An anti-disturbance control system for an omnidirectional mobile robot

Abstract

The present invention relates to the field of robot control technology, and specifically discloses an anti-disturbance control system for an omnidirectional mobile robot, comprising: a disturbance term extraction module, used for obtaining all disturbance term information of the omnidirectional mobile robot in a pre-period; a space equation generation module, used for generating a pre-disturbance expanded state space equation of the omnidirectional mobile robot; a control quantity correction module, used for performing disturbance correction on a target control quantity of the omnidirectional mobile robot in the pre-period to obtain an anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period; an anti-disturbance control module, used for treating the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period as an input control quantity of the omnidirectional mobile robot in the pre-period to obtain an anti-disturbance control result of the omnidirectional mobile robot; and used for performing high-precision disturbance correction on the target control quantity of the omnidirectional mobile robot in the pre-period to achieve high-precision anti-disturbance control of the omnidirectional mobile robot.

Description

Technical Field

[0001] The present invention relates to the technical field of robot control, and in particular to an anti-disturbance control system for an omnidirectional mobile robot. Background Art

[0002] At present, in recent years, with the rapid development of computer and microelectronics manufacturing technology, robotics has become an interdisciplinary and multi-integrated technical field. An omnidirectional robot is a robot that can move autonomously in all directions and is usually used for navigation and positioning. The omnidirectional mobile robot is a multi-input and multi-output nonlinear system. During the trajectory tracking control process, the robot's posture and motion state are affected by wheel slippage. The self-disturbance rejection controller has many parameters and is coupled. Improper parameter adjustment often affects the robustness of the controller. In disturbance rejection control, it is necessary to identify and eliminate common disturbances that may affect the stability and accuracy of the omnidirectional robot. The following are some common disturbances:

[0003] 1. Friction disturbance: When an omnidirectional robot moves, it is affected by the friction of the ground or other surfaces, which may cause speed changes, position deviation, etc.

[0004] 2. Inertial disturbance: The omnidirectional robot is affected by its own inertia, such as acceleration, deceleration, and turning, which may cause position deviation and posture distortion.

[0005] 3. Electromagnetic interference: Electronic equipment and sensors around the omnidirectional robot may be affected by electromagnetic interference, resulting in data errors and control command failure.

[0006] 4. Mechanical vibration and shock: The omnidirectional robot is subject to external mechanical vibration and shock, such as ground vibration, collision, etc., which may cause position deviation, speed change, etc.

[0007] 5. Environmental factors: The omnidirectional robot is affected by environmental factors such as temperature, humidity, air pressure, etc., which may cause the robot's performance to deteriorate or malfunction.

[0008] These disturbances may act on the omnidirectional robot individually or simultaneously. Therefore, these possible interference factors need to be considered during design and control, and corresponding measures need to be taken to reduce or eliminate their impact on the performance of the omnidirectional robot.

[0009] Anti-disturbance control is an important research direction in the field of robotics, especially in the application of omnidirectional mobile robots. The goal of anti-disturbance control is to reduce the impact of external disturbances on the robot while ensuring its stability, thereby improving the robot's positioning accuracy, trajectory tracking ability and robustness. However, traditional robot anti-disturbance control methods: This type of method mainly achieves anti-disturbance control by designing a robot system structure with good performance or using traditional control algorithms. Commonly used control algorithms include PID control, model predictive control (MPC), adaptive control, etc. However, when faced with complex actual environments and multiple disturbance sources, these methods often require a large amount of computing resources and time, and may have problems such as slow convergence and difficulty in parameter adjustment.

[0010] Therefore, the present invention proposes an anti-disturbance control system for an omnidirectional mobile robot. Summary of the Invention

[0011] The present invention provides an omnidirectional mobile robot anti-disturbance control system, which is used to perform high-precision disturbance correction on the target control quantity of the omnidirectional mobile robot within a pre-period based on all disturbance item information of the omnidirectional mobile robot within a pre-period and the analysis principle of the extended state observer, so as to realize high-precision anti-disturbance control of the omnidirectional mobile robot and improve the high performance, stability and robustness of the omnidirectional robot.

[0012] The present invention provides an omnidirectional mobile robot anti-disturbance control system, comprising:

[0013] A disturbance item extraction module is used to extract all disturbance item information of the omnidirectional mobile robot in a pre-pre-period from the observation environment data of the omnidirectional mobile robot in a pre-pre-period;

[0014] A space equation generation module is used to generate a pre-disturbance expanded state space equation of the omnidirectional mobile robot based on all disturbance term information of the omnidirectional mobile robot in a pre-disturbance period and the currently known state space equation of the system;

[0015] A control quantity correction module is used to perform disturbance correction on the target control quantity of the omnidirectional mobile robot in a pre-period based on the unknown quantity of the observer gain parameter and the nonlinear function and the pre-perturbation expanded state space equation of the omnidirectional mobile robot, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period;

[0016] The anti-disturbance control module is used to take the anti-disturbance control amount of the omnidirectional mobile robot in a pre-period as the input control amount of the omnidirectional mobile robot in the pre-period to obtain the anti-disturbance control result of the omnidirectional mobile robot.

[0017] Preferably, the disturbance term extraction module includes:

[0018] The sensor observation submodule is used to obtain the environmental sensor observation data of the current environment of the omnidirectional mobile robot based on multiple environmental observation sensors;

[0019] An environmental video acquisition submodule is used to acquire a preview environmental video of the omnidirectional mobile robot from the preview perspective of the omnidirectional mobile robot;

[0020] An environmental observation data acquisition submodule is used to obtain the environmental advance sensor observation data and environmental advance visual observation data of the omnidirectional mobile robot based on the environmental sensor observation data and advance environmental video of the omnidirectional mobile robot's current environment;

[0021] The disturbance item extraction submodule is used to extract all disturbance item information of the omnidirectional mobile robot in the pre-pre period from the environment pre-pre sensing observation data and environment pre-pre visual observation data of the omnidirectional mobile robot.

[0022] Preferably, the environmental observation data acquisition submodule includes:

[0023] A sensor data prediction unit is used to predict the environmental sensor observation data of the current environment of the omnidirectional mobile robot based on the pre-environmental video, and obtain the pre-environmental sensor observation data of the omnidirectional mobile robot;

[0024] The visual observation data extraction unit is used to extract the environmental advance visual observation data of the omnidirectional mobile robot from the environmental advance video of the omnidirectional mobile robot.

[0025] Preferably, the sensor data prediction unit includes:

[0026] The influence expression determination subunit is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on the historical environmental sensor observation gradient data and the environmental video within the corresponding gradient range;

[0027] The influence item information extraction subunit is used to extract all the influence item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot from the pre-environment video;

[0028] The observation data prediction subunit is used to substitute all the influencing item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot into the influence expression of all the influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data, so as to obtain the environmental advance sensor observation data of the omnidirectional mobile robot.

[0029] Preferably, the influence expression determination subunit includes:

[0030] The first position determination end is used to respectively use the actual spatial position of each observation value in each observation data amount collected in the historical environmental sensor observation gradual change data as the basis position of each observation value;

[0031] The second position determination end is used to determine the amount of extractable data for all items in the environmental video within the corresponding gradient range, and determine whether each value in each amount of extractable data has only one basis video frame. If so, the position point corresponding to the central pixel point of the corresponding video frame in the actual spatial position is used as the basis position of the corresponding value; otherwise, the position points corresponding to the central pixel points of the intermediate video frames in the basis video frames in the actual spatial position are used as the basis position of the corresponding value;

[0032] A representation curve generating end is used to fit a data representation curve of each observation data amount with a corresponding basis position as a horizontal coordinate value based on all observation values in each observation data amount, and at the same time, fit a data representation curve of each extractable data amount with a corresponding basis position as a horizontal coordinate value based on all values in each extractable data amount;

[0033] The influence expression mining end is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on the data representation curve of all observation data quantities and the data representation curve of all extractable data quantities.

[0034] Preferably, the impact expression mining end includes:

[0035] The curve segmentation sub-terminal is used to analyze all high-probability inflection points in each data representation curve based on the curve gradient characteristics of each data representation curve, segment each data representation curve based on all the high-probability inflection points in each data representation curve, and obtain all segmented data representation curves of each data representation curve;

[0036] The influence item screening sub-terminal is used to screen out all extractable data quantities from all extractable data quantities, and to find out all extractable data quantities whose corresponding segmented data representation curves have the same or opposite increasing and decreasing characteristics as all segmented data representation curves of the single observation data quantity, and use them as all first screened data quantities of the corresponding observation data quantity;

[0037] The influence expression mining sub-end is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on all segmented data representation curves of each observation data amount and all segmented data representation curves of all corresponding first filtered data amounts.

[0038] Preferably, the method for the influence expression mining sub-end to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on all segmented data representation curves of each observation data amount and all segmented data representation curves of all corresponding first filtered data amounts includes:

[0039] Calculating the Pearson product-moment correlation coefficient between any observation value of each observation data amount and any value in each corresponding first screening data amount, and constructing a Pearson product-moment correlation coefficient matrix for each first screening data amount based on all Pearson product-moment correlation coefficients between each observation data amount and each corresponding first screening data amount;

[0040] All first-screened data quantities corresponding to the mode in the rank of the Pearson product-moment correlation coefficient matrix of all first-screened data quantities of each observation data quantity are regarded as all influencing items of the corresponding observation data quantity;

[0041] Multivariate regression analysis is performed on all segmented data representation curves of each observation data quantity and all segmented data representation curves of all corresponding influencing items to obtain the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data.

[0042] Preferably, the space equation generation module includes:

[0043] The known parameter determination submodule is used to determine the known disturbance state equations of all disturbance terms;

[0044] The state space expansion submodule is used to generate the pre-disturbance expanded state space equation of the omnidirectional mobile robot based on the known disturbance state equations of all disturbance terms and the current known state space equation of the system.

[0045] Preferably, the control amount correction module includes:

[0046] An observer function generation submodule is used to generate an expanded state nonlinear observer function based on unknown quantities of observer gain parameters and nonlinear functions and a pre-disturbance expanded state space equation of the omnidirectional mobile robot;

[0047] The disturbance term representation submodule is used to determine the expressions of all disturbance terms based on the extended state nonlinear observer function;

[0048] The control quantity correction submodule is used to perform disturbance correction on the target control quantity of the omnidirectional mobile robot in the pre-period based on the expressions of all disturbance terms, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period.

[0049] Preferably, the control amount correction submodule includes:

[0050] A disturbance control amount determination unit is used to substitute all disturbance item information into the expressions of all disturbance items to determine the disturbance action control amounts of all disturbance items;

[0051] The control quantity correction unit is used to perform the same term difference between the target control quantity of the omnidirectional mobile robot in the pre-period and the disturbance effect control quantity of all disturbance items, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period.

[0052] The beneficial effects of the present invention compared to the prior art are as follows: based on all disturbance item information of the omnidirectional mobile robot in a pre-period and the analysis principle of the extended state observer, high-precision disturbance correction is performed on the target control quantity of the omnidirectional mobile robot in the pre-period, so as to realize high-precision anti-disturbance control of the omnidirectional mobile robot and improve the high performance, stability and robustness of the omnidirectional robot.

[0053] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0054] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0056] Figure 1 Schematic diagram of the internal functional modules of the anti-disturbance control system of an omnidirectional mobile robot in an embodiment of the present invention;

[0057] Figure 2 Schematic diagram of internal functional submodules of the disturbance term extraction module in an embodiment of the present invention;

[0058] Figure 3 This is a schematic diagram of the internal functional submodules of the spatial equation generation module in an embodiment of the present invention;

[0059] Figure 4 Schematic diagram of internal functional submodules of a control amount correction module in an embodiment of the present invention;

[0060] Figure 5 FIG. 4 is a logic diagram of a state observer in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0062] Example 1:

[0063] The present invention provides an omnidirectional mobile robot anti-disturbance control system, referring to Figure 1 ,include:

[0064] A disturbance item extraction module is used to extract all disturbance item information of the omnidirectional mobile robot in a pre-pre-period from the observation environment data of the omnidirectional mobile robot in a pre-pre-period;

[0065] A space equation generation module is used to generate a pre-disturbance expanded state space equation of the omnidirectional mobile robot based on all disturbance term information of the omnidirectional mobile robot in a pre-disturbance period and the currently known state space equation of the system;

[0066] A control quantity correction module is used to perform disturbance correction on the target control quantity of the omnidirectional mobile robot in a pre-period based on the unknown quantity of the observer gain parameter and the nonlinear function and the pre-perturbation expanded state space equation of the omnidirectional mobile robot, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period;

[0067] The anti-disturbance control module is used to take the anti-disturbance control amount of the omnidirectional mobile robot in a pre-period as the input control amount of the omnidirectional mobile robot in the pre-period to obtain the anti-disturbance control result of the omnidirectional mobile robot.

[0068] In this embodiment, the omnidirectional mobile robot is a special robot that has the ability to move freely in any direction without changing its orientation. This makes omnidirectional robots ideal for applications such as navigation and positioning. Omnidirectional robots can operate in a variety of environments, from flat surfaces to rugged terrain, and can easily move through confined spaces, making them very useful mobile platforms.

[0069] Unlike wheeled robots, omnidirectional robots do not use rolling mechanisms such as wheels or tracks. Instead, they use multi-jointed legs or multi-wheel drive systems with greater maneuverability. This design enables omnidirectional robots to move flexibly in various terrains and environments, and they can also maintain a relatively stable posture without changing their posture.

[0070] In this embodiment, the prediction period is a period starting from the current time and lasting for a period of time.

[0071] In this embodiment, the observed environment data includes the environment advance sensor observation data and environment advance visual observation data of the omnidirectional mobile robot.

[0072] In this embodiment, the disturbance item information is interference information that can be extracted from the observed environment data and affects the accuracy of the omnidirectional mobile robot control system, such as changes in ground resistance on the ground ahead or mechanical impact in the environment ahead.

[0073] In this embodiment, the currently known state space equation of the omnidirectional mobile robot system is a state equation of a linear control system using X to represent the system state variable, U to represent the system input, and Y to represent the system output, where X is an n-dimensional vector:

[0074]

[0075] In this embodiment, the unknown quantity of the observer gain parameter is the unknown quantity representing the observer gain parameter in the nonlinear function correction term when a nonlinear function correction term is added to the current known state space equation of the omnidirectional mobile robot system for compensation. This parameter determines the accuracy of the observer.

[0076] In this embodiment, when the nonlinear function is added to the current known state space equation of the omnidirectional mobile robot system to compensate, the nonlinear function g in the nonlinear function correction term is i (e), its expression is:

[0077]

[0078] In this embodiment, the target control quantity of the omnidirectional mobile robot in the pre-cycle is the parameter U input to the control system of the omnidirectional mobile robot in the original control instruction issued by the control end.

[0079] In this embodiment, the anti-disturbance control amount of the omnidirectional mobile robot in the pre-period is the parameter U obtained by correcting and compensating the target control amount after taking into account the disturbance caused by the disturbance item information in the environment to the posture and state of the omnidirectional mobile robot.

[0080] In this embodiment, the input control variable is the parameter U actually input into the control system of the omnidirectional mobile robot.

[0081] The beneficial effects of the above technology are: based on all disturbance item information of the omnidirectional mobile robot in the pre-period and the analysis principle of the extended state observer, high-precision disturbance correction is performed on the target control quantity of the omnidirectional mobile robot in the pre-period, so as to realize high-precision anti-disturbance control of the omnidirectional mobile robot and improve the high performance, stability and robustness of the omnidirectional robot.

[0082] Example 2:

[0083] Based on Example 1, the disturbance term extraction module, refer to Figure 2,include:

[0084] The sensor observation submodule is used to obtain the environmental sensor observation data of the current environment of the omnidirectional mobile robot based on multiple environmental observation sensors;

[0085] An environmental video acquisition submodule is used to acquire a preview environmental video of the omnidirectional mobile robot from the preview perspective of the omnidirectional mobile robot;

[0086] An environmental observation data acquisition submodule is used to obtain the environmental advance sensor observation data and environmental advance visual observation data of the omnidirectional mobile robot based on the environmental sensor observation data and advance environmental video of the omnidirectional mobile robot's current environment;

[0087] The disturbance item extraction submodule is used to extract all disturbance item information of the omnidirectional mobile robot in the pre-pre period from the environment pre-pre sensing observation data and environment pre-pre visual observation data of the omnidirectional mobile robot.

[0088] In this embodiment, the environment observation sensor is a sensor that can observe the environment and obtain corresponding environment data, such as a temperature sensor, a humidity sensor, an air pressure sensor, etc.

[0089] In this embodiment, the environmental sensing observation data is data obtained by using an environmental observation sensor and includes multiple pieces of environmental data.

[0090] In this embodiment, the preview viewing angle of the omnidirectional mobile robot is a viewing angle parallel to the direction in which the omnidirectional mobile robot will next drive.

[0091] In this embodiment, the preview environment video of the omnidirectional mobile robot is a video containing details of the actual space that the omnidirectional robot will enter next.

[0092] In this embodiment, all disturbance item information of the omnidirectional mobile robot in the pre-prediction period is extracted from the pre-prediction environmental sensor observation data and pre-prediction environmental visual observation data of the omnidirectional mobile robot, for example:

[0093] The road ahead may contain a road surface with large resistance locally, or rainy weather in the road ahead that will cause large resistance to the robot's movement, which is extracted from the environmental advance visual sensing data.

[0094] The beneficial effects of the above technology are: based on multiple environmental observation sensors and the preliminary environmental video of the omnidirectional mobile robot, a large-scale acquisition of the environmental data in the space that the omnidirectional mobile robot passes through during the preliminary period is achieved, and it is easy to extract the disturbance item information that will cause errors in the robot's control or movement.

[0095] Example 3:

[0096] Based on Example 2, the environmental observation data acquisition submodule includes:

[0097] A sensor data prediction unit is used to predict the environmental sensor observation data of the current environment of the omnidirectional mobile robot based on the pre-environmental video, and obtain the pre-environmental sensor observation data of the omnidirectional mobile robot;

[0098] The visual observation data extraction unit is used to extract the environmental advance visual observation data of the omnidirectional mobile robot from the environmental advance video of the omnidirectional mobile robot.

[0099] In this embodiment, the environmental pre-sensing observation data is the predicted performance values of multiple environmental data in the current environmental sensor observation data of the omnidirectional mobile robot in the pre-sessing period.

[0100] In this embodiment, the environmental pre-viewing visual observation data of the omnidirectional mobile robot is extracted from the pre-viewing environmental video of the omnidirectional mobile robot, including:

[0101] The data of dust density, rainfall and road friction on the road ahead are extracted from the preview environment video of the omnidirectional mobile robot as the preview visual observation data of the environment of the omnidirectional mobile robot.

[0102] The beneficial effects of the above technology are: obtaining the environmental advance sensor observation data and environmental advance visual observation data of the omnidirectional mobile robot.

[0103] Example 4:

[0104] Based on embodiment 3, the sensor data prediction unit includes:

[0105] The influence expression determination subunit is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on the historical environmental sensor observation gradient data and the environmental video within the corresponding gradient range;

[0106] The influence item information extraction subunit is used to extract all the influence item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot from the pre-environment video;

[0107] The observation data prediction subunit is used to substitute all the influencing item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot into the influence expression of all the influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data, so as to obtain the environmental advance sensor observation data of the omnidirectional mobile robot.

[0108] In this embodiment, the historical environmental sensor observation gradient data and the environmental video within the corresponding gradient range are the values of the environmental sensor observation data obtained by the omnidirectional robot in a previous continuous period of time that change with time, as well as the preliminary environmental video obtained in an even previous continuous period of time, which includes the space passed by the omnidirectional mobile robot in this continuous period of time.

[0109] In this embodiment, all influencing items of each environmental sensor observation data are data classes extracted from the preliminary environmental video that affect each environmental sensor observation data, such as the travel height extracted from the preliminary environmental video, which affects the ambient temperature and humidity of the space it is about to enter, and the rainfall extracted from the preliminary environmental video, which affects the ambient temperature and humidity of the space it is about to enter.

[0110] In this embodiment, the influence expression is an expression that represents the numerical relationship between all the influence terms of each environmental sensor observation data and the corresponding environmental sensor observation data.

[0111] In this embodiment, all influencing information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot is extracted from the preliminary environmental video, such as the specific values of the front travel height and travel distance, the specific value of the front rainfall, etc.

[0112] In this embodiment, all influencing item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot is substituted into the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data to obtain the environmental preliminary sensor observation data of the omnidirectional mobile robot, including:

[0113] The performance values of the corresponding influencing items contained in all the influencing item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot at different travel positions (i.e., positions at different distances from the current position in the front travel path) are substituted into the influence expression of all the influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data to obtain the performance values of the environmental data of the omnidirectional mobile robot in the environment at different travel positions, such as the temperature value in the space where the omnidirectional mobile robot is located when it is at a position 50 meters ahead from the current position.

[0114] The beneficial effects of the above technology are: to achieve accurate determination of the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data, and based on all influencing item information and influence expressions of each environmental sensor observation data of the current environment of the omnidirectional mobile robot, to achieve accurate prediction of the environmental advance sensor observation data of the omnidirectional mobile robot.

[0115] Example 5:

[0116] On the basis of Example 4, the influence expression determination subunit includes:

[0117] The first position determination end is used to respectively use the actual spatial position of each observation value in each observation data amount collected in the historical environmental sensor observation gradual change data as the basis position of each observation value;

[0118] The second position determination end is used to determine the amount of extractable data for all items in the environmental video within the corresponding gradient range, and determine whether each value in each amount of extractable data has only one basis video frame. If so, the position point corresponding to the central pixel point of the corresponding video frame in the actual spatial position is used as the basis position of the corresponding value; otherwise, the position points corresponding to the central pixel points of the intermediate video frames in the basis video frames in the actual spatial position are used as the basis position of the corresponding value;

[0119] A representation curve generating end is used to fit a data representation curve of each observation data amount with a corresponding basis position as a horizontal coordinate value based on all observation values in each observation data amount, and at the same time, fit a data representation curve of each extractable data amount with a corresponding basis position as a horizontal coordinate value based on all values in each extractable data amount;

[0120] The influence expression mining end is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on the data representation curve of all observation data quantities and the data representation curve of all extractable data quantities.

[0121] In this embodiment, the amount of extractable data for all items in the environmental video within the gradual change range refers to all data items that can be extracted from the environmental video within the gradual change range.

[0122] In this embodiment, the video frame on which the numerical value is based is the video frame that determines the numerical value. For example, when 20 mm of rainfall occurs at a position 10 meters away from the current position in the preliminary environmental video, the video frame that reads the information is the video frame on which the numerical value "20 mm occurs at a position 10 meters away from the current position" is based.

[0123] In this embodiment, the central pixel point of the video frame is the pixel point located at the center of the video frame.

[0124] In this embodiment, when the total number of all the based video frames is an odd number, the center pixel point of the middlemost video frame among all the based video frames is regarded as the center pixel point of the middle video frames among all the based video frames;

[0125] When the total number of all the based video frames is an even number, the center pixel of the video frame that is ordered earlier among the two middle video frames among all the based video frames is regarded as the center pixel of the middle video frame among all the based video frames.

[0126] In this embodiment, the data representation curve of each observation data amount is a curve fitted by all the values of each observation data amount, the horizontal axis of which is the basis position of the value in the observation data amount, and the vertical axis is the value in the observation data amount.

[0127] In this embodiment, the data representation curve of each extractable data amount is a curve fitted by all the values of each extractable data amount, the horizontal axis of which is the basis position of the value in the extractable data amount, and the vertical axis is the value in the extractable data amount.

[0128] The beneficial effects of the above technology are: by reasonably determining the basis position of each observation value in each observation data quantity in the historical environmental sensor observation gradient data and the basis position of each value of all extractable data quantities in the environmental video within the corresponding gradient range, it is convenient to fit the data representation curves of all observation data quantities and the data representation curves of all extractable data quantities that represent their values changing with the basis position, so as to facilitate the subsequent accurate mining of the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data.

[0129] Example 6:

[0130] Based on Example 5, the impact expression mining end includes:

[0131] The curve segmentation sub-terminal is used to analyze all high-probability inflection points in each data representation curve based on the curve gradient characteristics of each data representation curve, segment each data representation curve based on all the high-probability inflection points in each data representation curve, and obtain all segmented data representation curves of each data representation curve;

[0132] The influence item screening sub-terminal is used to screen out all extractable data quantities from all extractable data quantities, and to find out all extractable data quantities whose corresponding segmented data representation curves have the same or opposite increasing and decreasing characteristics as all segmented data representation curves of the single observation data quantity, and use them as all first screened data quantities of the corresponding observation data quantity;

[0133] The influence expression mining sub-end is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on all segmented data representation curves of each observation data amount and all segmented data representation curves of all corresponding first filtered data amounts.

[0134] In this embodiment, the curve gradient feature of the data representation curve is the first derivative function of the representation function of the data representation curve.

[0135] In this embodiment, based on the curve gradient characteristics of each data representation curve, all high-probability inflection points in each data representation curve are analyzed, including:

[0136] The horizontal coordinate of the breakpoint in the function curve of the first derivative function of the data representation curve is the horizontal coordinate of the high probability inflection point;

[0137] All high-probability inflection points in the data representation curve are determined based on the abscissas of the high-probability inflection points.

[0138] In this embodiment, the high probability inflection point is a point in the data representation curve whose abscissa is consistent with the abscissa of a breakpoint in the function curve of the first derivative function of the representation function of the data representation curve.

[0139] In this embodiment, each data representation curve is segmented based on all high-probability inflection points in each data representation curve, namely:

[0140] All high-probability inflection points in each data representation curve are regarded as segmentation positions, and each data representation curve is segmented.

[0141] In this embodiment, the increasing and decreasing change characteristics are characteristics indicating the monotonically increasing or monotonically decreasing or the order and number of occurrences of the monotonically increasing and monotonically decreasing curves of the segmented data representation curve.

[0142] The beneficial effects of the above technology are: by identifying all high-probability inflection points appearing in the data representation curve and segmenting the data representation curve based on this, it is convenient to perform subsequent segmented comparison of the data representation curve, and by performing segmented corresponding comparison of the data representation curves of all extractable data quantities and all observed data quantities, the first step of rough screening of all influencing items of each environmental sensor observation data is achieved, and based on all segmented data representation curves of each observation data quantity and the roughly screened first-selected data quantity and all segmented data representation curves, the accurate determination of the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data is achieved.

[0143] Example 7:

[0144] Based on Example 6, the method for the influence expression mining sub-end to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on all segmented data representation curves of each observation data amount and all segmented data representation curves of all corresponding first filtered data amounts includes:

[0145] Calculating the Pearson product-moment correlation coefficient between any observation value of each observation data amount and any value in each corresponding first screening data amount, and constructing a Pearson product-moment correlation coefficient matrix for each first screening data amount based on all Pearson product-moment correlation coefficients between each observation data amount and each corresponding first screening data amount;

[0146] All first-screened data quantities corresponding to the mode in the rank of the Pearson product-moment correlation coefficient matrix of all first-screened data quantities of each observation data quantity are regarded as all influencing items of the corresponding observation data quantity;

[0147] Multivariate regression analysis is performed on all segmented data representation curves of each observation data quantity and all segmented data representation curves of all corresponding influencing items to obtain the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data.

[0148] In this embodiment, a Pearson product-moment correlation coefficient matrix of each first screening data volume is constructed based on all Pearson product-moment correlation coefficients between each observation data volume and each corresponding first screening data volume, where:

[0149] The number of rows of the Pearson product-moment correlation coefficient matrix is equal to the total number of observations contained in the corresponding item observation data volume, and the number of columns of the Pearson product-moment correlation coefficient matrix is equal to the total number of values contained in the corresponding first screening data volume;

[0150] Furthermore, the Pearson product-moment correlation coefficient between the first observation value included in the corresponding item observation data volume and the first value included in the corresponding first screening data volume is regarded as the value of the first row and first column in the Pearson product-moment correlation coefficient matrix;

[0151] The Pearson product-moment correlation coefficient between the second observation value contained in the corresponding item observation data volume and the first value contained in the corresponding first screening data volume is regarded as the value in the second row and first column of the Pearson product-moment correlation coefficient matrix;

[0152] The Pearson product-moment correlation coefficient between the first observation value contained in the corresponding item observation data volume and the second value contained in the corresponding first screening data volume is regarded as the value in the first row and second column of the Pearson product-moment correlation coefficient matrix;

[0153] By analogy, the Pearson product-moment correlation coefficient matrix corresponding to the first screening data volume is further determined.

[0154] In this embodiment, multivariate regression analysis is performed on all segmented data representation curves of each observation data quantity and all segmented data representation curves of all corresponding influencing items to obtain the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data, which is:

[0155] The existing multivariate regression analysis technology (such as multiple regression analysis or multivariate multiple regression analysis) is used to perform multivariate regression analysis on all segmented data representation curves of each observation data quantity and all segmented data representation curves of all corresponding influencing items, and then analyze the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data.

[0156] The beneficial effects of the above technology are: based on all Pearson product-moment correlation coefficients between each observation data amount and each corresponding first screening data amount, a Pearson product-moment correlation coefficient matrix of each first screening data amount is constructed, and the mode in the rank of the Pearson product-moment correlation coefficient matrix of all first screening data amounts of each observation data amount is used as the screening principle to accurately screen out the influencing items of each observation data amount, thereby improving the accuracy of the variable categories in the final determination of the influencing expression, and then using the multivariate regression analysis method to accurately analyze the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data.

[0157] Example 8:

[0158] Based on Example 1, the space equation generation module, refer to Figure 3 ,include:

[0159] The known parameter determination submodule is used to determine the known disturbance state equations of all disturbance terms;

[0160] The state space expansion submodule is used to generate the pre-disturbance expanded state space equation of the omnidirectional mobile robot based on the known disturbance state equations of all disturbance terms and the current known state space equation of the system.

[0161] In this embodiment, the known disturbance state equations of all disturbance terms are determined, including:

[0162] That is, in the n-dimensional state space, expand the state variables of the n+1th dimension:

[0163]

[0164] The above formula is the state equation of the known disturbance.

[0165] In this embodiment, based on the known disturbance state equations of all disturbance terms and the current known state space equation of the system, the pre-disturbance expanded state space equation of the omnidirectional mobile robot is generated, namely:

[0166]

[0167] The beneficial effects of the above technology are: by expanding the disturbance variables to the control system of the omnidirectional mobile robot, the observation of the disturbance is expanded based on the principle of the extended state observer, which can greatly improve the performance of the controller.

[0168] Example 9:

[0169] On the basis of Example 1, the control amount correction module, refer to Figure 4 ,include:

[0170] An observer function generation submodule is used to generate an expanded state nonlinear observer function based on unknown quantities of observer gain parameters and nonlinear functions and a pre-disturbance expanded state space equation of the omnidirectional mobile robot;

[0171] The disturbance term representation submodule is used to determine the expressions of all disturbance terms based on the extended state nonlinear observer function;

[0172] The control quantity correction submodule is used to perform disturbance correction on the target control quantity of the omnidirectional mobile robot in the pre-period based on the expressions of all disturbance terms, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period.

[0173] In this embodiment, an extended state nonlinear observer function is generated based on unknown observer gain parameters and nonlinear functions and the pre-disturbance extended state space equation of the omnidirectional mobile robot, including:

[0174] Based on the nonlinear function compensation method, the unknown observer gain parameter and nonlinear function are introduced into the pre-disturbance extended state space equation of the omnidirectional mobile robot to obtain the generated extended state nonlinear observer function:

[0175]

[0176] Among them, β i is the unknown observer gain parameter, g i (e) is a nonlinear function, i∈[1,n].

[0177] In this embodiment, expressions of all disturbance terms are determined based on the extended state nonlinear observer function, including:

[0178] Based on the extended state nonlinear observer function, a nonlinear observer model is built in Simulink, and state estimation is performed based on the values of different groups of unknown observer gain parameters. The convergence effect is observed in the estimation results, and a suitable group of unknown observer gain parameters is selected. The values of this group of unknown observer gain parameters are substituted into the pre-disturbance extended state space equation to obtain the generated extended state nonlinear observer function.

[0179] The beneficial effects of the above technology are: by expanding the disturbance variables to the control system of the omnidirectional mobile robot, the observation of the disturbance is expanded based on the principle of the extended state observer, and the observation results are offset through the input, which can greatly improve the performance of the controller.

[0180] Example 10:

[0181] Based on Example 9, the control amount correction submodule includes:

[0182] A disturbance control amount determination unit is used to substitute all disturbance item information into the expressions of all disturbance items to determine the disturbance action control amounts of all disturbance items;

[0183] The control quantity correction unit is used to perform the same term difference between the target control quantity of the omnidirectional mobile robot in the pre-period and the disturbance effect control quantity of all disturbance items, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period.

[0184] In this embodiment, all disturbance term information is substituted into the expressions of all disturbance terms to determine the disturbance action control quantities of all disturbance terms, including:

[0185] Substitute the values in all disturbance term information into the expressions of all disturbance terms to determine the disturbance effect control quantities of all disturbance terms.

[0186] In this embodiment, the target control amount of the omnidirectional mobile robot in the pre-period is subtracted from the disturbance effect control amount of all disturbance items to obtain the anti-disturbance control amount of the omnidirectional mobile robot in the pre-period, including:

[0187] The difference between each value in the target control quantity of the omnidirectional mobile robot in the preliminary period and the sum of the corresponding same value in the disturbance effect control quantity of all disturbance items is regarded as the anti-disturbance control quantity of the omnidirectional mobile robot in the preliminary period.

[0188] The beneficial effects of the above technology are: offsetting the disturbance control quantity caused by the disturbance term in the target control quantity of the omnidirectional mobile robot in the pre-period, achieving high-precision disturbance correction of the target control quantity of the omnidirectional mobile robot in the pre-period, so as to achieve high-precision anti-disturbance control of the omnidirectional mobile robot and improve the high performance, stability and robustness of the omnidirectional robot.

[0189] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An omnidirectional mobile robot anti-disturbance control system, characterized in that: include: The disturbance item extraction module is used to extract all disturbance item information of the omnidirectional mobile robot in the pre-pre-period from the observation environment data of the omnidirectional mobile robot in the pre-pre-period, including: The sensor observation submodule is used to obtain environmental sensor observation data of the current environment of the omnidirectional mobile robot based on multiple environmental observation sensors; An environmental video acquisition submodule is used to acquire a preview environmental video of the omnidirectional mobile robot from the preview perspective of the omnidirectional mobile robot; The environmental observation data acquisition submodule is used to obtain the environmental pre-sensing observation data and environmental pre-visual observation data of the omnidirectional mobile robot based on the environmental sensor observation data and pre-environmental video of the omnidirectional mobile robot's current environment, including: The sensor data prediction unit is used to predict the environmental sensor observation data of the current environment of the omnidirectional mobile robot based on the pre-environmental video, and obtain the pre-environmental sensor observation data of the omnidirectional mobile robot, including: The influence expression determination subunit is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on the historical environmental sensor observation gradient data and the environmental video within the corresponding gradient range, including: The first position determination end is used to respectively use the actual spatial position of each observation value in each observation data amount collected in the historical environmental sensor observation gradual change data as the basis position of each observation value; The second position determination end is used to determine the amount of extractable data for all items in the environmental video within the corresponding gradient range, and determine whether each value in each amount of extractable data has only one basis video frame. If so, the position point corresponding to the central pixel point of the corresponding video frame in the actual spatial position is used as the basis position of the corresponding value; otherwise, the position points corresponding to the central pixel points of the intermediate video frames in the basis video frames in the actual spatial position are used as the basis position of the corresponding value; A representation curve generating end is used to fit a data representation curve of each observation data amount with a corresponding basis position as a horizontal coordinate value based on all observation values in each observation data amount, and at the same time, fit a data representation curve of each extractable data amount with a corresponding basis position as a horizontal coordinate value based on all values in each extractable data amount; The influence expression mining end is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on the data representation curve of all item observation data quantities and the data representation curve of all item extractable data quantities, including: The curve segmentation sub-terminal is used to analyze all high-probability inflection points in each data representation curve based on the curve gradient characteristics of each data representation curve, segment each data representation curve based on all the high-probability inflection points in each data representation curve, and obtain all segmented data representation curves of each data representation curve; The influence item screening sub-terminal is used to screen out all extractable data quantities from all extractable data quantities, and to find out all extractable data quantities whose corresponding segmented data representation curves have the same or opposite increasing and decreasing characteristics as all segmented data representation curves of the single observation data quantity, and use them as all first screened data quantities of the corresponding observation data quantity; The influence expression mining sub-end is used to mine the influence expressions of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data based on all segmented data representation curves of each observation data volume and all segmented data representation curves of all corresponding first-screened data volumes, including: Calculating the Pearson product-moment correlation coefficient between any observation value of each observation data amount and any value in each corresponding first screening data amount, and constructing a Pearson product-moment correlation coefficient matrix for each first screening data amount based on all Pearson product-moment correlation coefficients between each observation data amount and each corresponding first screening data amount; All first-screened data quantities corresponding to the mode in the rank of the Pearson product-moment correlation coefficient matrix of all first-screened data quantities of each observation data quantity are regarded as all influencing items of the corresponding observation data quantity; Perform multivariate regression analysis on all segmented data representation curves of each observation data quantity and all segmented data representation curves of all corresponding influencing items to obtain the influence expression of all influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data; The influence item information extraction subunit is used to extract all the influence item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot from the pre-environment video; The observation data prediction subunit is used to substitute all the influencing item information of each environmental sensor observation data of the current environment of the omnidirectional mobile robot into the influence expression of all the influencing items of each environmental sensor observation data on the corresponding environmental sensor observation data, and obtain the environmental pre-sensing observation data of the omnidirectional mobile robot. A visual observation data extraction unit is used to extract the environmental advance visual observation data of the omnidirectional mobile robot from the environmental advance video of the omnidirectional mobile robot; The disturbance term extraction submodule is used to extract all disturbance term information of the omnidirectional mobile robot in the pre-prediction period from the environmental pre-prediction sensor observation data and the environmental pre-prediction visual observation data of the omnidirectional mobile robot; A space equation generation module is used to generate a pre-disturbance expanded state space equation of the omnidirectional mobile robot based on all disturbance term information of the omnidirectional mobile robot in a pre-disturbance period and the currently known state space equation of the system; A control quantity correction module is used to perform disturbance correction on the target control quantity of the omnidirectional mobile robot in a pre-period based on the unknown quantity of the observer gain parameter and the nonlinear function and the pre-perturbation expanded state space equation of the omnidirectional mobile robot, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period; an anti-disturbance control module, configured to use the anti-disturbance control amount of the omnidirectional mobile robot in a pre-period as the input control amount of the omnidirectional mobile robot in the pre-period to obtain an anti-disturbance control result of the omnidirectional mobile robot; Among them, the control amount correction module includes: An observer function generation submodule is used to generate an expanded state nonlinear observer function based on unknown quantities of observer gain parameters and nonlinear functions and a pre-disturbance expanded state space equation of the omnidirectional mobile robot; The disturbance term representation submodule is used to determine the expressions of all disturbance terms based on the extended state nonlinear observer function; The control quantity correction submodule is used to perform disturbance correction on the target control quantity of the omnidirectional mobile robot in the pre-period based on the expressions of all disturbance terms, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period.

2. The anti-disturbance control system of the omnidirectional mobile robot according to claim 1, characterized in that: Space equation generation module, including: The known parameter determination submodule is used to determine the known disturbance state equations of all disturbance terms; The state space expansion submodule is used to generate the pre-disturbance expanded state space equation of the omnidirectional mobile robot based on the known disturbance state equations of all disturbance terms and the current known state space equation of the system.

3. The anti-disturbance control system of the omnidirectional mobile robot according to claim 1, characterized in that: The control amount correction submodule includes: A disturbance control amount determination unit is used to substitute all disturbance item information into the expressions of all disturbance items to determine the disturbance action control amounts of all disturbance items; The control quantity correction unit is used to perform the same term difference between the target control quantity of the omnidirectional mobile robot in the pre-period and the disturbance effect control quantity of all disturbance items, so as to obtain the anti-disturbance control quantity of the omnidirectional mobile robot in the pre-period.

Images