Fuzzy self-adaptive control system for gear shifting manipulator of robot based on impedance control

Abstract

The invention discloses a fuzzy self-adaptive control system based on an impedance control robot gear shifting manipulator, which comprises a gear shifting manipulator and an impedance controller, and is characterized in that the impedance controller comprises the following three parts: an outer ring is an impedance control system, and an inner ring is a fuzzy PID self-adaptive controller and an Eemd in-place detection device. According to the invention, gear shifting of the tractor can be controlled, so that'machine-to-person 'in the gear shifting process is realized, and the flexible gear shifting action in the unmanned working process of the tractor is realized. The control method has the advantages of being simple in structure, low in cost and capable of effectively solving the problems existing in a front-mounted type, the control scheme can also enable the gear shifting process to have the self-adaptive effect, and smooth gear shifting is achieved. The gear shifting manipulator adopts a rear-mounted design and can be mounted on a traditional tractor without damage, so that the equipment production cost is controlled, and meanwhile, the existing traditional tractor is effectively utilized. The problem of shortage of labor force can be effectively solved through the full-unmanned driving process, and agricultural productivity is improved.

Description

Technical Field

[0001] The invention relates to the technical field of gear shifting of rear-mounted tractors, and in particular to a fuzzy adaptive control system based on an impedance-controlled robot gear shifting manipulator. Background Art

[0002] At present, the existing control of the shifting part of the rear-mounted tractor design is still based on the traditional PID control, which is difficult to adapt to complex farmland scenes. Different from roads, the farmland environment is more complex and changeable, and traditional tractors are not as advanced as the current cars. At present, many intelligent tractor devices pay more attention to the design itself, but ignore the control effect in the actual farmland environment. Assuming that in an environment with greater forward resistance such as a raised small earth slope or a mud field, it is difficult to complete a smooth shift like a flat road. Many tractors still rely on gear structures to achieve switching. Larger shift gears correspond to greater shift resistance and are more difficult to control. At this time, it is easy to fail to engage in gear or go out of gear. Even if it can be engaged, it will cause damage to the manipulator and the tractor's own structure due to "dead force". At this time, a device that can realize active and smooth control of the shifting process is needed. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a fuzzy adaptive control system based on an impedance-controlled robot shifting manipulator, including a shifting manipulator and an impedance controller; the impedance controller includes the following three parts: an outer loop is an impedance control system, which is used to ensure the flexibility of the shifting process and avoid physical damage; an inner loop is a fuzzy PID adaptive controller; and an EEmd in-position detection device is used to prevent offside.

[0004] Preferably, the shifting robot comprises an electric push rod, a linear guide slide, a stepping motor, an elastic shifting module and an operating lever.

[0005] The elastic shift module includes a slider A, a slider B and a spring. The slider A is sleeved on the upper thread of the operating rod, and the slider B is sleeved on the outside of the slider A. The centers of both ends of the slider A are respectively provided with cylindrical slots, and the inner side of the slider B is oppositely provided with cylindrical slots. The two ends of the spring are respectively located in the two opposite cylindrical slots; two threaded holes are respectively provided on the left and right sides of the slider A, and sliding grooves are correspondingly provided on both sides of the slider B, so that the four bolts fixing the slider B can slide freely;

[0006] The head of the electric push rod is connected to the slider B through a buckle, and the tail of the electric push rod is installed on the mounting frame. The mounting frame is installed and fixed through the original hole position inside the cab. The other end of the mounting frame is fixedly connected to the slide. The slide is located on the linear guide rail, and the linear guide rail is located on the output shaft of the stepper motor. The slide can slide along the linear guide rail driven by the stepper motor; the gear shifting robot is provided with an encoder, which is used to measure the motor rotation angle in real time.

[0007] Preferably, the threaded hole is an M3 threaded hole, the width of the slide groove is 3.5 mm; and the compression ratio of the spring is 50%.

[0008] Preferably, the impedance control M d , B d , K d represent the inertia matrix, damping matrix and stiffness matrix of the target impedance model respectively; X represents the position vector of the shift robot, X r represents the desired position of the shift robot, F c Represents the contact force between the end of the shift robot and the environment, F d represents the expected contact force between the end of the shift robot and the environment;

[0009]

[0010] The impedance control system detects the motor's current signal instead of the force signal, superimposes the position correction signal and the expected position signal, and then uses PID control on the position signal.

[0011] Preferably, the fuzzy PID adaptive controller includes a PID controller on the forward channel and a fuzzy controller on the feedback channel. The fuzzy PID adaptive controller is used to adjust the PID parameters in real time to complete adaptive control; the ID controller is used for the fuzzy controller to perform PID parameter self-tuning; the input of the fuzzy controller is the motor current I and the rate of change of the current The output is K p , K i , K d The increments of, namely, △Kp, △Ki, △Kd;

[0012] The PID controller obtains the actual control current size and uses fuzzy rules to update the PID control parameters online; the current feedback value is turned on and off through the EEMD for in-position detection. If it is detected to be in position, the motor drive is stopped, otherwise, the set motor forward or reverse rotation is continued to control the handle forward and backward; the system's initial input expected position is determined by the unmanned driving decision or the remote driving operating lever position.

[0013] Preferably, the real-time motor current signal is defined as I, and the rate of change of I The range of variation is the domain of the fuzzy set, which is: I∈[-5,5], Fuzzy subset The elements NB, NM, NS, ZE, PS, PM, and PB in the subset represent negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, respectively. The basic domain of the output control variable is [-3, 3], and its fuzzy subset is {NB, NM, NS, ZE, PS, PM, PB}; the membership functions of the input and output NB and PB are all Gaussian functions in their respective domains, using formula (2), and the rest all use triangular membership functions;

[0014]

[0015] Preferably, EMD decomposes the input signal into several eigenmode functions and a residual, which is composed of the following formula (3):

[0016]

[0017] Where I(n) represents the input signal, Indicates M th The eigenmode function, Res M (n) represents the residual; the process of extracting IMF is called screening, and the screening process is as follows:

[0018] Mark the local extreme points;

[0019] Connect the maximum points to form the upper envelope, and connect the minimum points to form the lower envelope;

[0020] Find the mean signal line of the upper and lower envelopes;

[0021] Subtract the upper and lower envelope mean signals from the input signal to get the intermediate signal:

[0022] X(t)-m 1 =h 1 (4)

[0023] Determine whether the intermediate signal meets the two conditions of IMF. If so, the signal is an IMF component; if not, repeat the analysis of the first four steps based on the signal.

[0024] Based on EMD, EEMD adds white noise before decomposing the signal and averages the components obtained afterwards to reduce the interference caused by the noise.

[0025] EEMD decomposition consists of four steps:

[0026] Set the number of times the original signal is processed m;

[0027] Add random white noise to the m original signals respectively to form a series of new signals;

[0028] Perform EMD decomposition on this series of new signals to obtain a series of IMF components;

[0029] The IMF components of the corresponding modes are averaged respectively to obtain the EEMD decomposition results.

[0030] The EEMD technology is used to determine whether the gear shift is completed. When the gear shift is completed, the motor is immediately controlled to stop, so that the manipulator and the gear shift handle stop. At this time, the gear shift operation is successful;

[0031] The gear shift detection based on EEMD is divided into the following three steps:

[0032] Calculate the absolute value of the real-time shift current

[0033] Determine the peak value of the shift current

[0034] Determine the first minimum value after the shift current peak.

[0035] Preferably, it also includes an impedance control system, which includes a control module, a body recognition perception and positioning module, a motor module, a data acquisition module, a shift simulator module and a power module. The control module includes a main controller IPC, a host computer and a DC motor driver; the data acquisition module includes a current acquisition card, a Hall sensor and a digital quantity acquisition card; the body recognition perception and positioning module is connected to the main controller IPC through a communication module;

[0036] The surrounding recognition perception and positioning module is used to dynamically perceive the working environment of the tractor and perform visual obstacle avoidance and navigation through image recognition algorithms;

[0037] The main controller IPC generates the expected position of the tractor operating lever and enters the trained impedance-controlled fuzzy PID control system program of the gear shift manipulator, then outputs the control result and transmits it to the DC motor driver;

[0038] The DC motor driver receives instructions and controls the motor to rotate forward or reverse according to the level to realize forward or backward movement.

[0039] Preferably, the surrounding recognition perception and positioning module includes navigation, radar, ultrasound and camera;

[0040] The camera is used to collect real-time surrounding video, and the main controller IPC determines whether to perform a gear shift operation based on the real-time surrounding video transmitted back by the camera;

[0041] The gear shift simulator module sends data to the Alibaba Cloud; the main controller IPC receives the data sent by the gear shift simulator and sends a gear shift instruction;

[0042] Navigation provides the unmanned tractor with high-precision positioning data and body posture information, and then predicts the next gear action based on the specific situation of the vehicle;

[0043] The data acquisition module provides information feedback for the closed-loop control of the shift module;

[0044] The gear shift simulator module uses the Logitech G29 gear shift simulator;

[0045] The power module is a parallel connection of the tractor's own 12V battery and a 12V power supply;

[0046] The motor module adopts LK-150 DC motor.

[0047] Preferably, the measurement and control system includes:

[0048] The main control layer is composed of the main controller IPC, which has two modes: automatic driving and remote control driving;

[0049] The main task of the decision-making layer is to analyze the real-time feedback data from various sensors on the robot. If it is an unmanned driving measurement and control system, the data is processed based on multi-sensor information fusion technology and fed back to the main control layer; according to the decision of the main control layer, the shift control command is issued to complete the corresponding action; if it is a remote-operated measurement and control system, the operator controls the simulator according to the data fed back by the host computer to achieve the corresponding forward and backward actions;

[0050] The sensor layer is composed of sensors carried by the robot; the laser radar, ultrasonic ranging module and surrounding camera dynamically perceive the working environment of the tractor, and perform visual obstacle avoidance and navigation through image recognition algorithms; the combined navigation module provides high-precision positioning data and body posture information for the unmanned tractor, and then predicts the next gear action according to the specific situation of the vehicle; the Hall position sensor and current sensor provide information feedback for the closed-loop control of the shift module;

[0051] The bottom-level actuator is composed of a shift control module, a motor driver, and a shift control motor. This layer mainly completes the specific shifting action of the shift robot.

[0052] In view of the problems existing in the actual production and promotion of unmanned tractors, the present invention provides a shifting robot based on impedance control, which can control the gear shifting of the tractor, realize the "machine replacing people" in the gear shifting process, and realize smooth gear shifting during the unmanned working process of the tractor. The advantages of low cost and after-installation can effectively solve the problems existing in the front-installation type. The control scheme of the present invention can also make the gear shifting process have an adaptive effect and realize smooth gear shifting. The gear shifting robot adopts a rear-installation design and can be installed on traditional tractors without damage, which not only controls the production cost of the equipment, but also effectively utilizes the existing traditional tractors. The fully unmanned driving process can effectively solve the problem of labor shortage and improve agricultural productivity. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The technical solution of the present invention will be further described below in conjunction with the accompanying drawings:

[0054] Figure 1 It is a structural schematic diagram of the gear shift elasticity module.

[0055] Figure 2 It is the overall structure diagram of the gear shift robot.

[0056] Figure 3 It is the gear shifting operation space of traditional PID control.

[0057] Figure 4 It is the principle block diagram of the fuzzy adaptive control system of the gear shifting manipulator of the tractor driving robot based on impedance control.

[0058] Figure 5 It is the fuzzy controller structure of the gear shift robot.

[0059] Figure 6 It is the EEMD flow chart.

[0060] Figure 7 It is the effect diagram of the overall control system controlling the forward process. DETAILED DESCRIPTION

[0061] 1. Gear shifting robot configuration

[0062] Through theoretical calculation, the three-dimensional spherical motion of the joystick is reduced to two-dimensional linear motion, so the motion trajectory of the electric push rod is limited to a straight line along the direction of the handle movement. In order to make the shifting process smoother, the shifting device is given the degree of rotation around the axis at the front and end of the electric push rod. For the rod-type joystick, the joystick has a small range of movement after the gear is shifted into place. According to this feature, an elastic shift module is designed, such as Figure 1As shown, the structure consists of a slider A1, a slider B2 and a spring 3 with a compression ratio of 50%. The slider A1 is sleeved on the upper thread of the shift lever. Since the shift handle itself has threads, the slider A1 can be pressed by the threads. Two blue springs 3 with a compression ratio of 50% are placed in two cylindrical slots opened at the center of the slider A1 and the slider B2, respectively, to connect the slider A1 and the slider B2. At the same time, it is also necessary to open two M3 threaded holes on the left and right sides of the slider A1, and open a 3.5mm wide slide groove on both sides of the slider B2, so that the four bolts fixing the slider B2 can slide freely. This elastic structure reduces the rigidity of the entire module when entering the gear, making the shifting process more reasonable and close to manual shifting.

[0063] This module installs the tail of the electric push rod 5 for gear shifting on the mounting frame. The head of the electric push rod 5 is connected to the slider B2 through a quick-release buckle (for easy disassembly and manual intervention). By pressing the top button of the quick-insert buckle, the steel ball of the bayonet can be retracted into the inside of the bayonet, so that the buckle can be quickly removed, and when the button is released, the steel column pops out and can be stuck and will not loosen. The gear shift module mounting frame is installed and fixed through the original hole position inside the cab. In the X direction, the CY-42YT200 linear light rod guide rail slide 8 is selected, and it is equipped with a 42H4812 two-phase hybrid stepper motor 6. The friction of the lead screw movement is very small, and relatively high-precision position control is achieved; in the Y direction, the Changzhou Louis LK-150 electric push rod with a gain code disk is used. The encoder can measure the motor rotation angle in real time, and further calculate the telescopic length of the push rod. This constitutes Figure 2 The entire shift module is shown.

[0064] like Figure 3 As shown in the figure, the position repeatability in the X direction is high, and the friction force is much smaller than the thrust of the stepper motor. Therefore, the position repeatability in the Y direction is poor when the stepper motor is controlled by the traditional Pid, ​​and there are time-varying and nonlinear characteristics in the gear shifting process. Therefore, the fuzzy adaptive control system based on impedance control of the gear shifting manipulator of the tractor driving robot is used to control the extension and retraction of the push rod to complete the gear shifting action.

[0065] 2. Impedance controller

[0066] The fuzzy PID adaptive control system of the traditional tractor power shift manipulator based on impedance control of the present invention is based on current feedback signals instead of force feedback in impedance control. Because the current is within a certain range, it can be known from motor knowledge that the current of the motor is proportional to the torque, and current feedback is faster than force feedback, and can better meet the real-time requirements of the system shifting. Compared with the force sensor, conversion module and analog quantity acquisition card required for force feedback, it is only necessary to connect the current acquisition card in series in the circuit of the motor to conveniently measure its current value, which is more convenient and convenient. After the impedance controls the large outer loop, the actual position of the current shift handle can be obtained, and the expected position at the beginning can be subtracted to obtain The difference to the position enters the next cycle and the forward path of the control system. The corrected difference can be obtained after calculation with the new position of the handle, which is a basic position control of the inner loop. The actual control current can be obtained by the PID position controller. The PID controller used in the present invention uses fuzzy rules to update the PID control parameters online in order to increase the adaptability and robustness of the entire system and help the manipulator cope with the position environment. The establishment of fuzzy rules is formulated through data obtained from multiple experiments; the feedback value of the current will be turned on and off through the EEMD for in-position detection. If it is detected to be in place, the motor drive will be stopped. Otherwise, the set motor will continue to rotate forward or reverse to control the handle forward and backward. The applicable conditions for the manipulator designed by the present invention are remote driving or unmanned driving, so the system's initial input expected position is determined by the unmanned driving decision or the position of the remote driving operating lever. The principle block diagram of the traditional tractor shift upgrade intelligent system based on impedance control is as follows. Figure 4 As shown, X r represents the desired position of the shift robot, X c Represents the actual position of the handle, F c represents the contact force between the end of the shift robot and the environment, I c represents the end current of the gear shift robot, e represents the position deviation, I represents the current, n represents the speed, △Kp, △Ki, and △Kd represent the correction amounts of the PID parameters respectively.

[0067] The fuzzy adaptive control system of the gear shifting manipulator of the tractor driving robot based on impedance control is mainly composed of three parts: (1) the outer loop is an impedance control system to ensure the flexibility of the gear shifting process and avoid physical damage from the software level; (2) the inner loop is a fuzzy PID adaptive controller to improve the performance index of the gear shifting process, which is fast, accurate and stable. (3) EEmd in-position detection technology to prevent offside.

[0068] (1) Impedance control of the gear shift robot

[0069] Ordinary position control is to control the end of the robot to the designated location. This kind of control is too "rigid", which will bring many problems. Take the shifting process of the tractor as an example. Suppose the shifting manipulator is controlled to move forward 1 cm. In the complex and changeable farmland environment, can the manipulator really move forward 1 cm, and is the previously set 1 cm reasonable? If the actual gear position requires the manipulator to move forward only 0.95 cm, but the command received by the manipulator is to move forward 1 cm, then the shift handle has been in contact with the tractor mechanical structure, but the motor is still moving forward. The final result will only be that the motor is always blocked and scrapped, or the tractor's own shifting device is damaged. So at this time, it is necessary to avoid such a situation through active compliance control. Active compliance control is divided into impedance control, admittance control and mixed force position control. In the present invention, impedance control is a general term for active compliance.

[0070] Impedance Control M d , B d , K d represent the inertia matrix, damping matrix and stiffness matrix of the target impedance model respectively; X represents the position vector of the shift robot, X r represents the desired position of the shift robot, F c Represents the contact force between the end of the shift robot and the environment, F d Represents the expected contact force between the end of the shift robot and the environment.

[0071]

[0072] In formula (1), it is assumed that the position control is error-free, but in fact it is impossible, so the present invention will improve the traditional impedance control. Through the above formula, the contact force between the end of the manipulator and the environment can be converted into a position signal. When the shifting manipulator has reached the actual required position, the contact force will become larger and larger, and the position control forward error will become smaller and smaller. Finally, it will stop at the position where we actually want it to stop, and will not cause the mechanical damage mentioned above. At the same time, because within a certain range, the current size is linearly related to the push rod thrust size, the force signal can be replaced by detecting the motor current signal, and the correction signal of this position and the expected position signal are superimposed, and then the position signal is controlled by PID.

[0073] (2) Fuzzy PID adaptive controller for gear shifting robot

[0074] At present, many control algorithms have emerged in intelligent control, such as neural network algorithms, genetic algorithms, etc., which can be applied to complex nonlinear signal processing. However, most of these algorithms have the disadvantages of slow learning and training, and they are trained through offline models, which is not suitable for tractor shifting control in complex and changeable field operation environments. Fuzzy PID can adjust parameters online in real time, maintaining the reliability and real-time performance of traditional PID systems in position control, while having better adaptability and robustness. The entire controller is divided into two parts, namely the PID controller on the forward channel and the fuzzy controller on the feedback path. The role of fuzzy control is to adjust the parameters of PID in real time to complete adaptive control, because the PID parameters are not the same in the environment. Such as Figure 5 As shown, the input of the fuzzy controller is the motor current I and the rate of change of the current The output is K p , K i , K d The increments of, that is, △Kp, △Ki, △Kd.

[0075] The initial PID parameter values ​​are measured through simulation experiments, and the fuzzy rules are trained by obtaining effective data through direct human control of the actuator. The real-time motor current signal is defined as I, and the rate of change of I The range of variation is the domain of the fuzzy set, which is: I∈[-5,5], Fuzzy subset The elements NB, NM, NS, ZE, PS, PM, and PB in the subset represent negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, respectively. The basic domain of the output control variable is [-3, 3], and its fuzzy subset is {NB, NM, NS, ZE, PS, PM, PB}; the membership functions of the input and output NB and PB are all Gaussian functions in their respective domains, using formula (2), and the rest all use triangular membership functions;

[0076]

[0077] (3)EEMD in-place detection

[0078] Empirical Mode Decomposition (EMD) is a way of processing signals without setting basis functions, and only decomposes the signal according to its own time scale characteristics. The method is fully adaptive and generates a basis representing the data only from and based on these data. The basis functions, called Intrinsic Mode Functions (IMFs) represent a set of local orthogonal basis functions whose amplitude and frequency vary with time. EMD decomposition is similar to traditional Fourier or wavelet decomposition, but EMD represents a completely data-driven unsupervised signal decomposition without the need for any a priori defined basis system. EMD also satisfies the perfect reconstruction property, that is, all extracted IMFs are superimposed with the residual slow trend to reconstruct the original signal without information loss and distortion.

[0079] EMD decomposes the input signal into several intrinsic mode functions and a residual, which is composed of the following formula (3).

[0080]

[0081] Where I(n) represents the input signal, Indicates M th The eigenmode function, Res M (n) represents the residual.

[0082] The process of extracting IMF is called screening, and the screening process is as follows:

[0083] (1) Mark the local extreme points;

[0084] (2) Connect the maximum points to form the upper envelope, and connect the minimum points to form the lower envelope;

[0085] (3) Find the mean signal line of the upper and lower envelopes;

[0086] (4) Subtract the upper and lower envelope mean signals from the input signal to obtain the intermediate signal:

[0087] X(t)-m 1 =h 1 (4)

[0088] (5) Determine whether the intermediate signal meets the two conditions of IMF. If so, the signal is an IMF component. If not, repeat the analysis of (1) to (4) based on the signal. The acquisition of IMF components usually requires several iterations.

[0089] However, there are still some problems in the use of EMD. The most important problem is that it will produce modal aliasing, that is, signals of different modes are mixed together. EEMD is based on EMD. Before decomposing the signal, white noise is added, and the components obtained are averaged to reduce the interference caused by small noise.

[0090] EEMD decomposition is mainly divided into 4 steps, such as Figure 6 As shown:

[0091] Set the number of times the original signal is processed m;

[0092] Add random white noise to the m original signals respectively to form a series of new signals;

[0093] Perform EMD decomposition on this series of new signals to obtain a series of IMF components;

[0094] The IMF components of the corresponding modes are averaged respectively to obtain the EEMD decomposition results.

[0095] According to the structure of the tractor shifter, when the shift handle reaches the engagement point, the resistance will suddenly drop. Based on this feature, we can use EEMD technology to determine whether the shift is completed. When the shift action is completed, the motor is immediately controlled to stop, so that the manipulator and the shift handle stop. At this time, the shift operation is successful. The shift engagement detection based on EEMD is divided into the following three steps:

[0096] a. Calculate the absolute value of the real-time shift current;

[0097] b. Determine the peak value of the shift current;

[0098] c. Determine the first minimum value after the shift current peak.

[0099] 3. Measurement and control system

[0100] 3.1 Hardware Design

[0101] The hardware design of the impedance control system of the gear shift manipulator mainly designs the following modules: controller module, body recognition perception and positioning module, motor module, data acquisition module, and system and power management system design. The actuators of the four systems need to meet the working requirements of the driving robot and adapt to the overall measurement and control system of the robot; the control circuit of each actuator needs to complete accurate and fast control; the monitoring sensors of each system need to feedback the working status of each actuator to the main controller IPC in real time, and the entire measurement and control system forms a closed-loop control. The control module includes the main controller IPC, the host computer and the DC motor driver; the acquisition module includes the current acquisition card, the Hall sensor and the digital acquisition card; the gear shift simulator module is mainly the Logitech G29 gear shift simulator; the body recognition perception and positioning module mainly includes navigation, radar, ultrasound, and camera. The power module is a parallel connection of the tractor's own 12V battery and 12V power supply; the motor module uses an LK-150 DC motor.

[0102] When the system uses assisted unmanned driving tests, the operator will first determine whether to shift gears based on the real-time surrounding video sent back by the camera. After the operator controls the simulator, the data will be sent to the Alibaba Cloud. Then, the on-board controller IPC will receive the data and send a shift instruction. At this time, the homemade DC motor driver will receive the instruction and control the forward and reverse rotation of the motor to move forward or backward through the level. If a tractor is used to assist unmanned driving control, the on-board controller IPC will automatically fuse and process the data sent back by the tractor's sensors, make a decision, and send the control instruction to the driver to complete the control.

[0103] 3.2 Measurement and control system

[0104] The shifting manipulator is a post-installation modification for traditional tractors. The purpose of the modification is to better realize unmanned driving and remote driving. Therefore, in order to ensure the practicality of the design, the measurement and control system adopts auxiliary unmanned driving and teleoperated remote driving systems respectively. The purpose of measurement and control is to test whether the shifting manipulator can complete the specified tasks and achieve the expected results under these two working environments, and can also realize flexible shifting in the environment of the position.

[0105] (1) The main control layer of the system is composed of the on-board main controller IPC, which has two modes: automatic driving and remote control driving. The industrial computer is installed with the Ubuntu20.04 system. The robot software development is based on the ROS-noetic development environment to build a friendly human-computer interaction interface.

[0106] (2) The main task of the decision-making layer is to analyze the real-time feedback data from various sensors on the robot. If it is an unmanned driving measurement and control system, the data is processed based on multi-sensor information fusion technology and fed back to the main control layer. According to the decision of the main control layer, the shift control command is issued to complete the corresponding action. If it is a remote control system, the operator controls the simulator according to the data fed back by the host computer to achieve the corresponding forward and backward actions.

[0107] (3) The sensor layer consists of sensors carried by the robot. LiDAR, ultrasonic ranging modules and body cameras dynamically perceive the tractor's working environment and perform visual obstacle avoidance and navigation through image recognition algorithms. The combined navigation module provides high-precision positioning data and body posture information for the unmanned tractor, and then predicts the next gear action based on the specific situation of the vehicle. Hall position sensors and current sensors provide information feedback for the closed-loop control of the shift module.

[0108] (4) The bottom-level actuator is composed of a gear shift control module, a motor driver, and a gear shift control motor. This layer mainly completes the specific gear shifting action of the gear shift robot.

[0109] The communication between the master and slave controllers is mainly based on the RS485 bus. Because of its outstanding reliability, real-time performance and ultra-high transmission rate, it is widely used in the communication between different components of the tractor. It can provide reliable data interaction for the master and slave controllers of the driving robot. The bus structure enables the robot to reduce the laying of wiring harnesses and save costs. The communication of the sensor is mainly based on network port communication and RS232, and the communication method is determined by the sensor itself. The communication between the main controller and the host computer is realized through Alibaba Cloud's cloud communication, which can achieve communication without distance restrictions, but the disadvantage is that it is closely related to the network signal.

[0110] The measurement and control software of the shifting manipulator is written in the on-board controller IPC. The main controller can receive the cloud command in real time to remotely control the shifting, or it can shift gears autonomously according to the vehicle's surrounding conditions. Each sensor feeds back the operating status in real time. After transmitting the data to the main controller, the main controller will generate the expected position of the tractor's operating lever and enter the trained impedance-controlled fuzzy PID control system program of the shifting manipulator. Then, it outputs the control result and transmits it to the self-made DC motor driver to control the push rod and joystick through the level. At the same time, the underlying control software is encapsulated to facilitate the upper-level control software to call.

[0111] The teleoperation system allows the operator and the controlled robot to be separated by a certain distance, and data interaction is achieved through the Alibaba Cloud platform link. The system adopts a master-slave system model. The operator sends the shift control command through the simulated driving device. The control command is processed and transmitted to the slave tractor through the cloud communication system. The slave tractor interprets the control command as the control signal of the actuator and then controls the shift robot. At the same time, all sensor information in the slave tractor status information perception system is also transmitted back to the master operator through the cloud communication system, and the entire system forms a closed loop.

[0112] For various information on the main end, cloud end, and slave end of the tractor remote driving system, a visual representation of each information is designed through Qt and Baidu Maps - the human-computer interaction interface, which has an overall HUD-like effect and can provide the operator with a higher sense of visual presence.

[0113] 1. Simulation

[0114] Based on this control system, the experimental simulation is carried out. The input signals of the simulation experiment use a set of forward process current signals and a set of reverse process current signals. Figure 7 As shown, it can be seen that the signal of the whole process has no divergence, which is in line with the stability of the system; and the signal can be guaranteed to follow. Compared with the traditional PID, the overshoot is also reduced, which is in line with the accuracy of the system. In each stage of rapid change of the signal, the response speed of this system is faster than the traditional PID, which is in line with the rapid design of the system. In summary, the design of the system meets the requirements. At the same time, it can be seen that after the EEMD is detected in place, the control of the switch quantity is completed by selecting the module. After the in-place signal is detected, the control quantity of the system becomes 0.

[0115] According to the mechanical design of the present invention, parts such as the clamping device are processed, and components such as the motor, slide, guide rail, etc. are purchased for assembly.

[0116] 2. Remote control effect

[0117] Run the ROS node on the lower computer, run the communication node and Alibaba Cloud node on the upper computer, and complete the communication based on the RS485 bus.

[0118] 3. Unmanned driving control effect

[0119] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A fuzzy adaptive control system for a gear shifting manipulator based on impedance control, characterized in that: Including a shift robot and an impedance controller; The impedance controller includes the following three parts: The outer ring is an impedance control system, which is used to ensure the smoothness of the shifting process and avoid physical damage; The inner loop is a fuzzy PID adaptive controller; EEmd in-position detection device is used to prevent offside.

2. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 1 is characterized in that: The gear shifting robot comprises an electric push rod (5), a linear guide rail slide, a stepping motor (6), an elastic gear shifting module and an operating rod (4). The elastic shift module comprises a slider A (1), a slider B (2) and a spring (3), wherein the slider A (1) is sleeved on the upper thread of the operating rod (4), the slider B (2) is sleeved on the outer side of the slider A (1), the centers of the two ends of the slider A (1) are respectively provided with cylindrical slots, the inner side of the slider B (2) is provided with cylindrical slots opposite to each other, and the two ends of the spring (3) are respectively located in the two opposite cylindrical slots; the left and right sides of the slider A (1) are respectively provided with two threaded holes, and the two sides of the slider B (2) are correspondingly provided with sliding grooves, so that the four bolts fixing the slider B (2) can slide freely; The head of the electric push rod (5) is connected to the slider B (2) through a buckle, and the tail of the electric push rod (5) is installed on the mounting frame (7). The mounting frame (7) is installed and fixed through the original hole position inside the cab. The other end of the mounting frame (7) is fixedly connected to the slide (8). The slide (8) is located on the linear guide rail (9), and the linear guide rail (9) is located on the output shaft of the stepping motor (6). The slide (8) can slide along the linear guide rail (9) driven by the stepping motor (6); the gear shifting manipulator is provided with an encoder, and the encoder is used to measure the rotation angle of the motor in real time.

3. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 2 is characterized in that: The threaded hole is an M3 threaded hole, the width of the slide groove is 3.5 mm; and the compression ratio of the spring is 50%.

4. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 2 is characterized in that: Impedance Control M d , B d , K d represent the inertia matrix, damping matrix and stiffness matrix of the target impedance model respectively; X represents the position vector of the shift robot, X r represents the desired position of the shift robot, F c Represents the contact force between the end of the shift robot and the environment, F d represents the expected contact force between the end of the shift robot and the environment; The impedance control system detects the motor's current signal instead of the force signal, superimposes the position correction signal and the expected position signal, and then uses PID control on the position signal.

5. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 4 is characterized in that: The fuzzy PID adaptive controller includes a PID controller on the forward channel and a fuzzy controller on the feedback channel. The fuzzy PID adaptive controller is used to adjust the PID parameters in real time to complete adaptive control; the ID controller is used for the fuzzy controller to perform PID parameter self-tuning; the input of the fuzzy controller is the motor current I and the rate of change of the current The output is K p , K i , K d The increments of, namely, △Kp, △Ki, △Kd; The PID controller obtains the actual control current size and uses fuzzy rules to update the PID control parameters online; the current feedback value is turned on and off through the EEMD for in-position detection. If it is detected to be in position, the motor drive is stopped, otherwise, the set motor forward or reverse rotation is continued to control the handle forward and backward; the system's initial input expected position is determined by the unmanned driving decision or the remote driving operating lever position.

6. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 5 is characterized in that: Define the real-time motor current signal as I, the rate of change of I The range of variation is the domain of the fuzzy set, which is: I∈[-5,5], Fuzzy subset The elements NB, NM, NS, ZE, PS, PM, and PB in the subset represent negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, respectively. The basic domain of the output control variable is [-3, 3], and its fuzzy subset is {NB, NM, NS, ZE, PS, PM, PB}; the membership functions of the input and output NB and PB are all Gaussian functions in their respective domains, using formula (2), and the rest all use triangular membership functions; 7. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 6 is characterized in that: EMD decomposes the input signal into several eigenmode functions and a residual, which is composed of the following formula (3); Where I(n) represents the input signal, Indicates M th The eigenmode function, Res M (n) represents the residual; the process of extracting IMF is called screening, and the screening process is as follows: Mark the local extreme points; Connect the maximum points to form the upper envelope, and connect the minimum points to form the lower envelope; Find the mean signal line of the upper and lower envelopes; Subtract the upper and lower envelope mean signals from the input signal to get the intermediate signal: X(t)-m1=h1 (4) Determine whether the intermediate signal meets the two conditions of IMF. If so, the signal is an IMF component. If not, repeat the analysis of the first four steps based on the signal. Based on EMD, EEMD adds white noise before decomposing the signal and averages the components obtained afterwards to reduce the interference caused by the noise. EEMD decomposition consists of four steps: Set the number of times the original signal is processed m; Add random white noise to the m original signals respectively to form a series of new signals; Perform EMD decomposition on this series of new signals to obtain a series of IMF components; The IMF components of the corresponding modes are averaged to obtain the EEMD decomposition results; The EEMD technology is used to determine whether the gear shift is completed. When the gear shift is completed, the motor is immediately controlled to stop, so that the manipulator and the gear shift handle stop. At this time, the gear shift operation is successful; The gear shift detection based on EEMD is divided into the following three steps: Calculate the absolute value of the real-time shift current Determine the peak value of the shift current Determine the first minimum value after the shift current peak.

8. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 1 is characterized in that: It also includes an impedance control system, which includes a control module, a body recognition perception and positioning module, a motor module, a data acquisition module, a shift simulator module and a power module. The control module includes a main controller IPC, a host computer and a DC motor driver; the data acquisition module includes a current acquisition card, a Hall sensor and a digital quantity acquisition card; the body recognition perception and positioning module is connected to the main controller IPC through a communication module; The surrounding recognition perception and positioning module is used to dynamically perceive the working environment of the tractor and perform visual obstacle avoidance and navigation through image recognition algorithms; The main controller IPC generates the expected position of the tractor operating lever and enters the trained impedance-controlled fuzzy PID control system program of the gear shift manipulator, then outputs the control result and transmits it to the DC motor driver; The DC motor driver receives instructions and controls the motor to rotate forward or reverse according to the level to realize forward or backward movement.

9. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 8 is characterized in that: The surrounding recognition perception and positioning module includes navigation, radar, ultrasound and camera; The camera is used to collect real-time surrounding video, and the main controller IPC determines whether to perform a gear shift operation based on the real-time surrounding video transmitted back by the camera; The gear shift simulator module sends data to the Alibaba Cloud; the main controller IPC receives the data sent by the gear shift simulator and sends a gear shift instruction; Navigation provides the unmanned tractor with high-precision positioning data and body posture information, and then predicts the next gear action based on the specific situation of the vehicle; The data acquisition module provides information feedback for the closed-loop control of the shift module; The gear shift simulator module uses the Logitech G29 gear shift simulator; The power module is a parallel connection of the tractor's own 12V battery and a 12V power supply; The motor module adopts LK-150 DC motor.

10. The fuzzy adaptive control system of the gear shifting manipulator based on impedance control according to claim 8 is characterized in that: The measurement and control system includes: The main control layer is composed of the main controller IPC, which has two modes: automatic driving and remote control driving; The main task of the decision-making layer is to analyze the real-time feedback data from various sensors on the robot. If it is an unmanned driving measurement and control system, the data is processed based on multi-sensor information fusion technology and fed back to the main control layer; according to the decision of the main control layer, the shift control command is issued to complete the corresponding action; if it is a remote-operated measurement and control system, the operator controls the simulator according to the data fed back by the host computer to achieve the corresponding forward and backward actions; The sensor layer is composed of sensors carried by the robot; the laser radar, ultrasonic ranging module and surrounding camera dynamically perceive the working environment of the tractor, and perform visual obstacle avoidance and navigation through image recognition algorithms; the combined navigation module provides high-precision positioning data and body posture information for the unmanned tractor, and then predicts the next gear action according to the specific situation of the vehicle; the Hall position sensor and current sensor provide information feedback for the closed-loop control of the shift module; The bottom-level actuator is composed of a shift control module, a motor driver, and a shift control motor. This layer mainly completes the specific shifting action of the shift robot.

Images