Intelligent agricultural machine automatic driving platoon prediction control method for communication impaired scenarios

Abstract

The application claims a smart agricultural machine automatic driving platoon prediction control method for a communication impaired scene, comprising the following steps: S1, a single smart agricultural machine particle model is cited, a kinematics equation is established, and the kinematics equation is further expanded to the entire smart agricultural machine platoon; S2, a smart agricultural machine discrete model is established according to the kinematics equation, and a buffer is set to store data delay caused by communication impairment; S3, a front vehicle-follower platoon control structure mode is adopted, and safety performance constraint requirements of the smart agricultural machine under the conditions of communication impairment and external disturbance are established; S4, an optimal control problem is constructed, the actual running track and safety performance limit of the smart agricultural machine are considered, a maximum-minimum model predictive control strategy is introduced, an online optimal controller satisfying the conditions is solved, and optimal control input is obtained. The optimal controller under the communication impairment is obtained by considering online solving under the maximum disturbance, and efficient smart agricultural machine platoon collective work is realized.

Description

Technical Field

[0001] This invention relates to the field of agricultural automation and unmanned driving technology, and in particular to the research and implementation of a predictive control method for intelligent agricultural machinery autonomous driving formation in scenarios where communication is impaired. Background Technology

[0002] With the rapid development of global agricultural technology, traditional agriculture is gradually moving towards modernization and intelligence. Building upon the development of agricultural mechanization, smart agriculture and automation technologies are becoming important trends in global agricultural development, significantly improving agricultural production efficiency and resource utilization while reducing the labor intensity of farmers. Against this backdrop, intelligent agricultural machinery autonomous driving platooning technology has gradually become a hot research topic in the field of agricultural machinery. Intelligent agricultural machinery autonomous driving platooning technology refers to the use of high-precision navigation, real-time communication, path planning, and collaborative control technologies to enable multiple agricultural machines to operate synchronously and collaboratively in the field in a platoon formation. This technology is based on Global Navigation Satellite System (GNSS), sensors, artificial intelligence (AI), and automatic control algorithms to ensure that agricultural machinery efficiently and accurately completes agricultural production tasks such as tilling, sowing, fertilizing, and harvesting in an unmanned state.

[0003] However, despite the enormous potential of intelligent agricultural machinery platooning technology in multiple fields, its development still faces many challenges, especially in platooning control strategies. Platooning control involves the dynamic cooperation of multiple agricultural machines, requiring consideration of many complex factors, including: dynamics and kinematic models: each intelligent agricultural machine has its own unique dynamic characteristics, such as speed and acceleration. Under non-ideal conditions (such as wind speed and airflow), changes in these characteristics can affect the stability of the platoon and the efficiency of task execution. Communication latency and network communication: the effectiveness of intelligent agricultural machinery often depends on real-time information exchange between individual machines. However, in practical applications, communication between machines is often impaired by factors such as latency, interference, and data loss. Therefore, how to effectively conduct collaborative control while considering communication impairments becomes an important issue. The core of intelligent agricultural machinery platooning is the collaborative algorithm of multiple intelligent agricultural machines. Currently, various control algorithms have been applied to intelligent agricultural machinery platooning, such as the leader-follower model and virtual navigation. Although these algorithms have achieved success under certain conditions, there are still risks of failing to adequately address complex dynamic environments.

[0004] Currently, much research focuses on model-based control methods, such as model predictive control (MMC), which is considered capable of effectively enabling multi-machine collaboration in dynamic environments. In practical applications, by establishing a dynamic model of the system, predicting future states in real time, and performing optimized control, the predictive characteristics of the system can be utilized to effectively cope with delays and unexpected situations. However, existing MMC methods often do not fully consider the impact of communication delays on control performance. Coupled with uncertain external disturbances caused by factors such as wind and mechanical friction, this can lead to a decrease in the stability and reliability of the system in real-world environments. To address these technical backgrounds and challenges, this invention proposes a predictive control method for automated driving platooning of intelligent agricultural machinery in scenarios with impaired communication. This method aims to overcome the negative impacts of communication impairments and uncertain external disturbances, thereby improving the collaborative efficiency and stability of intelligent agricultural machinery platoons. By introducing communication delay modeling and a design based on max-min model predictive control, we plan to achieve a highly efficient and reliable intelligent agricultural machinery control system, providing a new solution for the widespread application of intelligent agricultural machinery technology.

[0005] The research of this invention not only aims to contribute to the development of intelligent agricultural machinery platoon control technology, but also hopes to improve the collaborative working ability and task execution efficiency of intelligent agricultural machinery in complex scenarios in practical applications.

[0006] A search revealed application publication number CN116466722A, which discloses a human-machine integrated intelligent agricultural machinery automatic driving platooning transfer system and method. The system is installed on a navigator and several follower vehicles. Each navigator and follower vehicle is equipped with the same driving status acquisition system. The navigator also has a platooning control terminal system, a router master station, and a host WiFi module connected in sequence. Each follower vehicle has a follower WiFi module that wirelessly communicates with the host WiFi module. All collected information is transmitted to the platooning control terminal system via the router master station. The onboard computer in the platooning control terminal system obtains the machine adjustment decisions from the follower vehicles. The navigator operator transmits the human decisions to the onboard computer through a human-machine interface. The onboard computer integrates the human decisions and machine adjustment decisions to generate platooning decisions and control commands for control input.

[0007] Despite its numerous advantages, the design and implementation of a human-machine integrated intelligent agricultural machinery automatic driving platooning transfer system and method still has some shortcomings. These mainly include the following aspects:

[0008] 1. Stability of communication and data transmission. In practical application scenarios, the complex farmland environment, signal interference, and multipath effects may lead to communication instability, affecting the real-time performance of data transmission; human-machine integrated systems may face delays in data acquisition and feedback processes, which can impact the system's real-time control and decision-making.

[0009] 2. Insufficient system robustness. The system is sensitive to external disturbances; under certain extreme conditions (such as severe weather or complex terrain), the automatic driving system may be affected, leading to decreased relocation efficiency. Compatibility and coordination between different models of agricultural machinery, sensors, and controllers may also present problems, impacting the overall system performance. 3. Safety and feasibility. Automatic driving technology may fail in certain situations, especially during relocation, potentially leading to equipment collisions or safety issues in the work area.

[0010] Therefore, communication impairment and interference issues can be considered during the modeling stage, along with security performance constraints to ensure safety. Simultaneously, the design of robust dynamic models and control algorithms should be improved. By incorporating environmental monitoring data, more robust control algorithms can be designed to adapt to different operating conditions, improving system reliability. Furthermore, the fusion processing of multiple sensors can enhance environmental perception capabilities, reduce reliance on single sensors, and increase the system's adaptability. Summary of the Invention

[0011] This invention aims to solve the problems of the prior art. It proposes a predictive control method for intelligent agricultural machinery automatic driving platooning in scenarios with impaired communication. The technical solution of this invention is as follows:

[0012] A predictive control method for automated driving formation of smart agricultural machinery in scenarios with impaired communication includes the following steps:

[0013] Step S1: Using a single smart agricultural machinery mass model, establish kinematic equations and further expand them to the entire smart agricultural machinery formation;

[0014] Step S2: Establish a discrete model of smart agricultural machinery under external disturbances including wind disturbance and mechanical friction based on the kinematic equations, and set up a buffer to store data lag caused by communication impairment; the lag data stored in the buffer is used to obtain the delay control sequence.

[0015] Step S3: Adopt the leader-follower formation control structure mode, establish the safety performance constraints of smart agricultural machinery under the conditions of communication impairment and external disturbance, including acceleration constraints, speed constraints, and safety distance constraints; consider the safety distance between two adjacent agricultural machines in the actual working process of smart agricultural machinery formation, and set safety performance constraints.

[0016] Step S4: Construct the optimal control problem, considering the actual operating trajectory and safety performance limitations of the smart agricultural machinery, introduce the maximum-minimum model predictive control strategy, solve for the online controller that meets the conditions, and obtain the optimal control input.

[0017] Furthermore, in step S1, assuming the intelligent agricultural machinery convoy is traveling in a straight line at a fixed interval, it is relatively convenient to design the controller using the relative kinematics of two consecutive adjacent machines. Therefore, the spatial kinematic equation of the j-th intelligent agricultural machine is:

[0018]

[0019] in x j (t)=[Δp j (t) Δv j (t) a j (t)] T This is the jth smart agricultural machinery information system.

[0020] x j-1 (t)=[Δp j-1 (t) Δv j-1 (t) a j-1 (t)] T Information for the (j-1)th smart agricultural machine at the front.

[0021] p represents the positional error between two adjacent agricultural machines. j-1 (t), p j (t) represents the position of the preceding machine and the current position of the agricultural machine, respectively, and A, B, and G represent appropriate matrices. d s These represent positional error and fixed spacing, respectively. Representing the relative speed between two adjacent agricultural machines, and considering the time constant ζ of the electric motor, we have: Configure controller u j (t) represents the desired acceleration a. des (t). a j (t) represents the derivative of acceleration and acceleration.

[0022] Furthermore, in step S1, centralized control of the intelligent agricultural machinery platoon is adopted, defining the state of the entire intelligent agricultural machinery platoon. Assuming that M intelligent agricultural machines are traveling in a straight line, then x(t) = col(x j (t),j∈M),u(t)=col(u j (t),j∈M), where col, x(t), and u(t) represent vector, state information, and the controller to be solved, respectively. The entire intelligent agricultural machinery platooning system is then...

[0023]

[0024] in, Represent an appropriate matrix, x 0(t) represents the navigator's attitude information. Here, all following agricultural machines use the adjacent smart agricultural machine in front as their tracking target, with the first smart agricultural machine serving as the navigator.

[0025] Furthermore, in step S2, in order to stabilize the dynamic error, the state equation system (2) is discretized using the forward Euler method, where Tr is the sampling time. Considering external disturbances, a more realistic dynamic model of the intelligent agricultural machinery is obtained.

[0026]

[0027] in

[0028] For a suitable matrix, T represents the appropriate matrix after sampling. r Indicates the sampling time, w k Due to wind disturbance and mechanical friction, and also Represents the set of disturbances.

[0029] Considering communication impairment, we introduce modeling by storing hysteresis data in a buffer, and then restate the above under impairment conditions as follows:

[0030]

[0031] in Solving for controller u k-2τ I M Represents the identity matrix, u k-τ This represents a delay controller, where τ is the delay and is related to the adoption time T. r They exhibit a certain linear relationship. An extended sequence of controller delays has been added. These represent the state information at time k+1 and time k, respectively.

[0032] Furthermore, in step S3, constraints are applied to acceleration, velocity, and safety distance to construct state constraint inequalities.

[0033]

[0034] E u f u , Represents the set of safety performance constraints, according to a min ≤a j (t)≤a max , a min =u min =u j (ttd )≤u max =a max It can be seen that requirements are made for position error, velocity error, and acceleration respectively.

[0035] ◇1 and ◇2 represent appropriate matrices, a max a min , These represent the maximum acceleration, minimum acceleration, maximum position error, minimum position error, maximum velocity error, and minimum velocity error, respectively. x , The safety performance constraints of intelligent agricultural machinery are described by the above polyhedral set, i.e., the linear matrix inequality, by representing an appropriate matrix; and satisfy its safety performance constraints.

[0036] Furthermore, the specific steps of step S4 are as follows:

[0037] S401. First, define the cost function using a max-min model predictive control strategy.

[0038]

[0039] Q, R, and P represent appropriate weight matrices. u k-τ+1|k-τ , These represent the state information and delay controller at different times, respectively.

[0040] At the same time, under this cost function condition, relevant safety performance constraints are added.

[0041]

[0042] The above constraints represent the constraints on controller u k+i|k Acceleration, velocity set x k+i|k And the constraints on the perturbation variable w, The stability that the model predictive control must satisfy at the end of the prediction time range is defined;

[0043] S402. To facilitate better analysis of the problem, the optimization problem (6) is expanded, and stacking variables are defined.

[0044]

[0045] At the same time, it also received The new equation of state is

[0046]

[0047] in Represent an appropriate matrix, and It is separate and The Kronecker product;

[0048] Therefore, optimization problem 6 can be rewritten as follows:

[0049]

[0050] In the formula The weight matrix is ​​represented by the following form, where the security constraints are rewritten as follows:

[0051]

[0052] S403. Find a definite value to replace the max-min problem, thus effectively generating the problem to be solved, therefore we have By effectively solving for the minimum value t, it becomes possible to resolve uncertain bounded external disturbances.

[0053] S404. According to Schur's complement theorem, Transform into an uncertain LMI with respect to U and t, where X and U, W are linearly parameterized;

[0054] S405. Definition of Insertion X Separate deterministic terms and uncertainty terms

[0055] S406. Considering the worst-case scenario of its uncertain external disturbance term, obtain the set of maximum values ​​at the current time. Establish the LMI linear matrix inequality for affine uncertainty using the maximum value principle;

[0056] S407. Based on the robust linear matrix inequality theorem for affine uncertainty, the uncertain LMI is transformed into an LMI containing only deterministic terms. State constraints and control constraints are added, and the optimal control input is determined by solving the problem using tools YALMIP and MOSEK.

[0057] S408. Define an invariant set of terminal constraints. If a terminal state constraint can be satisfied at time k, then it can also be satisfied at time k+1. Given a set of terminal constraints in a set of constant terms, use Schur's complement theorem to transform it into a bivariate product problem.

[0058] S409. Introduce a positive real scalar to eliminate uncertainty interference terms for the state constraint X∈X. N , and control constraints U∈U N , By performing congruential transformations to convert the problem into matrix inequalities, and given the value of a single variable in the product of two variables, the maximum volume invariant ellipsoid is solved using the maximum value theorem for linear functions in an ellipsoid, thus ensuring the stability of the system state.

[0059] A smart agricultural machine includes: a data link established between a follower smart agricultural machine and an adjacent preceding agricultural machine via a wireless communication module, enabling the follower smart agricultural machine to receive real-time position and speed information from the preceding machine and adjust its driving control according to an updated control law to ensure stable and safe platooning; and a storage unit, wherein a processor calls the executable program code stored in the storage unit to execute the predictive control method for autonomous driving platooning of smart agricultural machines in communication-impaired scenarios as described in any one of the claims.

[0060] A non-transitory computer-readable storage medium storing computer instructions, which, when invoked, are used to execute a predictive control method for intelligent agricultural machinery autonomous driving formation in scenarios with impaired communication, as described in any one of the claims.

[0061] The advantages and beneficial effects of this invention are as follows:

[0062] This invention proposes a predictive control method for autonomous driving platooning of intelligent agricultural machinery in scenarios with impaired communication. It considers external disturbances caused by wind and mechanical friction, thereby improving the computational method for predictive control of autonomous driving platooning of intelligent agricultural machinery. By accurately modeling communication impairment and external disturbances and employing an advanced max-min model predictive control strategy, this method can predict the state of intelligent agricultural machinery in real time and dynamically adjust the control input. This method significantly improves the stability of intelligent agricultural machinery platooning in the face of delays and disturbances, ensuring that the platoon can maintain a predetermined distance and formation, thus greatly reducing system instability caused by delays and external interference, which is the biggest innovation of this invention. An online iterative model predictive controller was developed, using YALMIP and MOSEK to solve the mini-max optimization problem, ensuring the safety performance constraints of the intelligent agricultural machinery platoon, greatly improving the efficiency of driving operations, and ensuring that agricultural machinery can efficiently and accurately complete agricultural production tasks such as tilling, sowing, fertilizing, and harvesting in an unmanned state. This is also the biggest innovation of this invention. Finally, to verify the practical feasibility of the proposed method, simulation experiments were first conducted in Matlab. After confirming the effectiveness of the method, it was imported into the Rflysim experimental platform, and a physical verification was performed using a formation of five agricultural machines. In summary, the intelligent agricultural machinery formation control method of this invention, by considering communication impairment, solves the communication delays and disturbances caused by real-world factors, significantly improving the system's stability, adaptability, and task execution efficiency. This will provide important support for the application and development of intelligent agricultural machinery formation technology in various fields. Attached Figure Description

[0063] Figure 1 This is a flowchart of a preferred embodiment of the intelligent agricultural machinery autonomous driving formation prediction control method under communication impairment provided by the present invention.

[0064] Figure 2 This is a spatial state model diagram of a smart agricultural machinery platoon.

[0065] Figure 3 Flowchart for storing lag data in a buffer.

[0066] Figure 4 A diagram showing the specific safety performance constraints for intelligent agricultural machinery platoons.

[0067] Figure 5 The flowchart for calculating the optimal controller is shown.

[0068] Figure 6 This is a schematic diagram showing the acceleration changes of a follower agricultural machine under different delay conditions.

[0069] Figure 7 This is a schematic diagram showing the speed changes of a follower agricultural machine under different delay conditions.

[0070] Figure 8 This is a schematic diagram showing the change in positional error between two adjacent agricultural machines under different delay conditions. Detailed Implementation

[0071] The technical solutions of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0072] The technical solution of the present invention to solve the above-mentioned technical problems is:

[0073] A predictive control method for autonomous driving platooning of smart agricultural machinery in scenarios with impaired communication is proposed. This method considers a single smart agricultural machine in a laboratory setting as a point mass model, describing it using a physical model in a Cartesian coordinate system to construct a kinematic model. This model is then expanded to a platoon of five smart agricultural machines. An error-based system model is established based on the kinematic equations, introducing communication delay modeling. A buffer is set to obtain a delay control sequence, and a max-min model predictive control strategy is employed. Uncertainties are separated, and the system optimization problem is solved using YALMIP and MOSEK methods. The specific steps include:

[0074] Step a: Set up a formation of five smart agricultural machines to travel in a straight line at a constant speed with a fixed distance. Use the actual position coordinates of the smart agricultural machines and the coordinates of the adjacent smart agricultural machines in front as the system error. Use the speed difference, position difference and acceleration as the system state variables. Establish the state equation of a single smart agricultural machine in a plane rectangular coordinate system.

[0075] Step b: Construct a set of state equations for the five smart agricultural machines to obtain the state equations for the entire smart agricultural machine formation system;

[0076] Step c: Establish a discrete model of smart agricultural machinery under external interference, introduce modeling of communication impairment, and store data delays caused by communication impairment through a buffer;

[0077] Step d: Establish acceleration constraints, speed constraints, and safe distance constraints between two adjacent agricultural machines under conditions of communication impairment and external disturbance;

[0078] Step e: Introduce the maximization-minimum model predictive control strategy. By separating the uncertain terms in the form of perturbation maximization, the problem is transformed into an optimization problem with only deterministic LMI terms. At the same time, safety constraints are added as the objective conditions to be achieved. The solution is obtained using YALMIP and MOSEK.

[0079] Step f: Define the terminal constraint invariant set such that the terminal constraint set is within a constant term 1. Use Schur's complement theorem to transform it into a bivariate product problem. Solve for the maximum volume invariant ellipsoid using the maximum value theorem for linear functions in an ellipsoid.

[0080] Step g: Experimental simulation and physical verification. This invention realizes the application of predictive control in the field of intelligent agricultural machinery autonomous driving formation in communication-impaired scenarios, improving practical feasibility.

[0081] In step a, according to Figure 2 Given the spatial state model diagram, the spatial kinematic equations of the j-th intelligent agricultural machine are as follows:

[0082]

[0083] in x j (t)=[Δp j (t)Δv j (t)a j (t)] T This is the jth smart agricultural machinery information system.

[0084] x j-1 (t)=[Δp j-1 (t)Δv j-1 (t)a j-1 (t)] T Information for the (j-1)th smart agricultural machine at the front.

[0085] This indicates the positional error between two adjacent agricultural machines. The relative speed between two adjacent agricultural machines is given, while also considering the time constant ζ of the electric motor. controller uj (t) represents the desired acceleration a. des (t).

[0086] In step b, the state equation of the entire intelligent agricultural machinery platoon is defined, and the five intelligent agricultural machinery are set to travel in a straight line in an unmanned platoon. Then we have x(t) = col(x j (t),j∈M),u(t)=col(u j (t),j∈5), the equation of the entire intelligent agricultural machinery formation system is:

[0087]

[0088] in, x 0 (t) = 0, where all following smart agricultural machines take the adjacent smart agricultural machine in front as the tracking target, and the first smart agricultural machine is the navigator.

[0089] In step c, to stabilize the dynamic error, the state equation system (2) is discretized using the forward Euler method, and the sampling time Tr is set to 0.05 s. The set of external disturbances in the actual environment, including wind disturbance and mechanical friction, is as follows: A more realistic drone dynamic model is as follows:

[0090]

[0091] in For an appropriate matrix. w k It is a collection of disturbances caused by wind and mechanical friction, and has

[0092] Store delayed data, such as in a process, using a buffer. Figure 3 As shown, the delay time is within an interval. After discretization, the sampling time T is set. r 0.05s, three different delay times are determined, and the previous control sequence can be restated under the delay conditions.

[0093]

[0094]

[0095] Solving for controller u k-2τ And output it.

[0096] In step d, according to Figure 4 Specific safety performance constraints for intelligent agricultural machinery platooning are clearly defined, including constraints on acceleration. min ≤a j (t)≤a max Constraining speed error Constraining position error Position error, velocity error, and acceleration are integrated into a state vector for the controller u. k-τ (Ideal acceleration) is constrained individually, and the following constraint inequality is constructed:

[0097]

[0098] in The safety performance constraints of intelligent agricultural machinery are described by the aforementioned polyhedral set, i.e., linear matrix inequalities. The proposed method subsequently addresses the issues of external disturbances and communication delays while satisfying these safety performance constraints.

[0099] In step e, refer to Figure 5 The controller algorithm calculation flowchart is as follows, and the specific implementation method is as follows:

[0100] A maximum-minimum model predictive control strategy is introduced, defining a cost function that considers external disturbances as maximum values, resulting in a controller with the minimum value obtained.

[0101]

[0102] Simultaneously, under this cost function condition, relevant safety performance constraints are added, including velocity error constraints, acceleration constraints, and position error constraints.

[0103]

[0104] The above constraints represent the constraints on controller u k+i|k The set of acceleration, velocity error, and position error x k+i|k And the constraints on the perturbation variable w, will It is defined as the stability that the model predictive control must satisfy at the end of the prediction time range.

[0105] The optimization problem (6) is expanded by defining stacking variables.

[0106]

[0107] At the same time, it also received The new equation of state is

[0108]

[0109] in and It is separate and The Kronecker product.

[0110]

[0111] Optimization problem 6 can be rewritten as follows:

[0112]

[0113] In the formula The safety performance constraints are rewritten in the following form

[0114]

[0115] Introducing the constant term t, we obtain the inequality By effectively solving for the minimum value t, the problem of uncertain bounded external disturbances can be resolved.

[0116] According to Schur's complement theorem, Transformed into an indeterminate LMI with respect to U and t.

[0117] Insert the definition of X Separate deterministic terms and uncertainty terms

[0118] Apart from itself, all other external disturbances are always uncertain and fluctuate, but they always fall within a definite set. Therefore, by setting the external disturbances to the worst-case scenario, we obtain the set of maximum values ​​at the current moment. We use the maximum principle to establish a robust linear matrix inequality for affine uncertainty to overcome the robustness of uncertain LMIs.

[0119] The robust linear matrix inequality is transformed into an LMI without any uncertain terms using LMI, and the state constraint X∈X is added. N , and control constraints U∈U N , The state constraints also contain uncertain external disturbances, which are then addressed by separating the deterministic terms. and uncertainty terms The solution involves using tools like YALMIP and MOSEK to determine the optimal control input.

[0120] In step f, a terminal constraint invariant set is defined. If the terminal state constraint can be satisfied at time k, then it can also be satisfied at time k+1. Given a terminal constraint set within a set of constant terms, the Schur complement theorem is used to transform it into a bivariate product problem. A positive real scalar is introduced, and the uncertainty interference term is eliminated through the LMI of affine uncertainty. The state constraint X∈X is then... N , and control constraints U∈U N , By performing congruential transformations to convert the problem into matrix inequalities, and given the value of a single variable in the product of two variables, the maximum volume invariant ellipsoid is solved using the maximum value theorem for linear functions in an ellipsoid, thus ensuring the stability of the system state.

[0121] In step g, based on the already determined maximum acceleration a of the smart agricultural machinery formation... max Minimum acceleration, a min Maximum speed error Minimum speed error Maximum position error Minimum position error And the given workshop spacing d s This study considers the operation and movement of intelligent agricultural machinery platoons under three different latency conditions. Software simulations using Coptersim and Matlab are conducted to ensure the proposed control method is effective in addressing communication impairments and external interference. Simultaneously, the QGC ground station records the simulated driving data of five intelligent agricultural machines, ensuring that the actual position, speed, and acceleration of each machine are within the required range, demonstrating a significant improvement. Next, the system connects to the Rflysim experimental platform, and a control module is built in Simulink. Commands are transmitted to the intelligent agricultural machines via Mavlink and UDP protocols. Experiments demonstrating the platoon's ability to move in a straight line at fixed intervals prove that the intelligent agricultural machinery platoon performs well under communication impairments and external interference, while also meeting safety requirements. Figures 6-8 The diagrams show the changes in acceleration, speed, and positional error between adjacent agricultural machines under three delay conditions: 0.05s, 0.1s, and 0.15s. The diagrams show that each following agricultural machine can follow the machine in front of it well, meeting the previously given safety performance constraints. Under traditional conditions, the distance between adjacent agricultural machines is greatly shortened, which greatly improves the working efficiency of the intelligent agricultural machine unmanned driving formation.

[0122] A predictive control method for autonomous driving platooning of smart agricultural machinery in scenarios with impaired communication is proposed. Based on the actual working conditions of smart agricultural machinery platooning, this method considers the impact of communication impairment over time and uncertain external factors. By introducing communication delay modeling, a maximum-minimum model predictive control strategy is adopted to optimize the driving performance of the smart agricultural machinery platooning and improve energy efficiency.

[0123] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0124] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0125] The above embodiments should be understood as illustrative only and not as limiting the scope of protection of the present invention. After reading the description of the present invention, those skilled in the art can make various alterations or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A method for intelligent agricultural vehicle automatic driving platoon prediction control in a communication impaired scenario, characterized in that, The method comprises the following steps: Step S1: a single intelligent agricultural machine particle model is cited to establish a kinematics equation, which is further expanded to the entire intelligent agricultural machine formation; Step S2: an intelligent agricultural machine discrete model under external disturbances including wind disturbance and mechanical friction is established according to the kinematics equation, and a buffer is set to store data delay caused by impaired communication; wherein the delay data stored in the buffer obtain a delay control sequence; Step S3: a leader-follower formation control structure mode is adopted to establish safety performance constraint requirements of the intelligent agricultural machine under the conditions of impaired communication and external disturbance, including acceleration constraint, speed constraint and safety distance constraint; the safety distance between adjacent two agricultural machines in the actual working process of the intelligent agricultural machine formation is considered, and safety performance constraint is set; Step S4: an optimal control problem is constructed, the actual running track and safety performance limit of the intelligent agricultural machine are considered, a maximum-minimum model predictive control strategy is introduced, an online controller satisfying the conditions is solved, and optimal control input is obtained; in the step S2, in order to make the dynamic error tend to be stable, the state equation system (2) is discretized by forward Euler method, Tr is the sampling time, external disturbance is considered, and a more real intelligent agricultural machine dynamic model is obtained （3） where , , , , , is an appropriate matrix, , , , denotes the sampled appropriate matrix, denotes the sampling time, is the wind disturbance and mechanical friction, and has ; denotes the set of disturbances; Considering the communication impairment, the dynamic model of intelligent agricultural machinery is re-expressed as （4） wherein , , , , solving the controller , denotes the identity matrix, denotes the delay controller, is the delay and takes the time presents a certain linear relationship, The extended sequence with the controller delay is added in the middle; , respectively represent the state information at k+1 and k time. The specific steps of the step S4 are: S401. First, a cost function is defined by the maximum-minimum model predictive control strategy （6） , , denote an appropriate weight matrix, , , , denote state information and a delay controller at different times, respectively; At the same time, under the condition of the cost function, relevant safety performance constraint conditions are added The above constraints represent constraints on the controller , the acceleration, velocity set , and the disturbance variable , define the stability that the model predictive control needs to satisfy at the end of the prediction time horizon. S402. In order to better analyze the problem, the optimization problem (6) is expanded, and a stacking variable is defined Also obtained are , and a new equation of state is (7) where , denotes an appropriate matrix, and is the Kronecker product of , respectively. Therefore, the optimization problem (6) can be rewritten as （8） wherein , , , denotes the weight matrix, wherein the safety performance constraint is rewritten in the form of S403. Find a deterministic value to replace the max-min problem, so as to effectively generate the problem to be solved, thus , by effectively solving the minimum value t, it is possible to solve the uncertain bounded external disturbance; S404. According to the Schur complement theorem, the inequality is converted into an uncertain LMI with respect to U and t, where, and , is linearly parameterized; S405. Insert Definitions Separate deterministic and non-deterministic terms ;​ S406. Considering the worst case of its uncertain external disturbance term, the maximum value set at the current time is obtained The LMI linear matrix inequality of affine uncertainty is established by using the maximum principle. S407. According to the affine uncertainty robust linear matrix inequality theorem, the LMI of the uncertainty is converted into an LMI containing only determined items, state constraints and control constraints are added, and the optimal control input is determined by solving through the tools YALMIP and MOSEK; S408. A terminal constraint invariant set is defined, if the terminal state constraint can be met at k time, it can also be met at k+1 time, a terminal constraint set is given in a constant item set, and the Schur complement theorem is used to convert it into a two-variable product problem; S409. Introduce a positive real scalar to eliminate the uncertain disturbance term, and state constraints and control constraints Perform congruence transformation, convert to matrix inequality, given the value of a single variable in the product of two variables, solve the maximum volume invariant ellipsoid according to the maximum value theorem of linear function in ellipsoid, and ensure the stability of the system state. 2.The method of claim 1, wherein, In the step S1, it is assumed that the intelligent agricultural machine formation is in a fixed-pitch straight-line driving, and it is convenient to design the controller by using the relative kinematics of the two adjacent agricultural machines; thus, the space kinematics equation of the jth intelligent agricultural machine is （1） wherein , , , is the jth intelligent agricultural machinery information, is the (j-1)th intelligent agricultural machinery information in front, denotes the position error between the adjacent two agricultural machines, , respectively denote the front machine position and the current agricultural machinery position information, A, B, G denote appropriate matrices, respectively denote the position error and the fixed interval, denotes the relative speed between the adjacent two agricultural machines, considering the motor time constant , , the controller is set is the desired acceleration , , respectively denote the acceleration derivative and the acceleration.

3. The intelligent agricultural vehicle automatic driving platoon prediction control method for a communication impaired scenario according to claim 1, wherein, In the step S1, the wisdom agricultural machine formation centralized control is adopted, the state of the whole wisdom agricultural machine formation is defined, it is assumed that M wisdom agricultural machines are in straight line driving, then , 、 、 respectively represent a vector, state information, a controller to be solved, and the whole wisdom agricultural machine formation system is （2） wherein, , , , , , denote appropriate matrices, is the leader pose information, here all the follower agricultural machines take the adjacent front intelligent agricultural machine as the tracking target, and the first intelligent agricultural machine is the leader.

4. The intelligent agricultural vehicle automatic driving platoon prediction control method for a communication impaired scenario according to claim 1, wherein, In the step S3, the acceleration, speed and safety distance are respectively constrained, and the state constraint inequality is constructed （5） 、 、 、 、 、 denotes a set of safety performance constraints, according to , , , It can be seen that the requirements for position error, velocity error, and acceleration are made respectively, wherein , , , , , , , , , denotes a suitable matrix, , , , , , denotes a set of maximum acceleration, minimum acceleration, maximum position error, minimum position error, maximum velocity error, minimum velocity error, , denotes a suitable matrix, and the safety performance constraints of the smart agricultural machine are described by the above polyhedral set, that is, a linear matrix inequality; and the smart agricultural machine satisfies the safety performance constraints.

5. A smart agricultural machine, characterized in that, It comprises: The follower intelligent agricultural machine and the adjacent front agricultural machine are linked through a wireless communication module, so that the follower intelligent agricultural machine can receive the position and speed information of the front in real time, and adjust the driving control according to the updated control law, so that the formation is stable and safe, and there is a storage, and the processor calls the executable program code stored in the storage to execute the intelligent agricultural machine automatic driving formation predictive control method for the impaired communication scene as claimed in any one of claims 1 to 4.

6. A non-transitory computer-readable storage medium, comprising: The computer readable storage medium stores computer instructions, which are invoked to perform the intelligent agricultural machine automatic driving platoon prediction control method for a communication impaired scenario according to any one of claims 1-4.

Images