A mobile robot shared control method and system based on SSVEP

By using the SSVEP-based shared control method for mobile robots, and combining stimulation devices and EEG recognition technology with TEB-based human-robot conflict resolution and multi-factor weight allocation, the problem of limited number and low frequency of EEG control commands is solved. This enables the robot to move freely and drive safely in a two-dimensional plane, enhancing the user's sense of control.

CN118466345BActive Publication Date: 2024-11-22SOUTH CHINA UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410696341.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-31
Publication Date
2024-11-22
Estimated Expiration
2044-05-31

AI Technical Summary

Technical Problem

Existing SSVEP-based EEG control methods for robot control suffer from limited number and low frequency of EEG control commands, as well as human-machine conflict issues, making it difficult to achieve free movement and safe driving of robots in a two-dimensional plane.

Method used

A shared control method for mobile robots based on SSVEP is adopted. SSVEP EEG signals in four directions are induced by a stimulation device. Combined with EEG recognition, state assessment and environmental perception, the human-robot conflict resolution method of TEB and multi-factor control weight allocation are used to achieve the integrated control of EEG and intelligent planning.

Benefits of technology

It enables robots to move freely and safely in a two-dimensional plane, improves the flexibility and reliability of brain-computer interface control, solves the human-machine conflict problem, and enhances the user's sense of control over the robot.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118466345B_ABST
    Figure CN118466345B_ABST
Patent Text Reader

Abstract

The present invention discloses a mobile robot shared control method and system based on SSVEP. The EEG control command obtained by the EEG control command generation method can make the robot move in any direction in a two-dimensional plane. At the same time, the human-machine conflict resolution method based on TEB is adopted to solve the human-machine conflict problem, so that the intelligent planning control command follows the EEG control command, and the user's sense of control over the robot is improved. In addition, the control weight allocation mechanism is deeply studied to ensure that it can better adapt to different application scenarios and user needs. The present invention fully combines the advantages of robots and EEG, improves the intelligent operation level of robots, and further promotes the implementation and application of brain-controlled mobile robots, which is of great significance to the development of this field.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of brain-controlled wheeled robots, and in particular to a SSVEP-based mobile robot shared control method and system. Background Art

[0002] At present, the SSVEP-based EEG signal has a high recognition accuracy, and many studies have proved that the solution of applying SSVEP to online robot control is feasible. However, in the actual process of robot control, there are still the following problems that need to be solved.

[0003] First, due to the limitation that SSVEP can only recognize signals generated by specific frequencies, the number of EEG control instructions is limited. There are also some research results that increase the number of EEG control instructions by using hierarchical instructions or multimodal brain-computer interfaces. However, these methods increase the number of EEG control instructions by a limited number, making it difficult to achieve free movement of the robot in a two-dimensional plane.

[0004] Secondly, the EEG control frequency is low. To reliably identify EEG signals, it is necessary to collect 2-3 seconds of EEG data. The shared control method can effectively solve the problem of low EEG control frequency and greatly improve the completion rate of tasks. How to reasonably allocate control weights and on what factors to allocate them is an important issue that has attracted much attention.

[0005] Finally, the human-machine conflict problem derived from the field of autonomous driving also exists in the field of EEG shared control. When the decisions of EEG and intelligent planning are basically the same, the final output control will continue to increase, so that the car can reach the destination faster; however, when EEG and intelligent planning make different decisions, the final planning result depends entirely on the control weight. If you want to solve the human-machine conflict problem, you need to conduct more in-depth research and exploration of shared control methods. Summary of the invention

[0006] The first purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and provide a mobile robot shared control method based on SSVEP. The EEG control instructions obtained by the EEG control instruction generation method can enable the robot to move in any direction in a two-dimensional plane. At the same time, the human-machine conflict resolution method based on TEB effectively solves the human-machine conflict problem, allowing the intelligent planning control instructions to follow the EEG control instructions, thereby improving the user's sense of control over the robot. In addition, the control weight allocation mechanism has been studied in depth to ensure that it can better adapt to different application scenarios and user needs.

[0007] The second object of the present invention is to provide a mobile robot sharing control system based on SSVEP.

[0008] The first object of the present invention is achieved by the following technical solution: A mobile robot shared control method based on SSVEP, which performs the following steps in each control cycle of the robot:

[0009] S1. The user looks at the stimulation device, which includes four small lights flashing at different frequencies, representing four directions, including forward, referred to as F, backward, referred to as B, left, referred to as L, and right, referred to as R. When the user looks at the stimulation device, the SSVEP original EEG signal is generated, which is collected in real time by the EEG device, and then the EEG recognition method is used to obtain the EEG recognition results of the four directions;

[0010] S2. Analyze the EEG recognition results, including using the EEG control command generation method to convert the EEG recognition results into the EEG control commands u required by the robot h , and use the EEG state evaluation method to obtain the EEG state evaluation value; at the same time, the robot obtains the information of the current environment and obstacles through sensors, and uses the human-machine conflict resolution method based on TEB to make the intelligent planning follow the EEG control instructions to obtain the intelligent planning control instructions u without human-machine conflict s , and using an environmental status assessment method to obtain an environmental status assessment value;

[0011] S3, using the multi-factor control weight allocation method, comprehensively considering the EEG state evaluation value and the environmental state evaluation value, obtain the current control weight, and then use the linear weighted fusion method to convert the EEG control command u h and intelligent planning control instructions s Use the current control weights to fuse and obtain the final control command u output to the robot c ;

[0012] S4, the robot follows the control instruction u c The user obtains the current information of the robot through the human-computer interaction interface, gets visual feedback, and then repeats the above steps to complete the operation of the next control cycle.

[0013] Further, in step S1, the EEG recognition method is specifically as follows:

[0014] First, preprocessing is performed, including 50Hz notch processing of the original EEG signal to remove power frequency interference, baseline drift, and use a bandpass filter to extract the effective components of the EEG signal; then, the FBCCA method is used to divide the preprocessed EEG signal into five frequency sub-bands for identification, and finally the correlation values ​​between the current signal and the four reference signals of different frequencies are calculated, which are p, F 、p B 、p L 、p R, as the EEG recognition result.

[0015] Further, in step S2, the method for generating the EEG control instruction is specifically as follows:

[0016] First, the four related values ​​are converted into velocity vectors in four directions using the following method:

[0017] v i =p i v imax

[0018] Where, i∈{F, B, L, R}, v imax is the maximum velocity in the i direction, p i is the correlation value in the i direction;

[0019] Then, the velocities in the four directions are vector-fused to parse the current EEG control instruction u h as follows:

[0020]

[0021] In the formula, v h is the expected linear velocity of EEG, w h is the expected angular velocity of EEG, t is the control time point, a1 and a2 are adjustment factors, θ h is the maximum expected angle of EEG, and its value range is It is used to increase the range of the expected angular velocity of EEG, thus enabling free movement in a two-dimensional plane.

[0022] Further, in step S2, the EEG state assessment method is specifically as follows:

[0023] First, calculate the EEG confidence based on the four correlation values:

[0024]

[0025] If the CS value is larger, it means that the EEG state is better, which means that the credibility of this recognition is higher and the user's attention is more focused, so the EEG state evaluation value is also larger. Therefore, the EEG state evaluation value λ1 is designed as:

[0026]

[0027] Where b1 and b2 are adjustment factors.

[0028] Further, in step S2, the human-machine conflict resolution method based on TEB is referred to as SCR-TEB method, which is as follows:

[0029] The SCR-TEB method describes path planning as a multi-objective optimization problem according to the constraints of the mobile robot:

[0030]

[0031] In the formula, x * is the optimal solution to the current problem, B is the combined state of the mobile robot’s position and time series, and f kt is the objective function corresponding to the kt-th constraint condition, including the objective function f of the speed constraint vel , the objective function f of the acceleration constraint acc , the objective function f of the kinematic constraints nh And the objective function f of the newly added conflict constraint h , σ kt is the corresponding weight of the objective function corresponding to the kt-th constraint;

[0032] The objective function f of the conflict constraint h , it occurs when human-machine conflict occurs, that is, the EEG control instruction u h and intelligent planning control instructions s When different, a penalty value can be given and the intelligent planning control instructions can be updated to keep the two consistent. When there is no human-machine conflict, the function f h The value of f remains zero, h The specific expression is as follows:

[0033] f h (B) = max(-w h w s ,0)

[0034] In the formula, w s is the expected angular velocity of intelligent planning, w h is the expected angular velocity of the EEG, and by judging w h 、w s The sign of f can tell whether a human-machine conflict occurs. When the angular velocities of the two are opposite and the product is less than 0, it means that a conflict occurs. h (B) = -w h w s ; If they are the same and the product is greater than 0, then f h (B) = 0; Finally, the G2O graph optimization method is used to optimize each position S of the robot. i And each control cycle ΔT i As a vertex, each objective function is used as an edge to construct a G2O graph for multi-objective optimization and obtain intelligent planning control instructions.

[0035] Further, in step S2, the environmental status evaluation value is specifically as follows:

[0036]

[0037] In the formula, λ2 represents the environmental status assessment value, s k is the current position of the robot, It's an obstacle. is the Euclidean distance between the robot and the obstacle, Represents the distance between the current robot and the nearest obstacle, δ safe is the safety distance, c1 and c2 are adjustment factors.

[0038] Further, in step S3, the multi-factor control weight allocation method is referred to as RAS method, which is as follows:

[0039] When the robot maintains a safe distance from the obstacle, the control weight is determined by the evaluation value of the EEG state; when the robot is close to the obstacle, the control weight needs to be reduced to ensure the robot's safe driving; therefore, the control weight λ is designed as follows:

[0040] λ=argmin(λ1,λ2)

[0041] Through the RAS method, the control weight can be adjusted in real time according to the current EEG state evaluation value and the environmental state evaluation value.

[0042] Further, in step S3, the linear weighted fusion method is specifically as follows:

[0043] The control weights are used to linearly weight the EEG control instructions and the intelligent planning control instructions, as described below:

[0044] u c =λu h +(1-λ)u s

[0045] In each control cycle, the linear fusion method will update the control instruction u according to the current control weight c .

[0046] The second object of the present invention is achieved by the following technical solution: a mobile robot shared control system based on SSVEP, used to implement the above-mentioned mobile robot shared control method based on SSVEP, which includes: a stimulation device, an electroencephalogram device, a host computer, a slave computer and a robot module, which interact with each other through socket communication, wherein:

[0047] The stimulation device is used to induce the subject to generate an electroencephalogram signal with a special pattern;

[0048] The EEG device is used to obtain the original SSVEP EEG signal and send it to the host computer;

[0049] The host computer is used to use an EEG recognition method on the original EEG data to obtain an EEG recognition result and send it to the slave computer;

[0050] The lower computer is used to parse the EEG instructions, obtain the EEG control instructions and the EEG state evaluation value, and at the same time obtain the intelligent planning control instructions and the environmental state evaluation value through the human-machine conflict resolution method based on TEB, and finally obtain the final control instructions of the robot through the multi-factor control weight distribution method and the linear weighted fusion method and send them to the robot module;

[0051] The robot module is used to obtain map information and sensor information, and control the robot chassis to move according to the final control instructions.

[0052] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0053] 1. The method for generating EEG control instructions adopted by the present invention uses the SSVEP EEG recognition results in four directions to generate EEG control instructions u in any direction. h , realizing the robot's free steering in a two-dimensional plane, greatly increasing the expected angular velocity w of the EEG h The value range of .

[0054] 2. The present invention adopts a multi-factor control weight adjustment method, which comprehensively considers the evaluation value of the EEG state and the evaluation value of the environmental state, dynamically and in real time adjusts the control weight, effectively solving the problem of low EEG signal control frequency, thereby ensuring the safe driving of the robot.

[0055] 3. The present invention adopts a human-machine conflict resolution method based on TEB. By adding a new conflict objective function, the method enables intelligent planning to follow the direction of EEG control instructions when a human-machine conflict occurs. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Flow chart of the method of the present invention.

[0057] Figure 2 Schematic diagram of the method for generating EEG control instructions.

[0058] Figure 3 Optimizing the hypergraph for the G2O graph of the SCR-TEB method.

[0059] Figure 4 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION

[0060] The present invention is further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0061] Example 1

[0062] This embodiment discloses a mobile robot shared control method based on SSVEP. During data collection, the user is required to keep looking at the same small light. After the experiment starts, the robot will start moving only after receiving the first "forward" command. Then the user selects the desired command from the four commands of "forward", "left", "right" and "stop". When the "stop" signal is received, the robot will stop immediately. Figure 1 As shown, the process of each control cycle of this method during operation is demonstrated.

[0063] S1. The user looks at the stimulation device, which includes four small lights flashing at different frequencies, representing four directions, including forward, referred to as F, backward, referred to as B, left, referred to as L, and right, referred to as R. When the user looks at the stimulation device, the SSVEP original EEG signal is generated, which is collected in real time by the EEG device, and then the EEG recognition method is used to obtain the EEG recognition results of the four directions;

[0064] S2. Analyze the EEG recognition results, including using the EEG control command generation method to convert the EEG recognition results into the EEG control commands u required by the robot h , and use the EEG state evaluation method to obtain the EEG state evaluation value; at the same time, the robot obtains the information of the current environment and obstacles through sensors, and uses the human-machine conflict resolution method based on TEB to make the intelligent planning follow the EEG control instructions to obtain the intelligent planning control instructions u without human-machine conflict s , and using an environmental status assessment method to obtain an environmental status assessment value;

[0065] S3, using the multi-factor control weight allocation method, comprehensively considering the EEG state evaluation value and the environmental state evaluation value, obtain the current control weight, and then use the linear weighted fusion method to convert the EEG control command u h and intelligent planning control instructions s Use the current control weights to fuse and obtain the final control command u output to the robot c ;

[0066] S4, the robot follows the control instruction u c The user obtains the current information of the robot through the human-computer interaction interface, gets visual feedback, and then repeats the above steps to complete the operation of the next control cycle.

[0067] Specifically, the EEG recognition method is as follows:

[0068] First, preprocessing is performed, including 50Hz notch processing of the original EEG signal to remove power frequency interference, baseline drift, and use of bandpass filters to extract the effective components of the EEG signal; then, the preprocessed EEG signal is divided into five frequency sub-bands using the FBCCA method for identification, and finally the correlation values ​​between the current signal and four reference signals of different frequencies are calculated, which are p, F 、p B 、p L 、p R , as the EEG recognition result.

[0069] Specifically, the method for generating the EEG control instruction is as follows:

[0070] First, the four related values ​​are converted into velocity vectors in four directions using the following method:

[0071] v i =p i v imax

[0072] Where, i∈{F, B, L, R}, v imax is the maximum velocity in the i direction, p i is the correlation value in the i direction;

[0073] Then, the velocities in the four directions are vector-fused to parse the current EEG control instruction u h as follows:

[0074]

[0075] In the formula, v h is the expected linear velocity of EEG, w h is the expected angular velocity of EEG, t is the control time point, a1 and a2 are adjustment factors, θ h is the maximum expected angle of EEG, and its value range is The range of the expected angular velocity of EEG is greatly increased, so free movement in a two-dimensional plane can be achieved. Figure 2 As shown, the method for generating EEG control instructions is demonstrated.

[0076] Specifically, the EEG state assessment method is as follows:

[0077] First, calculate the EEG confidence based on the four correlation values:

[0078]

[0079] If the CS value is larger, it means that the EEG state is better, which means that the credibility of this recognition is higher and the user's attention is more focused, so the EEG state evaluation value is also larger. Therefore, the EEG state evaluation value λ1 is designed as:

[0080]

[0081] Where b1 and b2 are adjustment factors.

[0082] Specifically, the human-machine conflict resolution method based on TEB is referred to as SCR-TEB method, which is as follows:

[0083] The SCR-TEB method describes path planning as a multi-objective optimization problem according to the constraints of the mobile robot:

[0084]

[0085] In the formula, x * is the optimal solution to the current problem, B is the combined state of the mobile robot’s position and time series, and f kt is the objective function corresponding to the kt-th constraint condition, including the objective function f of the speed constraint vel , the objective function f of the acceleration constraint acc , the objective function f of the kinematic constraints nh And the objective function f of the newly added conflict constraint h , σ kt is the corresponding weight of the objective function corresponding to the kt-th constraint;

[0086] The objective function f of the conflict constraint h , it occurs when human-machine conflict occurs, that is, the EEG control instruction u h and intelligent planning control instructions s When they are different, a certain penalty value can be given and the intelligent planning control instructions can be updated to keep the two consistent. When there is no human-machine conflict, the function f h The value of f remains zero, h The specific expression is as follows:

[0087] f h (B) = max(-w h w s ,0)

[0088] In the formula, w s is the expected angular velocity of intelligent planning, w h is the expected angular velocity of the EEG, and by judging w h 、w s The sign of f can tell whether a human-machine conflict occurs. When the angular velocities of the two are opposite and the product is less than 0, it means that a conflict occurs.h (B) = w h w s ; If they are the same and the product is greater than 0, then f h (B) = 0; Finally, the G2O graph optimization method is used to optimize each position S of the robot. i And each control cycle ΔT i As a vertex, each objective function is used as an edge to construct a G2O graph for multi-objective optimization to obtain intelligent planning control instructions. The hypergraph optimized by the G2O graph is as follows: Figure 3 shown.

[0089] Specifically, the environmental status assessment value is as follows:

[0090]

[0091] In the formula, λ2 represents the environmental status assessment value, s k is the current position of the robot, It's an obstacle. is the Euclidean distance between the robot and the obstacle, represents the distance between the current robot and the nearest obstacle, c1 and c2 are adjustment factors, δ safe It is the safety distance. It is necessary to set an appropriate safety distance. If the safety distance is set too large, it will greatly reduce the user's control. If it is set too small, the safety of the robot cannot be guaranteed. safe The value is taken as 5 times the radius D of the robot. a1, a2, b1, b2, c1, c2 are adjustment factors, and the values ​​set by the present invention are shown in Table 1 below:

[0092] Table 1 Reference values ​​of various parameters in the experiment

[0093]

[0094] Specifically, the multi-factor control weight allocation method is referred to as the RAS method, which is as follows:

[0095] When the robot maintains a safe distance from the obstacle, the control weight is determined by the evaluation value of the EEG state; when the robot is close to the obstacle, the control weight needs to be reduced to ensure the robot's safe driving; therefore, the control weight λ is designed as follows:

[0096] λ=argmin(λ1,λ2)

[0097] Through the RAS method, the control weight can be adjusted in real time according to the current EEG state evaluation value and the environmental state evaluation value.

[0098] Specifically, the linear weighted fusion method is as follows:

[0099] The control weights are used to linearly weight the EEG control instructions and the intelligent planning control instructions, as described below:

[0100] u c =λu h +(1-λ)u s

[0101] In each control cycle, the linear fusion method will update the control instruction u according to the current control weight c .

[0102] Example 2

[0103] This embodiment discloses a mobile robot sharing control system based on SSVEP, which is used to implement the mobile robot sharing control system method based on SSVEP described in Example 1. Figure 4 As shown, it includes: a stimulation device, an electroencephalogram device, a host computer, a slave computer and a robot module, which interact with each other through socket communication, wherein:

[0104] The stimulation device is used to induce the subject to generate EEG signals with special rules, and includes the following units: a stimulation control unit, which is used to control the stimulation operation process, maintain the stimulation operation state, and present paradigm stimulation and feedback functions; a communication unit, which is used to perform socket communication with the host computer.

[0105] The EEG device is used to obtain the original SSVEP EEG signal and send it to the host computer;

[0106] The host computer is used to use the EEG recognition method on the original EEG data to obtain the EEG recognition result and send it to the lower computer; it includes the following units: a data receiving unit, which obtains the SSVEP original EEG signal from the EEG device; an EEG recognition unit, which uses the EEG recognition method to preprocess, recognize and classify the original EEG signal to obtain the EEG recognition result.

[0107] The lower computer is used to parse EEG instructions, obtain EEG control instructions and EEG state evaluation values, and at the same time obtain intelligent planning control instructions and environmental state evaluation values ​​through a human-machine conflict resolution method based on TEB, and finally obtain the final control instructions of the robot through a multi-factor control weight allocation method and a linear weighted fusion method and send them to the robot module; it includes the following units: an EEG instruction parsing unit, which parses the EEG recognition results transmitted by the upper computer into the robot's EEG control instructions and EEG state evaluation values; an intelligent planning unit, which obtains intelligent planning control instructions and environmental state evaluation values ​​through a human-machine conflict resolution method based on TEB; a shared control unit, which obtains the final control instructions of the robot through a multi-factor control weight allocation method and a linear weighted fusion method and sends them to the robot module.

[0108] The robot module is used to obtain map information and sensor information, and control the robot chassis to move according to the final control instructions.

[0109] The above-described embodiments are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, all changes made according to the shape and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A mobile robot sharing control method based on SSVEP, characterized in that: The following steps are performed during each control cycle of the robot: S1. The user looks at the stimulation device, which includes four small lights flashing at different frequencies, representing four directions, including forward, referred to as F, backward, referred to as B, left, referred to as L, and right, referred to as R. When the user looks at the stimulation device, the SSVEP original EEG signal is generated, which is collected in real time by the EEG device, and then the EEG recognition method is used to obtain the EEG recognition results of the four directions; The EEG recognition method is specifically as follows: First, preprocessing is performed, including 50Hz notch processing of the original EEG signal to remove power frequency interference, baseline drift, and use a bandpass filter to extract the effective components of the EEG signal; then, the FBCCA method is used to divide the preprocessed EEG signal into five frequency sub-bands for identification, and finally the correlation values ​​between the current signal and the four reference signals of different frequencies are calculated, which are p, F 、p B 、p L 、p R , as the result of EEG recognition; S2. Analyze the EEG recognition results, including using the EEG control command generation method to convert the EEG recognition results into the EEG control commands u required by the robot h , and use the EEG state evaluation method to obtain the EEG state evaluation value; at the same time, the robot obtains the information of the current environment and obstacles through sensors, and uses the human-machine conflict resolution method based on TEB to make the intelligent planning follow the EEG control instructions to obtain the intelligent planning control instructions u without human-machine conflict s , and using an environmental status assessment method to obtain an environmental status assessment value; The EEG state assessment method is specifically as follows: First, calculate the EEG confidence based on the four correlation values: If the CS value is larger, it means that the EEG state is better, which means that the credibility of this recognition is higher and the user's attention is more focused, so the EEG state evaluation value is also larger. Therefore, the EEG state evaluation value λ1 is designed as: Where b1 and b2 are adjustment factors; The human-machine conflict resolution method based on TEB is referred to as SCR-TEB method, which is as follows: The SCR-TEB method describes path planning as a multi-objective optimization problem according to the constraints of the mobile robot: In the formula, x * is the optimal solution to the current problem, B is the combined state of the mobile robot’s position and time series, and f kt is the objective function corresponding to the kt-th constraint condition, including the objective function f of the speed constraint vel , the objective function f of the acceleration constraint acc , the objective function f of the kinematic constraints nh And the objective function f of the newly added conflict constraint h , σ kt is the corresponding weight of the objective function corresponding to the kt-th constraint; The objective function f of the conflict constraint h , it occurs when human-machine conflict occurs, that is, the EEG control instruction u h and intelligent planning control instructions s When different, a penalty value can be given and the intelligent planning control instructions can be updated to keep the two consistent. When there is no human-machine conflict, the function f h The value of f remains zero, h The specific expression is as follows: f h (B)=max(-w h w s ,0) In the formula, w s is the expected angular velocity of intelligent planning, w h is the expected angular velocity of the EEG, and by judging w h 、w s The sign of f can tell whether a human-machine conflict occurs. When the angular velocities of the two are opposite and the product is less than 0, it means that a conflict occurs. h (B) = -w h w s ; If they are the same and the product is greater than 0, then f h (B) = 0; Finally, the G2O graph optimization method is used to optimize each position S of the robot. i And each control cycle ΔT i As a vertex, each objective function is used as an edge to construct a G2O graph for multi-objective optimization and obtain intelligent planning control instructions; The environmental status assessment values ​​are as follows: In the formula, λ2 represents the environmental status assessment value, s k is the current position of the robot, It's an obstacle. is the Euclidean distance between the robot and the obstacle, Represents the distance between the current robot and the nearest obstacle, δ safe is the safety distance, c1 and c2 are adjustment factors; S3, using the multi-factor control weight allocation method, comprehensively considering the EEG state evaluation value and the environmental state evaluation value, obtain the current control weight, and then use the linear weighted fusion method to convert the EEG control command u h and intelligent planning control instructions s Use the current control weights to fuse and obtain the final control command u output to the robot c ; The multi-factor control weight allocation method is referred to as RAS method, which is as follows: When the robot maintains a safe distance from the obstacle, the control weight is determined by the evaluation value of the EEG state; when the robot is close to the obstacle, the control weight needs to be reduced to ensure the robot's safe driving; therefore, the control weight λ is designed as follows: λ=argmin(λ1,λ2) Through this RAS method, the control weight can be adjusted in real time according to the current EEG state evaluation value and the environmental state evaluation value; S4, the robot follows the control instruction u c The user obtains the current information of the robot through the human-computer interaction interface, gets visual feedback, and then repeats the above steps to complete the operation of the next control cycle.

2. A mobile robot shared control method based on SSVEP according to claim 1, characterized in that: In step S2, the method for generating the EEG control instruction is as follows: First, the four related values ​​are converted into velocity vectors in four directions using the following method: v i =p i v imax Where, i∈{F, B, L, R}, v imax is the maximum velocity in the i direction, p i is the correlation value in the i direction; Then, the velocities in the four directions are vector-fused to parse the current EEG control instruction u h as follows: In the formula, v h is the expected linear velocity of EEG, w h is the expected angular velocity of EEG, t is the control time point, a1 and a2 are adjustment factors, θ h is the maximum expected angle of EEG, and its value range is It is used to increase the range of the expected angular velocity of EEG, thus enabling free movement in a two-dimensional plane.

3. A mobile robot shared control method based on SSVEP according to claim 2, characterized in that: In step S3, the linear weighted fusion method is specifically as follows: The control weights are used to linearly weight the EEG control instructions and the intelligent planning control instructions, as described below: in c =λu h +(1-λ)u s In each control cycle, the linear fusion method will update the control instruction u according to the current control weight c .

4. A mobile robot sharing control system based on SSVEP, characterized in that: The SSVEP-based shared control method for a mobile robot used to implement any one of claims 1 to 3 comprises: a stimulation device, an electroencephalogram device, a host computer, a slave computer and a robot module, which interact with each other by means of socket communication, wherein: The stimulation device is used to induce the subject to generate an electroencephalogram signal with a special pattern; The EEG device is used to obtain the original SSVEP EEG signal and send it to the host computer; The host computer is used to use an EEG recognition method on the original EEG data to obtain an EEG recognition result and send it to the slave computer; The lower computer is used to parse the EEG instructions, obtain the EEG control instructions and the EEG state evaluation value, and at the same time obtain the intelligent planning control instructions and the environmental state evaluation value through the human-machine conflict resolution method based on TEB, and finally obtain the final control instructions of the robot through the multi-factor control weight distribution method and the linear weighted fusion method and send them to the robot module; The robot module is used to obtain map information and sensor information, and control the robot chassis to move according to the final control instructions.