Intelligent agricultural machine automatic driving formation prediction control method for communication damage scene

Abstract

The invention requests to protect an intelligent agricultural machine automatic driving formation prediction control method for a communication damage scene. Comprising the steps of S1, introducing a single smart agricultural machinery mass point model, establishing a kinematic equation thereof, and further expanding the kinematic equation to a whole smart agricultural machinery formation; s2, establishing a smart agricultural machine discrete model according to the kinematics equation, and setting a buffer to store data hysteresis caused by communication damage; s3, a front machine-follower formation control structure mode is adopted, and intelligent agricultural machine safety performance constraint requirements under the conditions of communication damage and external disturbance are established; and S4, constructing an optimal control problem, introducing a maximum-minimum model prediction control strategy by considering the actual running track and safety performance limitation of the intelligent agricultural machine, solving an online optimal controller meeting conditions, and obtaining optimal control input. According to the method, the optimal controller under the condition of communication damage is obtained by considering online solution under the maximum disturbance, and high-efficiency intelligent agricultural machinery formation collective work is realized.

Description

Technical Field

[0001] The present invention relates to the field of agricultural automation unmanned driving technology, and in particular to the research and implementation of a prediction control method for a self-driving formation of intelligent agricultural machinery in a communication-impaired scenario. Background Art

[0002] With the rapid development of global agricultural technology, traditional agriculture is gradually moving towards modernization and intelligence. Based on the development of agricultural mechanization, smart agriculture and automation technology are becoming an important trend in global agricultural development, greatly improving agricultural production efficiency and resource utilization, and reducing the labor intensity of farmers. Against this background, the technology of autonomous driving formation of smart agricultural machinery has gradually become a hot research direction in the field of agricultural machinery. The technology of autonomous driving formation of smart agricultural machinery refers to the use of high-precision navigation, real-time communication, path planning and collaborative control technologies to enable multiple agricultural machinery to achieve synchronous and collaborative automated operations in the field in the form of formation. This technology is based on the global navigation satellite system (GNSS), sensors, artificial intelligence (AI) and automatic control algorithms to ensure that agricultural machinery can efficiently and accurately complete agricultural production tasks such as ploughing, sowing, fertilizing and harvesting in an unmanned state.

[0003] However, although the intelligent agricultural machinery formation technology has shown great potential in many fields, its development still faces many challenges, especially in the formation control strategy. Formation control involves the dynamic collaboration of multiple agricultural machinery, and many complex factors need to be considered, including: Dynamics and kinematic models: Each intelligent agricultural machinery has its own unique dynamic characteristics, such as speed, acceleration, etc. Under non-ideal conditions (such as wind speed and airflow), changes in these characteristics will affect the stability of the formation and the efficiency of task execution. Communication delay and network communication, the effect of intelligent agricultural machinery often depends on the real-time information exchange between each intelligent agricultural machinery. However, in practical applications, the communication between agricultural machinery is often affected by communication impairment due to factors such as delay, interference and data loss. Therefore, how to effectively carry out collaborative control under the consideration of communication impairment has become an important issue. The core of the intelligent agricultural machinery formation is the collaborative algorithm of multiple intelligent agricultural machinery. At present, a variety of control algorithms have been applied to the intelligent agricultural machinery formation, such as the leader-follower model and virtual piloting. Although these algorithms have been successful under certain conditions, there is still a risk of not being fully addressed in complex dynamic environments.

[0004] At present, many studies focus on model-based control methods, such as model predictive control, which is believed to be able to effectively carry out multi-intelligent agricultural machinery collaboration in a dynamic environment. In practical applications, by establishing a system dynamic model, predicting the future state in real time and optimizing the control, the predictive characteristics of the system can be used to effectively deal with delays and emergencies. However, the existing model predictive control methods often do not fully consider the impact of communication delay on control performance, coupled with many external factors such as wind disturbance and uncertain external interference caused by mechanical friction, which may lead to a decrease in the stability and reliability of the system in a real environment. In view of the above technical background and challenges, the present invention proposes a predictive control method for the automatic driving formation of intelligent agricultural machinery in a communication-impaired scenario, aiming to overcome the negative effects of communication impairment and uncertain external interference, so as to improve the collaboration efficiency and stability of the intelligent agricultural machinery formation. By introducing the modeling of communication delay and the design based on the maximum-minimum model predictive control, we plan to realize an efficient and reliable intelligent agricultural machinery control system, and provide a new solution for the promotion and application of intelligent agricultural machinery technology.

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

[0006] After searching, the application publication number is CN116466722A, a human-machine integrated intelligent agricultural machinery automatic driving formation transfer system and method, which is installed on a pilot host and several follower slaves, the pilot host and each follower slave are equipped with the same driving status collection system, the pilot host is also equipped with a formation control terminal system, a router master station and a host WiFi module connected in sequence, each follower slave is equipped with a slave WiFi module that wirelessly communicates with the host WiFi module, all collected information is transmitted to the formation control terminal system via the router master station, the on-board computer in the formation control terminal system obtains the machine adjustment decision of the follower slave, the pilot host operator transmits the human decision to the on-board computer through the human-machine interaction interface, and the on-board computer integrates the human decision with the machine adjustment decision to generate the formation decision and the control instruction of the control quantity input;

[0007] In the design and implementation of a human-machine integrated intelligent agricultural machinery automatic driving formation transfer system and method, although there are many advantages, there are still some shortcomings, mainly including the following aspects:

[0008] 1. Stability of communication and data transmission. In actual application scenarios, the farmland environment is complex, and signal interference and multipath effects may cause unstable communication and affect the real-time data transmission. The human-machine fusion system may face delays in the data collection and feedback process, which affects the real-time control and decision-making of the system.

[0009] 2. Insufficient system robustness. Due to its sensitivity to external interference, the system's autonomous driving may be affected under certain extreme conditions (such as bad weather and complex terrain), resulting in reduced efficiency in field transfers. There may be problems with the compatibility and coordination between different types of agricultural machinery, sensors, and controllers, affecting the overall performance of the system. 3. Safety and feasibility Autonomous driving technology may fail in some cases, especially during field transfers, which may lead to equipment collisions or safety issues in the work area.

[0010] Therefore, the problems of communication impairment and interference can be considered in the modeling stage, and safety performance constraints can be considered to ensure safety. At the same time, the design of the dynamic model and control algorithm of the system robustness can be improved. In combination with environmental monitoring data, a more robust control algorithm can be designed to adapt to different working conditions and improve the reliability of the system. The addition of fusion processing of multiple sensors can enhance environmental perception, reduce dependence on a single sensor, and increase the adaptability of the system. Summary of the invention

[0011] The present invention aims to solve the above problems of the prior art. A prediction control method for the automatic driving formation of smart agricultural machinery in a communication-impaired scenario is proposed. The technical solution of the present invention is as follows:

[0012] A prediction control method for an intelligent agricultural machinery automatic driving formation in a communication-impaired scenario includes the following steps:

[0013] Step S1: refer to a single smart agricultural machinery particle model, establish kinematic equations, and further expand it to the entire smart agricultural machinery fleet;

[0014] Step S2: establishing a discrete model of the smart agricultural machinery under external disturbances including wind disturbance and mechanical friction according to the kinematic equation, and setting a buffer to store data hysteresis caused by communication damage; wherein the hysteresis data stored in the buffer is used to obtain a delay control sequence;

[0015] Step S3: adopt the leading machine-follower formation control structure mode to establish the safety performance constraint requirements of smart agricultural machinery under the conditions of communication damage and external disturbance, including acceleration constraint, speed constraint, and safety distance constraint; consider the safety distance problem between two adjacent agricultural machinery in the actual working process of the smart agricultural machinery formation, and set the safety performance constraint;

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

[0017] Furthermore, in step S1, assuming that the smart agricultural machinery formation is traveling in a straight line with a fixed spacing, it is relatively convenient to design the controller using the relative kinematics of two consecutive adjacent agricultural machinery. Therefore, the spatial kinematic equation of the jth smart agricultural machinery is:

[0018]

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

[0020] x j-1 (t) = [Δp j-1 (t) Δv j-1 (t) a j-1 (t)] T Information about the j-1st smart agricultural machine ahead.

[0021] represents the position error between two adjacent agricultural machines, p j-1 (t), p j (t) represents the position of the preceding machine and the current agricultural machine position information, A, B, G represent appropriate matrices, d s represent the position error and fixed spacing respectively, represents the relative speed between two adjacent agricultural machines, and considering the motor time constant ζ, we have Setting up the controller j (t) is the expected acceleration a des (t). a j (t) represents the acceleration derivative and acceleration.

[0022] Furthermore, in step S1, the state of the entire smart agricultural machinery formation is defined by using the centralized control of the smart agricultural machinery formation. Assuming that there are M smart agricultural machinery traveling in a straight line, then x(t)=col(x j (t),j∈M),u(t)=col(u j (t), j∈M), col, x(t), and u(t) represent vector, state information, and controller to be solved respectively. The entire smart agricultural machinery formation system is

[0023]

[0024] in, represents the appropriate matrix, x 0(t) is the leader’s attitude information. Here, all following agricultural machines take the adjacent front smart agricultural machine as the tracking target, and the first smart agricultural machine is the leader.

[0025] Furthermore, in step S2, in order to make the dynamic error tend to be stable, the state equation system (2) is discretized by the forward Euler method, Tr is the sampling time, and external interference is considered to obtain a more realistic dynamic model of smart agricultural machinery.

[0026]

[0027] in

[0028] is a suitable matrix, represents the appropriate matrix after sampling, T r represents the sampling time, w k For wind disturbance and mechanical friction, and represents a collection of perturbations.

[0029] Considering communication impairment, the buffer is used to store the hysteresis data, and the modeling of communication impairment is introduced. Under the condition of impairment, the above is restated as

[0030]

[0031] in Solve the controller u k-2τ , I M represents the identity matrix, u k-τ represents a delay controller, τ is the delay and is related to the time T r There is a certain linear relationship. An extended sequence of controller delays was added. Respectively represent the state information at time k+1 and time k.

[0032] Furthermore, in step S3, the acceleration, speed, and safety distance are constrained respectively to construct the state constraint inequality:

[0033]

[0034] E u 、f u , Represents a 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 the requirements for position error, velocity error, and acceleration are respectively:

[0035] ◇1, ◇2 represent appropriate matrices, a max 、a min , Respectively represent the maximum acceleration, minimum acceleration, maximum position error, minimum position error, maximum speed error, and minimum speed error, f x , Represents an appropriate matrix, and the safety performance constraints of smart agricultural machinery are described by the above polyhedron set, that is, the linear matrix inequality; satisfying its safety performance constraints.

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

[0037] S401. First, define the cost function through the maximum-minimum model predictive control strategy

[0038]

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

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

[0041]

[0042] The above constraints represent the controller u k+i|k , acceleration, velocity set x k+i|k , and the constraints on the disturbance variable w, Defines the stability that model predictive control needs to meet at the end of the forecast time horizon;

[0043] S402. In order to better analyze the problem, the optimization problem (6) is expanded to define stacking variables

[0044]

[0045] Also got The new state equation is

[0046]

[0047] in represents the appropriate matrix, and are respectively The Kronecker product of

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

[0049]

[0050] In the formula represents the weight matrix, where the safety performance constraint is rewritten as follows

[0051]

[0052] S403. Find a definite value to replace the maximum-minimum problem, so as to effectively generate the problem to be solved, so there is By effectively solving for the minimum value t, it becomes possible to solve uncertain bounded external disturbances;

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

[0054] S405. Insert the definition of X Separate deterministic terms and uncertainty

[0055] S406. Consider the worst case of uncertain external interference terms and obtain the maximum value set at the current moment The LMI linear matrix inequality with affine uncertainty is established using the maximum principle;

[0056] S407. According to the affine uncertainty robust linear matrix inequality theorem, the uncertain LMI is transformed into an LMI containing only certain terms, state constraints and control constraints are added, and the optimal control input is determined by solving the problem using the tools YALMIP and MOSEK;

[0057] S408. Define the terminal constraint invariant set. 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 in a constant term set, use Schur's complement theorem to transform it into a two-variable product problem.

[0058] S409. Introduce a positive real scalar to eliminate the uncertain interference term, and constrain the state X∈X N , and control constraints U∈U N , A congruence transformation is performed and converted into a matrix inequality. Given the value of a single variable in the product of two variables, the maximum volume invariant ellipsoid is solved according to the maximum value theorem of the linear function in the ellipsoid, thus ensuring the stability of the system state.

[0059] A smart agricultural machinery, comprising: establishing a data link between a follower smart agricultural machinery and an adjacent front agricultural machinery through a wireless communication module, so that the follower smart agricultural machinery can receive the position and speed information of the front in real time, and adjust the driving control according to the updated control law to make the formation stable and safe; and a storage, a processor calls the executable program code stored in the storage, and executes a smart agricultural machinery automatic driving formation prediction control method for a communication impaired scenario as described in any one of the items.

[0060] A non-transitory computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are called, they are used to execute a smart agricultural machinery automatic driving formation prediction control method for a communication-impaired scenario as described in any one of the items.

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

[0062] The present invention proposes a predictive control method for the automatic driving formation of smart agricultural machinery for communication-impaired scenarios, taking into account the external disturbance problems caused by wind disturbance and mechanical friction, thereby improving the calculation method of predictive control of the automatic driving formation of smart agricultural machinery. By accurately modeling the communication impairment and external disturbances and adopting the advanced maximum-minimum model predictive control strategy, this method can predict the state of smart agricultural machinery in real time and dynamically adjust the control input. This method significantly improves the stability of the smart agricultural machinery formation in the face of delays and disturbances, ensures that the formation can maintain a predetermined distance and formation, thereby greatly reducing the system instability caused by delays and external interference, which is the biggest innovation of the present invention. An online iterative model predictive controller was developed, and YALMIP and MOSEK were used to solve the "minimum-maximum" optimization problem, ensuring the safety performance constraints of the smart agricultural machinery formation, greatly improving the efficiency of driving work, and ensuring that agricultural machinery can efficiently and accurately complete agricultural production tasks such as ploughing, sowing, fertilizing, and harvesting in an unmanned state. This is also the biggest innovation of the present invention. Finally, in order to verify the practical feasibility of the proposed method, a simulation experiment was first carried out in Matlab. After confirming the effectiveness of the method, it was imported into the Rflysim experimental platform and five agricultural machinery formations were used for physical verification. In summary, the intelligent agricultural machinery formation control method of the present invention solves the communication delay and disturbance caused by realistic factors by considering the communication damage, significantly improves the stability, adaptability and task execution efficiency of the system, and will provide important support for the application and development of intelligent agricultural machinery formation technology in various fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a flow chart of a method for predicting and controlling an automatic driving formation of smart agricultural machinery under impaired communication, which is a preferred embodiment provided by the present invention.

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

[0065] Figure 3 Flowchart for storing hysteresis data for buffers.

[0066] Figure 4 This is a parameter diagram for the specific safety performance constraint requirements of smart agricultural machinery formations.

[0067] Figure 5 Flowchart for optimal control controller calculation.

[0068] Figure 6 Schematic diagram of the acceleration change of the follower agricultural machinery under different delay conditions.

[0069] Figure 7 Schematic diagram of the speed change of the follower agricultural machinery under different delay conditions.

[0070] Figure 8 Schematic diagram of the position error change between two adjacent agricultural machines under different delay conditions. DETAILED DESCRIPTION

[0071] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

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

[0073] A predictive control method for the autonomous driving formation of smart agricultural machinery in the communication-impaired scenario is proposed. The particle model of a single smart agricultural machinery in the laboratory is considered, described by a physical model in a plane rectangular coordinate system, and a kinematic model is constructed. The single smart agricultural machinery model is expanded to a five-unit unmanned driving formation of smart agricultural machinery. An error-based system model is established according to the kinematic equation, and the modeling of communication delay is introduced. A buffer is set to obtain a delay control sequence, and a maximum-minimum model predictive control strategy is adopted to separate uncertainties, and the system optimization problem is solved using YALMIP and MOSEK. The specific steps include:

[0074] Step a: Set up a formation of five smart agricultural machines to drive in a straight line at a fixed distance and at a uniform speed. Take the actual position coordinates of the smart agricultural machine and the coordinates of the adjacent front smart agricultural machine as the system error, take the speed difference, position difference and acceleration as the system state quantity, and establish the state equation of a single smart agricultural machine in the plane rectangular coordinate system;

[0075] Step b: construct a set of state equations of five smart agricultural machinery to obtain the state equation of the entire smart agricultural machinery formation system;

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

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

[0078] Step e: The maximum-minimum model predictive control strategy is introduced. By separating the uncertain terms in the form of disturbance maximization, it is transformed into an optimization problem with LMI containing only certain terms. At the same time, safety constraints are added as the target conditions to be achieved, and the problem is solved using YALMIP and MOSEK.

[0079] Step f: Define the terminal constraint invariant set so that the terminal constraint set is within a constant term 1. Use Schur's complement theorem to transform it into a two-variable product problem, and solve the maximum volume invariant ellipsoid based on the maximum value theorem of linear functions in the ellipsoid.

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

[0081] In step a, according to Figure 2 Given the spatial state model diagram, the spatial kinematic equation of the jth smart agricultural machine is:

[0082]

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

[0084] x j-1 (t) = [Δp j-1 (t)Δv j-1 (t)a j-1 (t)] T Information about the j-1st smart agricultural machine ahead.

[0085] Represents the position error between two adjacent agricultural machines, Represents the relative speed between two adjacent agricultural machines, and considering the electric machine time constant ζ, we have Controllerj (t) is the expected acceleration a des (t).

[0086] In step b, the state equation of the entire smart agricultural machinery formation is defined, and five smart agricultural machinery unmanned driving formations are set to travel in a straight line, then x(t)=col(x j (t),j∈M),u(t)=col(u j (t),j∈5), the whole smart agricultural machinery formation system equation 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 leader.

[0089] In step c, in order to make the dynamic error tend to be stable, the state equation system (2) is discretized by the forward Euler method, and the sampling time Tr is set to 0.05s. In the actual environment, the external disturbance set of wind disturbance and mechanical friction is A more realistic drone dynamic model is as follows

[0090]

[0091] in is a suitable matrix. k is the disturbance collection of wind disturbance and mechanical friction, and has

[0092] Storing hysteresis data in a buffer, such as process Figure 3 As shown, the delay time is within an interval. After discretization, the sampling time T is set. r 0.05s, determine three different delay times, under the delay condition, it can be restated the previous control sequence

[0093]

[0094]

[0095] Solve the controller u k-2τ And output.

[0096] In the step d, according to Figure 4 Clearly give the specific safety performance constraints of the smart agricultural machinery formation and constrain the acceleration min ≤a j (t)≤a max , constrain the speed error Constraining the position error Integrate the position error, velocity error, and acceleration into a state vector and apply it to the controller u k-τ (Ideal acceleration) performs a separate acceleration constraint and constructs the following constraint inequality

[0097]

[0098] in The safety performance constraints of smart agricultural machinery are described by the above polyhedron set, that is, linear matrix inequality. The safety performance constraints are met by solving the problems of external disturbance and communication delay through the subsequent proposed method.

[0099] In the step e, reference Figure 5 The controller algorithm calculation flow chart is as follows:

[0100] The maximum-minimum model predictive control strategy is introduced to define the cost function, the external disturbance is considered as the maximum value, and the controller solved is the minimum value, and we get

[0101]

[0102] At the same time, under this cost function condition, relevant safety performance constraints are added, including speed error constraints, acceleration constraints, and position error constraints.

[0103]

[0104] The above constraints represent the controller u k+i|k , acceleration, velocity error, position error set x k+i|k , and the constraints on the disturbance variable w, will Defined as the stability that the model predictive control needs to meet at the end of the prediction time horizon.

[0105] Expand the optimization problem (6) and define stacking variables

[0106]

[0107] Also got The new state equation is

[0108]

[0109] in and are respectively The Kronecker product of .

[0110]

[0111] Optimization problem 6 can be rewritten as

[0112]

[0113] In the formula The safety performance constraint is rewritten as follows

[0114]

[0115] Introducing the constant term t, we get the inequality Uncertain bounded external disturbances are resolved by efficiently solving for the minimum t.

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

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

[0118] Except for itself, all other external interference items are always uncertain and in a fluctuating state, but they are always within a certain set. Therefore, the external interference items are set to the worst case and the maximum value set at the current moment is obtained. The maximum principle is used to establish an affine uncertain robust linear matrix inequality to overcome the robustness satisfaction of uncertain LMI.

[0119] The robust linear matrix inequality is transformed into an LMI without any uncertain terms through LMI, and the state constraint X∈X is added N , and control constraints U∈U N , The state constraints also contain uncertain external disturbances, and we can separate the deterministic terms again. and uncertainty Solve, use the tools YALMIP and MOSEK to solve and determine the optimal control input.

[0120] In the 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 in a constant term set, the Schur complement theorem is used to transform it into a two-variable product problem. A positive real scalar is introduced. The uncertain interference term is eliminated through the LMI of affine uncertainty. The state constraint X∈X N , and control constraints U∈U N , A congruence transformation is performed and converted into a matrix inequality. Given the value of a single variable in the product of two variables, the maximum volume invariant ellipsoid is solved according to the maximum value theorem of the linear function in the ellipsoid, thus ensuring the stability of the system state.

[0121] In step g, according to the 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 , considering the working and driving conditions of the smart agricultural machinery formation under three delay conditions, and combining Coptersim and Matlab for software simulation to ensure that the proposed control method has obvious effects in solving the problems of communication damage and external interference. At the same time, the QGC ground station records the simulated driving data of five smart agricultural machinery, ensuring that the actual position, speed, and acceleration of each smart agricultural machinery are within the required range, which has a significant improvement. Then connect to the Rflysim experimental platform, build a control module in Simulink, and transmit instructions to the smart agricultural machinery through Mavlink and UDP protocols. The experiment of the fixed-distance straight-line driving of the smart agricultural machinery formation proves that the smart agricultural machinery formation performs well in solving the problems of communication damage and external interference, and meets the safety performance requirements. Figure 6 to Figure 8 The schematic diagrams show the acceleration change, speed change, and position error change of the follower agricultural machinery under the three delay conditions of 0.05s, 0.1s, and 0.15s. It can be seen from the figure that each following agricultural machinery can follow the front machine very well, meeting the safety performance constraint requirements given previously. Under traditional basic conditions, the distance between two adjacent agricultural machinery is greatly shortened, greatly improving the work efficiency of the smart agricultural machinery unmanned driving formation.

[0122] A predictive control method for the automatic driving formation of smart agricultural machinery for scenarios with impaired communication is proposed. Starting from the actual working situation of the smart agricultural machinery formation, the influence of communication impairment and uncertain external factors is considered. By introducing communication delay modeling, the maximum-minimum model predictive control strategy is used to optimize the driving performance of the smart agricultural machinery formation and improve the energy efficiency level.

[0123] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules 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 technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0124] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0125] The above embodiments should be understood to be only used to illustrate the present invention and not to limit the protection scope of the present invention. After reading the contents of the present invention, technicians can make various changes 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 prediction control method for automatic driving formation of smart agricultural machinery in communication-impaired scenarios, characterized in that: The following steps are involved: Step S1: refer to a single smart agricultural machinery particle model, establish kinematic equations, and further expand it to the entire smart agricultural machinery fleet; Step S2: establishing a discrete model of the smart agricultural machinery under external disturbances including wind disturbance and mechanical friction according to the kinematic equation, and setting a buffer to store data hysteresis caused by communication damage; wherein the hysteresis data stored in the buffer is used to obtain a delay control sequence; Step S3: adopt the leading machine-follower formation control structure mode to establish the safety performance constraint requirements of smart agricultural machinery under the conditions of communication damage and external disturbance, including acceleration constraint, speed constraint, and safety distance constraint; consider the safety distance problem between two adjacent agricultural machinery in the actual working process of the smart agricultural machinery formation, and set the safety performance constraint; Step S4: Construct the optimal control problem, consider the actual operation trajectory and safety performance limitations of the smart agricultural machinery, introduce the maximum-minimum model predictive control strategy, solve the online controller that meets the conditions, and obtain the optimal control input.

2. The method for controlling a UAV formation based on communication delay and external disturbance according to claim 1 is characterized in that: In step S1, it is assumed that the smart agricultural machinery formation is traveling in a straight line with a fixed spacing. It is relatively convenient to design the controller using the relative kinematics of two consecutive adjacent agricultural machinery. Therefore, the spatial kinematic equation of the jth smart agricultural machinery is: in x j (t) = [Δp j (t) Δv j (t) a j (t)] T is the information of the jth smart agricultural machine, x j-1 (t) = [Δp j-1 (t) Δv j-1 (t) a j-1 (t)] T Information about the j-1st smart agricultural machine ahead. represents the position error between two adjacent agricultural machines, p j-1 (t), p j (t) represents the position of the preceding machine and the current agricultural machine position information, A, B, G represent appropriate matrices, d s represent the position error and fixed spacing respectively, represents the relative speed between two adjacent agricultural machines, and considering the motor time constant ζ, we have Setting up the controller j (t) is the expected acceleration a des (t), a j (t) denote the acceleration derivative and acceleration respectively.

3. The method for predicting and controlling the formation of smart agricultural machinery automatic driving in communication-impaired scenarios according to claim 1 is characterized in that: In step S1, the state of the entire smart agricultural machinery formation is defined by using the centralized control of the smart agricultural machinery formation. Assuming that there are M smart agricultural machinery traveling in a straight line, then x(t)=col(x j (t),j∈M),u(t)=col(u j (t), j∈M), col, x(t), and u(t) represent vector, state information, and controller to be solved respectively. The entire smart agricultural machinery formation system is in,, x 0 (t) = 0, represents the appropriate matrix, x 0 (t) is the leader’s attitude information. Here, all following agricultural machines take the adjacent front smart agricultural machine as the tracking target, and the first smart agricultural machine is the leader.

4. The method for predicting and controlling the formation of smart agricultural machinery automatic driving in communication-impaired scenarios according to claim 1 is characterized in that: In step S2, in order to make the dynamic error tend to be stable, the state equation system (2) is discretized by the forward Euler method, Tr is the sampling time, and external interference is considered to obtain a more realistic dynamic model of smart agricultural machinery. in is a suitable matrix, represents the appropriate matrix after sampling, T r represents the sampling time, w k For wind disturbance and mechanical friction, and represents a set of perturbations; Considering communication impairment, the buffer is used to store the hysteresis data, and the modeling of communication impairment is introduced. Under the condition of impairment, the above is restated as in Solve the controller u k-2τ , I M represents the identity matrix, u k-τ represents a delay controller, τ is the delay and is related to the time T r There is a certain linear relationship. An extended sequence of controller delays was added. Respectively represent the state information at time k+1 and time k.

5. The method for predicting and controlling the automatic driving formation of smart agricultural machinery in a communication-impaired scenario according to claim 1 is characterized in that: In step S3, the acceleration, speed, and safety distance are constrained respectively to construct the state constraint inequality. Represents a set of safety performance constraints, according to a min =u min =u j (tt d )≤u max =a max It can be seen that the requirements for position error, velocity error, and acceleration are respectively: represents the appropriate matrix, a max 、a min , Respectively represent the maximum acceleration, minimum acceleration, maximum position error, minimum position error, maximum speed error, and minimum speed error, f x , Represents an appropriate matrix, and the safety performance constraints of smart agricultural machinery are described by the above polyhedron set, that is, the linear matrix inequality; satisfying its safety performance constraints.

6. The method for predicting and controlling the formation of smart agricultural machinery automatic driving in communication-impaired scenarios according to claim 1, characterized in that: The specific steps of step S4 are: S401. First, define the cost function through the maximum-minimum model predictive control strategy Q, R, P represent appropriate weight matrices, Respectively represent the state information and delay controller at different times; At the same time, under this cost function condition, relevant safety performance constraints are added The above constraints represent the controller u k+i|k , acceleration, velocity set x k+i|k , and the constraints on the disturbance variable w, Defines the stability that model predictive control needs to meet at the end of the forecast time horizon; S402. In order to better analyze the problem, the optimization problem (6) is expanded to define stacking variables Also got The new state equation is in represents the appropriate matrix, and are respectively The Kronecker product of Therefore, the optimization problem 6 can be rewritten as In the formula represents the weight matrix, where the safety performance constraint is rewritten as follows S403. Find a definite value to replace the maximum-minimum problem, so as to effectively generate the problem to be solved, so there is By effectively solving the minimum value t, it becomes possible to solve uncertain bounded external disturbances; S404. According to Schur's complement theorem, Transformed into uncertain LMI with respect to U and t, where X and U, W are linearly parameterized; S405. Insert the definition of X Separate deterministic terms and uncertainty S406. Consider the worst case of uncertain external interference terms and obtain the maximum value set at the current moment The LMI linear matrix inequality with affine uncertainty is established using the maximum principle; S407. According to the affine uncertainty robust linear matrix inequality theorem, the uncertain LMI is transformed into an LMI containing only certain terms, state constraints and control constraints are added, and the optimal control input is determined by solving the problem using the tools YALMIP and MOSEK; S408. Define the terminal constraint invariant set. 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 in a constant term set, use Schur's complement theorem to transform it into a two-variable product problem. S409. Introduce a positive real scalar to eliminate the uncertain interference term and constrain the state and control constraints A congruence transformation is performed and converted into a matrix inequality. Given the value of a single variable in the product of two variables, the maximum volume invariant ellipsoid is solved according to the maximum value theorem of the linear function in the ellipsoid, thus ensuring the stability of the system state.

7. A smart agricultural machine, characterized in that: include: A data link is established between the follower smart agricultural machinery and the adjacent front agricultural machinery through the wireless communication module, so that the follower smart agricultural machinery can receive the position and speed information in front in real time, and adjust the driving control according to the updated control law to make the formation stable and safe. There is also a storage device, and the processor calls the executable program code stored in the storage device to execute the smart agricultural machinery automatic driving formation prediction control method for communication impaired scenarios as described in any one of claims 1 to 6.

8. A non-transitory computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which, when called, are used to execute a smart agricultural machinery automatic driving formation prediction and control method for communication-impaired scenarios as described in any one of claims 1-6.

Images