Remote intelligent vehicle system based on predefined time sliding mode variable structure control

Abstract

The invention discloses a remote intelligent vehicle system based on predefined time sliding mode variable structure control, an intelligent vehicle adopts two sets of controllers, namely a student end and a teacher end, the student end can remotely burn and run a program written by a learner to meet the learning requirement of the learner; the teacher end is used for storing a trolley position control program and executing position control according to a remote instruction; the position control comprises the following steps: S1, establishing a system model of the remote Mecanum wheel type intelligent trolley; s2, the upper computer issues an expected position or an expected track to the main control chip, each sensor collects data and obtains motion parameters of the trolley, and the main controller calculates an error between a target pose and an actual pose; s3, establishing an integral sliding mode observer, and estimating the interference of the system; s4, establishing a sliding mode surface; and S5, establishing a control law, calculating control input, and enabling the trolley to run according to a specified track.

Description

Technical Field

[0001] The present invention relates to the field of robot trajectory tracking, and in particular to a remote intelligent vehicle system based on predefined time sliding mode variable structure control. Background Art

[0002] In the field of electronic information, an intelligent vehicle is a relatively typical motion control experimental device, which has broad development prospects in many engineering and technical fields and combines fun and theoretical research. However, the learning of an intelligent vehicle often requires a large site and a special experimental map, which are difficult to meet the requirements in general home and classroom environments. At the same time, learning an intelligent vehicle requires purchasing a series of accessories for learning, resulting in a relatively high learning cost. Therefore, a remote intelligent vehicle platform has emerged. It can burn the program written by the learner into the intelligent vehicle through remote online downloading, and at the same time provide a camera and an interaction interface, enabling the learner to observe the running status and internal data of the vehicle. The learner only needs a computer and an Ethernet, and can conduct research on the intelligent vehicle anywhere.

[0003] Since the position of the vehicle will change during the use of this platform, the learner needs to reset it after each use for the next operation. In addition, due to different experimental requirements, the learner may need to place the vehicle in different positions or run it along a specified trajectory. However, this platform is used in a remote connection mode, and the learner cannot manually place the vehicle position. Therefore, a method is needed to automatically control the position and movement trajectory of the vehicle.

[0004] In addition, considering that the main controller of the vehicle needs to be open to the learner for writing and running the learner's own program, this will cause the position control program to be overwritten. If the method of rewriting is adopted, the debugging efficiency will be greatly reduced. Therefore, how to schedule the learner's program and the position control program is a problem that needs to be solved.

[0005] Therefore, in view of the defects of the prior art, it is indeed necessary to propose a technical solution to solve the technical problems existing in the prior art. Summary of the Invention

[0006] In view of this, it is indeed necessary to provide a remote intelligent vehicle system based on predefined time sliding mode variable structure control. The learner can remotely log in to the experimental platform and control the intelligent vehicle through programming. The intelligent vehicle adopts two sets of controllers. Among them, the student side can remotely burn through the network and is used to store and run the program written by the learner; the teacher side is used to store the relevant programs for vehicle position control, and an integral sliding mode observer is introduced and a new type of predefined time sliding mode controller is constructed, so that the intelligent vehicle can quickly converge within a predefined time, avoiding control output saturation while improving control accuracy.

[0007] To solve the technical problems existing in the prior art, the technical solution of the present invention is as follows:

[0008] A remote intelligent vehicle system based on predefined time sliding mode variable structure control, in which learners can remotely log in to the local experimental terminal to operate the intelligent vehicle platform. Among them, the intelligent vehicle adopts two sets of controllers, namely the student end and the teacher end. Among them, the student end can remotely burn and run the program written by the learner to meet the learning needs of the learner; the teacher end is used to store the vehicle position control program and execute position control according to remote instructions;

[0009] The position control includes the following steps:

[0010] Step S1, establish a system model of the remote Mecanum wheeled intelligent vehicle;

[0011] Step S2, the host computer issues the desired position or desired trajectory to the main control chip, and each sensor collects data to obtain the motion parameters of the vehicle. The main controller calculates the error between the target pose and the actual pose;

[0012] Step S3, establish an integral sliding mode observer to estimate the interference of the system;

[0013] Step S4, establish a sliding mode surface;

[0014] Step S5, establish a control law, calculate the control input, and make the vehicle run according to the specified trajectory.

[0015] As a further improvement scheme, in step S1, the kinematic model of the intelligent vehicle system is expressed as:

[0016]

[0017] where δ = [δ1, δ2, δ3, δ4] T is the lumped interference of the system, ζ = [x, y, ψ] T is the actual pose of the vehicle body, are its first and second derivatives respectively, u = [u1, u2, u3, u4] T is the control input of the system, and A and g(ψ) are related matrices, which are specifically expressed as:

[0018]

[0019] where j0 is the nominal equivalent inertia moment of the wheel, b0 is the nominal viscous friction force, K t0 is the nominal motor constant, r is the tire radius of the Mecanum wheel, and a and b represent the projections of the line from the geometric center of the wheel to the geometric center of the robot on the local coordinate axes X b and Y b respectively,

[0020] As a further improvement, in step S2, the pose error of the system is expressed as:

[0021] e = [e1, e2, e3] T = ζ r - ζ

[0022] where e represents the pose error of the system, and ζ r = [x r , y r , ψ r T is the desired pose, input by the learner through the host computer, and ζ = [x, y, ψ] T is obtained through the intelligent vehicle sensor.

[0023] As a further improvement, in step S3, an integral sliding mode observer is designed:

[0024]

[0025] where Z = [Z1, Z2, Z3] T is the state variable of the observer, is the estimated value of the lumped disturbance, and each of its components is:

[0026]

[0027] k1 > 0, k2 > η are positive coefficients, is the exponential term; where, is the auxiliary variable; sign(·) is the sign function; for any real number x, a, the function sig(x) a = |x| a sign(x), and for the matrix then

[0028] As a further improvement, in step S4, a predefined-time sliding mode surface is designed as:

[0029]

[0030] where 0 < ρ1 < 0.5, T c1 > 0 is the predefined time for the sliding mode surface to converge to 0, and α s is a constant.

[0031] As a further improvement, in step S5, the control law of the sliding mode is designed as:

[0032] u = u​eq +u r

[0033] Among them, assuming the perturbation δ i = 0, let The equivalent control input can be obtained as follows:

[0034]

[0035] Among them, g(ψ) + is the generalized inverse matrix of g(ψ), and its expression is:

[0036]

[0037]

[0038] Moreover, M1 and M2 are expressed as:

[0039]

[0040] An integral sliding mode observer is introduced, and the control input is designed as:

[0041]

[0042] Among them, 0 < ρ2 < 1, T c2 > 0 is the attitude error after the sliding surface converges, and the perturbation error

[0043] As a further improvement scheme, the encoder and motor on the intelligent trolley are shared by the teacher side and the student side; each motor on the intelligent trolley has two data channels, one connected to the teacher side and one connected to the student side; the switches of the two data channels are controlled by a group of analog switches, and the teacher side selects the data channel through the analog switch.

[0044] As a further improvement scheme, during the use of the intelligent trolley platform, the teacher side is always in operation, and controls the analog switch to close the data channel of the teacher side connected to the motor and open the student side channel. At this time, the control right of the trolley movement is located at the student side, and the learner controls the trolley movement by writing his own program; when the upper computer issues the target position, the teacher side closes the student side data channel and opens the teacher side channel. At this time, the control right of the trolley movement is located at the teacher side until the trolley moves to the specified position, and then the data channel is switched back to the student side, so as to realize the control of the position without affecting the student side program.

[0045] Compared with the prior art, the technical solution of the present invention has the following technical effects:

[0046] 1. The automatic control of the trolley is realized, and the control process is optimized. The user only needs to provide the target position or path, and the trolley can automatically obtain its own pose information and achieve trajectory tracking.

[0047] 2. Compared with the traditional sliding mode control algorithm, the present invention introduces an integral sliding mode observer to estimate and compensate for disturbances, effectively suppressing the chattering problem of the sliding mode.

[0048] 3. The present invention constructs a new predefined-time sliding mode controller, enabling the trolley to reach the specified position within a predefined time. Compared with the existing control methods, this method can provide better control accuracy and faster convergence speed. Description of the Drawings

[0049] Figure 1 It is the system architecture block diagram of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention;

[0050] Figure 2 It is the schematic diagram of the platform physical object in a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention;

[0051] Figure 3 It is the schematic diagram of the intelligent trolley physical object in a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention;

[0052] Figure 4 It is the control flow chart of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention;

[0053] Figure 5 It is the control structure schematic diagram of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention;

[0054] Figure 6 It is the schematic diagram of the motion model of the intelligent trolley in a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention;

[0055] Figure 7 It is the trajectory tracking effect diagram of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention at different starting poses;

[0056] Figure 8 It is the pose error diagram of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention at different starting poses;

[0057] Figure 9 It is the trajectory tracking effect diagram of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention by different control methods;

[0058] Figure 10 This is the pose error graph of a remote intelligent vehicle system based on predefined time sliding mode variable structure control in an embodiment of the present invention through different control methods.

[0059] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. Specific Embodiments

[0060] The following will further illustrate the technical solutions provided by the present invention in conjunction with the drawings.

[0061] Participate Figure 1 , as shown is the structural block diagram of a remote intelligent vehicle system based on predefined time sliding mode variable structure control provided by the present invention, including:

[0062] A local experimental terminal, used for learners to remotely log in and be able to program and control the intelligent vehicle platform through locally written software. Among them, it at least includes a host computer, and the host computer transmits data with the intelligent vehicle platform in a wired manner;

[0063] The intelligent vehicle platform includes: a power supply, used to supply power to the system; a monitoring camera, installed above the platform, used to monitor the running status of the intelligent vehicle and display the image on the host computer in real time; a downloader, used to burn the program of the intelligent vehicle according to the control instructions of the host computer; an intelligent vehicle, used to control the movement of the vehicle according to the program;

[0064] The intelligent vehicle adopts a Mecanum wheel type intelligent vehicle, that is, the remote vehicle platform selects a Mecanum wheel type omnidirectional vehicle model. Due to its unique wheel hub structure, the Mecanum wheel type vehicle can achieve movement modes such as forward movement, lateral movement, diagonal movement, rotation and their combinations, which makes it extremely flexible in narrow or complex environments. The motion control of the Mecanum wheel vehicle is relatively simple and easy to implement.

[0065] Among them, the intelligent vehicle adopts two sets of controllers, namely the student end and the teacher end. Among them, the student end is connected to the downloader, and can remotely burn and run the programs written by learners, and set a variety of peripheral resources to meet the different learning needs of learners; the teacher end is used to store the relevant programs for vehicle position control to perform position control.

[0066] See Figure 2 and Figure 3, as shown in the figure, is a physical connection diagram of the control system of the remote intelligent vehicle platform of the present invention. Among them, the upper computer uses the LabVIEW upper computer and communicates with the intelligent vehicle through the serial port module; the power supply, the downloader, and the serial port module are connected to the intelligent vehicle through a bus. Learners can remotely access the upper computer through the network. The upper computer displays the image of the platform in real time and integrates functions such as program downloading, serial port data transmission, and position control, realizing the remote use and learning of the platform by learners.

[0067] As a further improvement scheme, peripheral devices such as infrared sensors, temperature and humidity sensors, and cameras are installed on the student side to meet the needs of different learners.

[0068] As a further improvement scheme, the teacher side is implemented using a Raspberry Pi. Programs related to the position control of the vehicle are pre-stored. To implement the position control function, a Raspberry Pi, a radar, and a nine-axis acceleration sensor are installed. After the upper computer sends the target position to the teacher side through the serial port, the ROS system is installed and run on the Raspberry Pi. The current position is obtained through the radar and the nine-axis acceleration sensor, and the slam mapping and positioning are used to calculate the current position and attitude information of the vehicle. Then, the sliding mode control algorithm is run to calculate the magnitude of the control input and transmit it to the teacher side through the serial port. The teacher side then converts the control input into a PWM signal and controls each motor through the motor drive, thereby realizing position control.

[0069] As a further improvement scheme, the encoder and motor on the vehicle are shared by the teacher side and the student side. Each motor on the vehicle has two data channels, one connected to the teacher side and one connected to the student side. The switches of the two data channels are controlled by a group of analog switches, and the teacher side can select the data channels through the analog switches. During the use of the platform, the teacher side is always in the running state and controls the analog switches to close the data channel of the motor on the teacher side and open the student side channel. At this time, the control right of the vehicle movement is located on the student side, and learners can control the vehicle movement with their own programs. When the upper computer issues the target position, the teacher side closes the student side data channel and opens the teacher side channel. At this time, the control right of the vehicle movement is located on the teacher side until the vehicle moves to the specified position, and then the data channel is switched back to the student side, thereby realizing the control of the position without affecting the student side program.

[0070] See Figure 4 , as shown in the figure, is a control flow block diagram of a remote intelligent vehicle system based on predefined time sliding mode variable structure control provided by the present invention. See Figure 5, shown is a schematic diagram of the control structure of a remote intelligent vehicle system based on predefined-time sliding mode variable structure control according to an embodiment of the present invention. First, an accurate model of the Mecanum wheeled intelligent vehicle is established; an integral sliding mode observer (ISMO) is used to compensate for the lumped disturbance and potential error of the system caused by control input saturation, so that the observation error converges to a known boundary. A predefined-time sliding mode surface is constructed, and a predefined-time control law is designed in combination with the observer, so that the vehicle can reduce chattering while quickly reaching the specified position. The specific steps are as follows:

[0071] Step S1, establish a system model of the remote Mecanum wheeled intelligent vehicle;

[0072] Step S2, the host computer sends the desired position or desired trajectory to the main control chip, each sensor collects data, obtains the motion parameters of the vehicle, and the main controller calculates the error between the target pose and the actual pose;

[0073] Step S3, establish an integral sliding mode observer to estimate the disturbance of the system;

[0074] Step S4, establish a sliding mode surface;

[0075] Step S5, establish a control law, calculate the control input, and make the vehicle run according to the specified trajectory.

[0076] Among them, in step S1, the intelligent vehicle adopts a Mecanum wheeled intelligent vehicle. The Mecanum wheeled robot system is a typical multi-input multi-output nonlinear system. Sliding mode control is considered to be one of the most effective methods for dealing with uncertain nonlinear systems, with fast response, high precision and strong robustness. In actual use, it is very important to make the vehicle reach the desired state within a limited control time. However, the traditional terminal sliding mode control is affected by the initial state and cannot estimate the exact upper limit of the convergence time. Although the predefined-time sliding mode method can estimate the convergence time, due to different initial states, the predefined-time control method will generate too large a control input at the beginning, resulting in actuator saturation and affecting the control process. The common phenomenon of control input saturation not only reduces the dynamic performance of the system, but also interferes with the stability and control accuracy of the closed-loop system, making it one of the key problems affecting the system stability.

[0077] The specific process of the dynamic model equation for the Mecanum wheeled intelligent vehicle is as follows:

[0078] Figure 6 This is a schematic diagram of the motion model of the intelligent vehicle in the method of the present invention. The motion coordinates of the MWOMR shown in the figure take the center of the rectangular frame of the vehicle as the geometric center, where OXY represents the global coordinate system (GCF), O b X b Y bDenotes the body coordinate system (BCF). Here, ζ = [x, t, ψ] T , ζ b = [x b , y b , ψ b T Are respectively defined as the poses of the MWOMR in the GCF and BCF. Is the speed of the cart in the BCF. ω = [ω1, ω2, ω3, ω4] T Respectively represent the angular velocities of the four wheels. According to the inverse kinematics formula of the MWOMR, the relationship between the speed in the BCF and the angular velocities of the wheels is:

[0079]

[0080] Among them, r is the tire radius of the Mecanum wheel, and a and b represent the projection lengths of the line from the geometric center of the wheel to the geometric center of the robot on the local coordinate axes X b and Y b .

[0081] By converting between the GCF and BCF, we can obtain:

[0082]

[0083] Therefore, it is easy to obtain the relationship between the speed of the cart in the GCF and the wheel speeds as follows:

[0084]

[0085] h(ψ) is the transformation matrix, and h(ψ) + Is its generalized inverse matrix, and their expressions are respectively:

[0086]

[0087] Among them, Furthermore, the classical model of the motor is as follows:

[0088]

[0089] In the formula, j0 is the nominal equivalent inertia of the wheel, b0 is the nominal viscous friction, d = [d1, d2, d3, d4] is the disturbance on the four wheels, and K t0 Is the nominal motor constant. v = [v1, v2, v3, v4] T Is the input voltage of the four corresponding motors.

[0090] Taking the derivative of formula (3) can obtain:

[0091] ​

[0092] where H is expressed as:

[0093]

[0094] Combining formulas (4), (7), and (8), the model equation can be rewritten as:

[0095]

[0096]

[0097] where A and g(ψ) are expressed as:

[0098]

[0099] Considering the input saturation phenomenon of the MWOMR system, the motor voltage has both upper and lower limits. Define the upper bound of the input voltage as v max , and the saturation function sat(u i ) is expressed as follows:

[0100]

[0101] Introducing input saturation, the final kinematic model is expressed as follows:

[0102]

[0103] where

[0104]

[0105] In step S2, to better design the control input, first determine the tracking error, ζ r = [x r , y r , ψ r is the desired pose, input by the learner through the host computer, and ζ = [x, y, ψ] T is the actual pose of the vehicle body, obtained by the vehicle through sensors. e represents the pose error of the system, and the expression is:

[0106] e = [e1, e2, e3] T = ζ r - ζ (16)

[0107] In step S3, to address the adverse effects caused by disturbances due to external factors and parameter uncertainties, and to compensate for the input error caused by control input saturation, an integral sliding mode observer is designed.

[0108] Assume that the disturbance δ is bounded, and there exists a positive number for η such that Define the state variable of the observer as Z and introduce an auxiliary variable Define the initial value of Z According to the dynamic model of the system, the state equation of the observer can be designed as follows:

[0109]

[0110] where is the estimated value of the lumped disturbance, and each component is:

[0111]

[0112] κ1>0, κ2>η are positive coefficients, is an exponential term. Where, sign(·) is the sign function. For any real numbers x, a, the function sig(x) a = |x| a sign(x), and for the matrix then

[0113] Select the following Lyapunov function:

[0114] V0 = V 01 + V 02 + V 03 (19)

[0115] Each component Then its derivative with respect to time is:

[0116]

[0117] where Therefore, the system can converge to zero, and the time to reach the origin for the first time is:

[0118]

[0119] The convergence speed and accuracy of the observer are mainly determined by κ1>0, κ2>η. Appropriately increasing κ1 can accelerate the convergence speed of the observer and reduce the steady-state error. However, too large κ1 may cause chattering in the disturbance estimation. Increasing κ2 can also accelerate the convergence speed but will increase the steady-state error. Therefore, in practical applications, it is crucial to balance the convergence speed, steady-state error, and system stability and select appropriate parameter values.

[0120] In step S4, in order to achieve the convergence of the system error within a predetermined time, according to the predefined time principle, design the predefined time sliding surface as:

[0121]

[0122] where \(0 < \rho_1 < 0.5\), \(T\) c1 \(> 0\) is the predefined time for the sliding surface to converge to \(0\), and \(\alpha\) s is a constant.

[0123] In step S5, assuming the disturbance \(\delta\) i \(= 0\), let the equivalent control input can be obtained:

[0124]

[0125] where \(g(\psi)\) + is the generalized inverse matrix of \(g(\psi)\), and its expression is:

[0126]

[0127] Moreover, \(M_1\) and \(M_2\) can be expressed as:

[0128]

[0129] Furthermore, an integral sliding mode observer is introduced, and the reaching control input is designed as:

[0130]

[0131] where \(0 < \rho_2 < 1\), \(T\) c2 \(> 0\) is the attitude error after the sliding surface converges, and the disturbance error

[0132] Then the total control input is:

[0133] \(u = u\) eq \(+ u\) r (28)

[0134] To prove that the designed controller can converge within the predefined time, two Lyapunov equations need to be designed:

[0135]

[0136] Taking the derivatives of each of them respectively, we can get:

[0137]

[0138]

[0139] According to the theorem related to the predefined time, it can be proved that the controller can achieve convergence, and its convergence time \(T\) c \(< T\) c1 \(+ T\) c2 .

[0140] The parameters in the controller significantly affect the tracking performance. The corresponding selection criteria for the control parameters are as follows:

[0141] The predefined time is defined by T c1 and T c2 which is the most critical index of the controller. The smaller the T c1 and T c2 the faster the convergence time. However, the actual settlement time is much less than T c1 +T c2 . If the values of T c1 and T c2 are too small, it will lead to too large control input.

[0142] The exponential terms ρ1 and ρ2 are responsible for the convergence speed of the position error and the convergence speed of the sliding surface respectively. To select appropriate values, the trade-off between the convergence speed and the control input should be considered, because under the condition of a large initial value, a smaller value can achieve a faster convergence speed, but it will also generate a larger control input, which may lead to control input saturation.

[0143] To control the motor movement, the control input needs to be converted into a PWM wave output:

[0144]

[0145] where PWM i is the output of the i-th PWM port, PWM max is the upper limit of the PWM output duty cycle, and V max is the maximum voltage of the motor output.

[0146] The effectiveness of the control method of the present invention is verified through experiments as follows:

[0147] 1) Predefined time convergence test

[0148] Figure 7 The figure shows the motion trajectories of the intelligent vehicle moving from the stationary state to the target trajectory under three different initial conditions. Figure 8 The figure shows the tracking errors of the vehicle's x, y, and ψ axes. Among them, the initial states are respectively The target trajectory is ζ r =[-0.5cos(0.5t), 0.5sin(0.5t), 0] T .

[0149] The experimental results show that despite different initial conditions, the tracking errors of the vehicle can converge within about four seconds, which is less than the set predefined time of 6 seconds. This demonstrates the predefined time characteristic of the control algorithm.

[0150] 2) Performance Comparison of Different Control Methods

[0151] Figure 9 As shown, the initial pose of the trolley is The target trajectory is ζ r =[-0.5cos(0.25t), 0.5sin(0.5t), 0] T , and the motion trajectories of moving from the stationary state to the target trajectory with different control methods. Figure 10 As shown are the tracking errors of the trolley in the x, y, and ψ axes. Among them, the control algorithms used are: Sliding Mode Control (SMC), Terminal Sliding Mode Control (TSMC), Traditional Predefined-Time Sliding Mode Control (Traditional PTSMC), and the Sliding Mode Control described in this paper (Proposed PTSMC).

[0152] The experimental results show that compared with other control algorithms, the proposed control algorithm has a faster convergence speed. To further verify the performance of the proposed control method, evaluation indexes of Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) are used for comparison, and their expressions are respectively:

[0153]

[0154] The specific test data are shown in Table 1. It can be seen that the control accuracy of this method is higher.

[0155] Table 1 Test Data

[0156]

[0157]

[0158] The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of the present invention, several improvements and modifications can still be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

[0159] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A remote intelligent vehicle system based on predefined time sliding mode variable structure control, characterized in that, Learners operate the intelligent vehicle platform by remotely logging in to the local experimental terminal. Among them, the intelligent vehicle adopts two sets of controllers, namely the student end and the teacher end. Among them, the student end can remotely burn and run the programs written by learners to meet the learning needs of learners; the teacher end is used to store the vehicle position control program and execute position control according to remote instructions. The position control includes the following steps: Step S1, establish a system model of a remote Mecanum wheeled intelligent vehicle; Step S2, the host computer sends the desired position or desired trajectory to the main control chip, each sensor collects data, obtains the motion parameters of the vehicle, and the main controller calculates the error between the target pose and the actual pose; Step S3, establish an integral sliding mode observer to estimate the interference of the system; Step S4, establish a sliding mode surface; Step S5, establish a control law, calculate the control input, and make the vehicle run according to the specified trajectory.

2. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that In step S1, the kinematic model of the intelligent vehicle system is expressed as: where δ = [δ1, δ2, δ3, δ4] T is the lumped interference of the system, ζ = [x, y, ψ] T is the actual pose of the vehicle body, which are its first and second derivatives respectively, u = [u1, u2, u3, u4] T is the control input of the system, A and g(ψ) are related matrices, specifically expressed as: Among them, j0 is the nominal equivalent inertia moment of the wheel, b0 is the nominal viscous friction force, K t0 is the nominal motor constant, r is the tire radius of the Mecanum wheel, and a and b represent the projection lengths of the lines from the geometric center of the wheel to the geometric center of the robot on the local coordinate axes X b and X b respectively.

3. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that, In step S2, the pose error of the system is expressed as: e = [e1, e2, e3] T = ζ r -ζ where e represents the pose error of the system, and ζ r = [x r , y r , ψ r T is the desired pose, input by the learner through the host computer, and ζ = [x, y, ψ] T is obtained through the sensors of the intelligent vehicle.​ 4. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that, In step S3, design an integral sliding mode observer: where \(Z = [Z_1, Z_2, Z_3]\) T is the state variable of the observer, is the estimated value of the lumped disturbance, and each component thereof is: κ1 > 0, κ2 > η are positive coefficients, is an exponential term; where, is an auxiliary variable; sign(·) is a sign function; for any real numbers x, a, the function sig(x) a = |x| a sign(x), and for a matrix then 5. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that, In step S4, design a predefined time sliding mode surface as: where \(0 < \rho_1 < 0.5\), T c1 > 0 is the predefined time for the sliding surface to converge to 0, and \(\alpha\) s is a constant.

6. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that, In step S5, design the control law of the sliding mode as: u = u eq + u r Among them, assume the perturbation δ i = 0, let The equivalent control input can be obtained as follows: where \(g(\psi)\) + is the generalized inverse matrix of \(g(\psi)\), and its expression is: And, M1 and M2 are expressed as: Introduce an integral sliding mode observer, and the control input is designed as: where \(0 < \rho_2 < 1\), T c2 > 0 is the attitude error after the sliding mode surface converges, and the disturbance error 7. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that The teacher end and the student end share the encoder and motor on the intelligent vehicle; each motor on the intelligent vehicle has two data channels, one connected to the teacher end and one connected to the student end; the switches of the two data channels are controlled by a group of analog switches, and the teacher end selects the data channel through the analog switch.

8. The remote intelligent vehicle system based on predefined time sliding mode variable structure control according to claim 1, characterized in that, During the use of the intelligent vehicle platform, the teacher end is always in operation, and controls the analog switch to close the data channel of the motor of the teacher end and open the student end channel. At this time, the control right of the vehicle movement is at the student end, and the learner controls the vehicle movement by writing his own program; when the host computer sends the target position, the teacher end closes the student end data channel and opens the teacher end channel. At this time, the control right of the vehicle movement is at the teacher end until the vehicle moves to the specified position, and then the data channel is switched back to the student end, so as to realize the position control without affecting the student end program.

Images