower underwater robots i and j, then there exists a ij =a ji ≠0; if there is no coupling relationship, then there is a ij =a ji =0;

[0060] In order to obtain the optimal gain of the finite-time formation controller of the follower underwater robot, an optimization problem is solved by using the LMI toolbox to obtain the gain K of the finite-time formation controller.

[0061] The specific steps are as follows:

[0062] 1) Define matrix variables, and use the lmivar function to define variables μ≥1, Variable Matrix and positive definite variable matrices P and Q;

[0063] 2) Construct LMI constraints and construct the following inequality conditions based on system parameters:

[0064]

[0065] in, I is the identity matrix, and They represent the square of the inner product of the initial state difference vector between the follower and the leader, the square of the inner product of the external interference vector, and the square of the inner product of the decoding error vector;

[0066] 3) Set the optimization goal∈ 2 , the optimization goal is: min(∈ 2 );

[0067] 4) Solve the LMI optimization problem and obtain the optimal matrix using the solver provided by the LMI toolbox Finally, calculate by the following formula:

[0068]

[0069] The gain K of the finite-time formation controller is obtained.

[0070] One of the above technical solutions includes the following beneficial effects: 1. The bit rate allocation mechanism based on the hybrid IVYA-FA optimization algorithm proposed in this patent can achieve the optimal bit rate allocation for robots with different characteristics in an underwater limited bandwidth environment. Compared with the existing methods, it can effectively improve the optimal utilization of limited channel resources by robots and improve the stability of the system; 2. The encoding and decoding strategy based on dynamic quantization of error information proposed in this patent can minimize the impact of quantization saturation on the stability of the system under limited channel resources. Compared with the existing methods, it can make full use of limited underwater channel resources, weaken the impact of signal errors caused by quantization saturation on the system, and improve the robustness and coordination performance of the multi-underwater robot system in a limited bit rate environment. 3. The limited time domain formation control strategy designed by this patent can achieve the underwater robot formation to maintain stable control performance in a preset time domain. Compared with the existing methods, this patent designs a limited time domain formation control method for a multi-underwater robot system in a limited bit rate environment, which improves the safety and stability of formation control of a multi-underwater robot system in a specific time domain. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a schematic diagram of the overall steps of an embodiment of the present invention;

[0072] Figure 2 is a schematic diagram of a bit rate allocation mechanism according to an embodiment of the present invention;

[0073] Figure 3 is a schematic diagram of the steps of a hybrid IVYA-FA optimization algorithm according to an embodiment of the present invention;

[0074] Figure 4 It is a schematic diagram of the mixed population iterative optimization step of an embodiment of the present invention. DETAILED DESCRIPTION

[0075] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0076] like Figure 1 As shown, a method for controlling an underwater robot formation based on dynamic coding and decoding comprises the following steps:

[0077] S100, constructing a multi-underwater robot system based on a leader robot and a follower robot architecture, wherein a leader interacts with multiple follower robots through an underwater communication network, the leader sends information, the followers receive information from the leader, and the followers send and receive information to each other;

[0078] S200, based on the hybrid IVYA-FA optimization algorithm, bit rate allocation is performed, and according to the characteristics of the underwater digital communication network, the underwater network bandwidth is characterized in the form of bit rate; according to the communication requirements of different underwater robots, the bit rate of each underwater robot is allocated based on the distribution type MAC protocol, and the bit rate allocation mechanism of the hybrid IVYA-FA optimization algorithm is used for allocation;

[0079] S300, based on the coding and decoding strategy of dynamic quantization of error information, the follower underwater robot estimates its own state information through the attitude sensor, integrates it into an estimated state vector, and quantizes the estimated state vector; the dynamic quantizer realizes the dynamic quantization of the underwater robot transmission data by dynamically adjusting parameters; the processor synthesizes the estimated state vector and the dynamic adjustment parameters into a data packet, and sends it to the communication network through the underwater acoustic sensor;

[0080] S400, finite time domain formation control strategy, combining the bit rate allocation mechanism in step S200 and the dynamic encoding and decoding strategy in step S300 to design a finite time domain formation controller of the follower underwater robot, and solving the proposed linear matrix inequality conditions through the LMI toolbox to obtain the finite time domain formation controller gain;

[0081] S500, realize formation control of multiple underwater robot systems, design the controller of each follower underwater robot based on the finite time domain formation control gain solved in step S400, calculate the control input of the underwater robot at the next movement moment according to the system state information obtained in the dynamic encoding and decoding process of step S300, and transmit it to the actuator of the underwater robot to realize formation control of multiple underwater robot systems in a finite time domain.

[0082] Beneficial effects: 1. The bit rate allocation mechanism based on the hybrid IVYA-FA optimization algorithm proposed in this patent can achieve the optimal bit rate allocation for robots with different characteristics in an underwater limited bandwidth environment. Compared with the existing methods, it can effectively improve the optimal utilization of limited channel resources by robots and improve the stability of the system; 2. The encoding and decoding strategy based on dynamic quantization of error information proposed in this patent can minimize the impact of quantization saturation on the stability of the system under limited channel resources. Compared with the existing methods, it can make full use of limited underwater channel resources, weaken the impact of signal errors caused by quantization saturation on the system, and improve the robustness and coordination performance of multi-underwater robot systems in a limited bit rate environment. 3. The limited time domain formation control strategy designed by this patent can achieve stable control performance of underwater robot formations in a preset time domain. Compared with the existing methods, this patent designs a limited time domain formation control method for a multi-underwater robot system in a limited bit rate environment, which improves the safety and stability of formation control of a multi-underwater robot system in a specific time domain.

[0083] like Figure 2 As shown, the bit rate allocation mechanism of the hybrid IVYA-FA optimization algorithm in step S200 is as follows:

[0084] S210, initialize system parameters, consider the system as a multi-underwater robot system including a leader and N followers; the system state variable of the leader is x l , the follower's system state variable is x i , the initial value of the follower's system state is x i0 , calculate the sum of the initial values ​​of the state variables of the follower robot s 0 , set the quantization limit value d of the follower robot i , set the minimum dynamic adjustment parameters of the follower robot

[0085] S220, construct an optimization objective function, and construct the following optimization objective function according to the system preset parameters:

[0086]

[0087] Among them, f p (R) is the penalty function, n x Represents the number of elements in the state variable vector, R = [R 1 ,R 2 ,…,R N ];

[0088] S230, perform optimal bit rate allocation, perform optimal bit rate allocation of the follower underwater robot by using the hybrid IVYA-FA optimization algorithm, and calculate the optimal bit rate allocation result R best , set the maximum bit rate of the corresponding follower underwater robot.

[0089] like Figure 3 As shown, the hybrid IVYA-FA optimization algorithm includes the following steps:

[0090] S231, Initialization parameters, Initialization parameters of ivy population: X min , X max , T max and N po p; let T = 1;

[0091] Initialize firefly population parameters: light absorption intensity γ, maximum attraction θ max , random term coefficient a, maximum number of iterations T max and the firefly population size N pop ;

[0092] Calculating the ivy population The initial value of X is calculated by: i =X min +rand(1,D)⊙(X max -X min );

[0093] Randomly initialize N pop The position of the firefly Y i (i=1,2,…,N pop );

[0094] S232, calculate the growth direction of the ivy, calculate the growth direction of the ivy ΔGv i , calculation formula:

[0095] S233, calculate the objective function value of all individuals in the initial population, according to the fitness function F(X i ) Calculate the objective function values ​​of all individuals in the initial population and sort them from best to worst according to their fitness values, and finally set X best =X 1 ;

[0096] S234. Calculate the luminous intensity of the firefly population; calculate N pop The luminous intensity of a firefly is L i (i=1,2,…,N pop );

[0097] like Figure 4As shown, S235, mixed population iterative optimization, when T is less than or equal to T max When , execute the ivy iterative optimization function, firefly iterative optimization function and mixed population optimization result and construct the optimal allocation set function until T is greater than Tmax;

[0098] S2351, Ivy search iteration, i starts from 1, and the value of i increases by 1 after each execution of the Ivy search function until the value of i is greater than N pop When , the firefly population optimization iteration is executed;

[0099] S2352, Ivy League search, select the best neighbor X li , sort the members of population X from high to low according to their fitness values ​​to form a new set: The best neighbor is selected as follows:

[0100]

[0101] Calculate β = (2 + rand) / 2 and ΔGv i =rand 2 ·(N(1,D)⊙ΔGv i );

[0102] If F(X i )<β·F(X best ), then execute formula (1). If F(X i )<β·F(X best ), then execute formula (2);

[0103]

[0104] calculate The fitness value of

[0105] Will Add to set X as the i-th element new and set its growth direction to

[0106] S2353, Firefly population optimization iteration, j starts from 1, and the value of j increases by 1 after each execution of the firefly population optimization until the value of j is greater than N pop ;

[0107] S2354, Firefly population optimization, if L is satisfied i <L j , execute the ivy optimization iteration, otherwise, do nothing; i starts from 1, and the value of i increases by 1 after each ivy optimization iteration until the value of i is greater than Npop ;

[0108] Execution (3)

[0109]

[0110] Among them, r ij is the Cartesian distance between two fireflies;

[0111] S2355, Mix the population optimization results and build the optimal allocation set, merge the two populations X new and Y new , expressed as M = {X new ,Y new};

[0112] Calculate the fitness value of each member in M ​​according to the fitness function, and sort them from high to low to form a new set M s , then, order

[0113] S236, output the optimal bit rate allocation result, the mixed population iterates and loops until T is greater than Tmax, jumps out of the loop and outputs the optimal bit rate allocation result according to the last optimization result of S5, and obtains the final

[0114] The hybrid IVYA-FA optimization algorithm effectively enhances the overall efficiency of the system's search for the optimal solution by increasing population diversity, and helps to quickly find the optimal bit rate allocation scheme based on the differential characteristics of system nodes, thereby making full use of limited underwater bandwidth resources to achieve formation control of multiple underwater robot systems.

[0115] In addition, in step S300, the estimated state vector is The dynamic adjustment parameter is μ i , the selection rules of dynamic adjustment parameters are as follows:

[0116]

[0117] in, The setting rules meet the following requirements: Usually choose Since the decoding end also needs to obtain the dynamic adjustment parameter μ in real time i Therefore, the encoder needs to i Perform binary encoding; dynamically adjust parameter μ i It is expressed by the following formula:

[0118]

[0119] Among them, α i and β iExpressed as the encoder pair dynamically adjusts the parameter μ i (k) the number of bits used in binary encoding;

[0120] Because μ i The value of (k) is not unique, so we set the following i (k) value selection rules:

[0121]

[0122] Considering the limited communication bandwidth, in and The value is set according to the number of bits allocated in actual situation.

[0123] In order to address the quantization saturation problem that exists when an underwater robot transmits its own state value, this patent proposes a dynamic encoding and decoding strategy based on a limited bit rate, and performs dual-end encoding and decoding on the dynamic adjustment factor based on the state information, which can effectively reduce the quantization error caused by the quantization saturation phenomenon, thereby improving the formation control performance of the entire multi-underwater robot system under limited communication resources.

[0124] In addition, in step S400,

[0125] Design the finite-time formation controller of the follower underwater robot as follows:

[0126]

[0127] Among them, K i is the control gain of the ith follower underwater robot, a ij is the connection weight between follower underwater robot i and follower underwater robot j; b i is the connection weight between the leader and the follower underwater robot;

[0128] In addition, the communication topology of the follower layer subsystem of the multi-underwater robot system is composed of a symmetric and irreducible matrix Description, as follows:

[0129]

[0130] If there is a coupling relationship between follower underwater robots i and j, then there exists a ij =a ji ≠0; if there is no coupling relationship, then there is a ij =a ji =0;

[0131] In order to obtain the optimal gain of the finite-time formation controller of the follower underwater robot, an optimization problem is solved by using the LMI toolbox to obtain the gain K of the finite-time formation controller.

[0132] The specific steps are as follows:

[0133] 1) Define matrix variables, and use the lmivar function to define variables μ≥1, Variable Matrix and positive definite variable matrices P and Q;

[0134] 2) Construct LMI constraints and construct the following inequality conditions based on system parameters:

[0135]

[0136] in, I is the identity matrix, and They represent the inner product of the difference vectors of the follower and the leader’s initial states.

[0137] The square of the inner product of the external interference vector and the square of the inner product of the decoding error vector;

[0138] 3) Set the optimization goal∈ 2 , the optimization goal is: min(∈ 2 );

[0139] 4) Solve the LMI optimization problem and obtain the optimal matrix using the solver provided by the LMI toolbox Finally, calculate by the following formula:

[0140]

[0141] The gain K of the finite-time formation controller is obtained.

[0142] Combined with the optimal bit rate allocation results and the constraints of dynamically adjusting parameters, the LMI toolbox is used to solve the optimal problem of the minimum error upper bound and obtain the finite time domain formation controller gain K. The controller gain K solved in this way fully considers the communication resource limitation and quantization saturation problem under the bit rate constraint environment, and can ensure the stability and robustness of the multi-underwater robot system in the formation control process.

[0143] The technical principle of the present invention is described above in conjunction with specific embodiments. These descriptions are only for explaining the principle of the present invention and cannot be interpreted as limiting the scope of protection of the present invention in any way. Based on the explanations herein, those skilled in the art can associate other specific implementations of the present invention without paying creative labor, and these methods will fall within the scope of protection of the present invention.

Claims

1. A method for controlling underwater robot formation based on dynamic coding and decoding, characterized in that: The steps include: S100, constructing a multi-underwater robot system based on a leader robot and a follower robot architecture, wherein a leader interacts with multiple follower robots through an underwater communication network, the leader sends information, the followers receive information from the leader, and the followers send and receive information to each other; S200, based on the hybrid IVYA-FA optimization algorithm, bit rate allocation is performed, and according to the characteristics of the underwater digital communication network, the underwater network bandwidth is characterized in the form of bit rate; according to the communication requirements of different underwater robots, the bit rate of each underwater robot is allocated based on the distribution type MAC protocol, and the bit rate allocation mechanism of the hybrid IVYA-FA optimization algorithm is used for allocation; S300, based on the coding and decoding strategy of dynamic quantization of error information, the follower underwater robot estimates its own state information through the attitude sensor, integrates it into an estimated state vector, and quantizes the estimated state vector; the dynamic quantizer realizes the dynamic quantization of the underwater robot transmission data by dynamically adjusting parameters; the processor synthesizes the estimated state vector and the dynamic adjustment parameters into a data packet, and sends it to the communication network through the underwater acoustic sensor; S400, finite time domain formation control strategy, combining the bit rate allocation mechanism in step S200 and the dynamic encoding and decoding strategy in step S300 to design a finite time domain formation controller of the follower underwater robot, and solving the proposed linear matrix inequality conditions through the LMI toolbox to obtain the finite time domain formation controller gain; S500, realize formation control of multiple underwater robot systems, design the controller of each follower underwater robot based on the finite time domain formation control gain solved in step S400, calculate the control input of the underwater robot at the next movement moment according to the system state information obtained in the dynamic encoding and decoding process of step S300, and transmit it to the actuator of the underwater robot to realize formation control of multiple underwater robot systems in a finite time domain.

2. The underwater robot formation control method based on dynamic coding and decoding according to claim 1 is characterized in that: The bit rate allocation mechanism of the hybrid IVYA-FA optimization algorithm in step S200 is as follows: S210, initialize system parameters, consider the system as a multi-underwater robot system including a leader and N followers; the system state variable of the leader is x l , the follower's system state variable is x i , the initial value of the follower's system state is x i0 , calculate the sum of the initial values ​​of the state variables of the follower robot s0, and set the quantization limit value d of the follower robot i , set the minimum dynamic adjustment parameters of the follower robot S220, construct an optimization objective function, and construct the following optimization objective function according to the system preset parameters: Among them, f p (R) is the penalty function, n x Represents the number of elements in the state variable vector, R = [R1, R2, …, R N ]; S230, perform optimal bit rate allocation, perform optimal bit rate allocation of the follower underwater robot by using the hybrid IVYA-FA optimization algorithm, and calculate the optimal bit rate allocation result R best , set the maximum bit rate of the corresponding follower underwater robot.

3. The underwater robot formation control method based on dynamic coding and decoding according to claim 1 is characterized in that: The hybrid IVYA-FA optimization algorithm consists of the following steps: S231, Initialization parameters, Initialization parameters of ivy population: X min , X max 、T max and N pop ; Let T = 1; Initialize firefly population parameters: light absorption intensity γ, maximum attraction θ max , random term coefficient a, maximum number of iterations T max and the firefly population size N pop ; Calculating the ivy population The initial value of is calculated as: X i =X min +rand(1,D)⊙(X max -X min ); Randomly initialize N pop The position of the firefly Y i (i=1,2,…,N pop ); S232. Calculate the growth direction of the ivy, calculate the growth direction ΔGv of the ivy i , calculation formula: S233, calculate the objective function value of all individuals in the initial population, according to the fitness function F(X i ) Calculate the objective function values ​​of all individuals in the initial population and sort them from best to worst according to their fitness values, and finally set X best =X1; S234. Calculate the luminous intensity of the firefly population; calculate N pop The luminous intensity of a firefly is L i (i=1,2,…,N pop ); S235, mixed population iterative optimization, when T is less than or equal to T max When , execute the ivy iterative optimization function, firefly iterative optimization function and mixed population optimization result and construct the optimal allocation set function until T is greater than Tmax; S2351, Ivy search iteration, i starts from 1, and the value of i increases by 1 after each execution of the Ivy search function until the value of i is greater than N pop When , execute the firefly population optimization iteration; S2352, Ivy League search, select the best neighbor X li , sort the members of population X from high to low according to their fitness values ​​to form a new set: The best neighbor is selected as follows: Calculate β = (2 + rand) / 2 and ΔGv i =rand 2 ·(N(1,D)⊙ΔGv i ); If F(X i )<β·F(X best ), then execute formula (1). If F(X i )<β·F(X best ), then execute formula (2); calculate The fitness value of Will Add to set X as the i-th element new and set its growth direction to S2353, Firefly population optimization iteration, j starts from 1, and the value of j increases by 1 after each execution of the firefly population optimization until the value of j is greater than N pop ; S2354, Firefly population optimization, if L is satisfied i <L j , execute the ivy optimization iteration, otherwise, do nothing; i starts from 1, and the value of i increases by 1 after each ivy optimization iteration until the value of i is greater than N pop ; Execution (3) Among them, r ij is the Cartesian distance between two fireflies; S2355, Mix the population optimization results and build the optimal allocation set, merge the two populations X new and Y new , expressed as M = {X new ,Y new }; Calculate the fitness value of each member in M ​​according to the fitness function, and sort them from high to low to form a new set M s , then, order S236, output the optimal bit rate allocation result, the mixed population iterates and loops until T is greater than Tmax, jumps out of the loop and outputs the optimal bit rate allocation result according to the last optimization result of S5, and obtains the final 4. The underwater robot formation control method based on dynamic coding and decoding according to claim 1 is characterized in that: In step S300, the estimated state vector is The dynamic adjustment parameter is μ i , the selection rules of dynamic adjustment parameters are as follows: in, The setting rules meet the following requirements: Usually choose Since the decoding end also needs to obtain the dynamic adjustment parameter μ in real time i Therefore, the encoder needs to i Perform binary encoding; dynamically adjust parameter μ i It is expressed by the following formula: Among them, α i and β i Expressed as the encoder pair dynamically adjusts the parameter μ i (k) the number of bits used in binary encoding; Because μ i The value of (k) is not unique, so we set the following i (k) value selection rules: Considering the limited communication bandwidth, in and The value is set according to the number of bits allocated in actual situation.

5. The underwater robot formation control method based on dynamic coding and decoding according to claim 1 is characterized in that: In step S400, Design the finite-time formation controller of the follower underwater robot as follows: Among them, K i is the control gain of the ith follower underwater robot, a ij is the connection weight between follower underwater robot i and follower underwater robot j; b i is the connection weight between the leader and the follower underwater robot; In addition, the communication topology of the follower layer subsystem of the multi-underwater robot system is composed of a symmetric and irreducible matrix Description, as follows: If there is a coupling relationship between follower underwater robots i and j, then there exists a ij =a ij ≠0; if there is no coupling relationship, then there is a ij =a ij =0; In order to obtain the optimal gain of the finite-time formation controller of the follower underwater robot, an optimization problem is solved by using the LMI toolbox to obtain the gain K of the finite-time formation controller. The specific steps are as follows: 1) Define matrix variables, and use the lmivar function to define variables μ≥1, Variable Matrix and positive definite variable matrices P and Q; 2) Construct LMI constraints and construct the following inequality conditions based on system parameters: in, I is the identity matrix, and They represent the square of the inner product of the initial state difference vector between the follower and the leader, the square of the inner product of the external interference vector, and the square of the inner product of the decoding error vector; 3) Set the optimization goal∈ 2 , the optimization goal is: min(∈ 2 ); 4) Solve the LMI optimization problem and obtain the optimal matrix using the solver provided by the LMI toolbox Finally, calculate by the following formula: The gain K of the finite-time formation controller is obtained.

Images