Mobile robot positioning method based on multi-data analysis
By combining the data of the odometer and navigation system in indoor robots, and using data coupled prediction algorithms for global coordinate system analysis, the problem of low indoor positioning accuracy is solved and more accurate robot positioning is achieved.
Patent Information
- Application Number
- CN202111215125.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-10-19
AI Technical Summary
The complex indoor environment and signal interference lead to low accuracy of indoor robot positioning technology. The existing technology relies on a single sensor to cause large errors, making it difficult to provide an accurate indoor robot positioning method.
Using a multi-data analysis method, combined with the data of the odometer and navigation system, analyses and calculations are performed under the global coordinate system through a data coupled prediction algorithm to obtain the positioning information of the robot and reduce the impact of environmental interference.
It improves the accuracy of robot positioning, reduces the error caused by a single sensor, and enhances the accuracy of indoor positioning.
Smart Images

Figure CN113984057B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of robot technology, and in particular to a mobile robot positioning method based on multi-data analysis. Background Art
[0002] At present, positioning technology in outdoor scenarios relies on fixed sensors or GPS-based sensors, and GPS has become the most widely used satellite-based positioning system because it can provide the largest coverage area. However, the following problems have caused great limitations in the application of outdoor positioning technology in indoor scenarios: the indoor environment is more complex than the outdoor environment; there is strong signal interference and reflection indoors, and the positioning error based on signals will be large; indoor positioning is highly dependent on the indoor environment; indoor communication links become unreliable.
[0003] At present, the more common indoor positioning technology mainly determines the robot's posture state based on prior environmental information, combined with the current robot position information and information obtained by the sensor. Indoor relative positioning determines the current posture state by calculating the posture change relative to the initial posture. Common ones include odometers and navigation systems. If the positioning technology relies on a single sensor, there will be a large error. The odometer, navigation system, etc. have large cumulative errors and poor positioning accuracy over a long period of time or on a large scale. Therefore, how to provide an accurate robot indoor positioning method has become a technical problem that technicians in this field need to solve urgently. Summary of the invention
[0004] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide a mobile robot positioning method based on multi-data analysis with improved positioning accuracy.
[0005] The technical solution adopted by the present invention to solve the technical problem is: a mobile robot positioning method based on multi-data analysis, wherein the robot is provided with an odometer and a navigation system, wherein the odometer is used to record the robot's forward distance and the robot's rotation angle, and the navigation system is used to record the robot's current rotation direction angle and acceleration;
[0006] The robot's forward distance and the robot's rotation angle recorded by the odometer are recorded as odometer data;
[0007] The robot's current rotation angle and acceleration recorded by the navigation system are recorded as navigation data;
[0008] The following steps are involved:
[0009] Step 1: Convert the odometer data and navigation data to the global coordinate system to obtain the representation of the relevant mobile data information in the global coordinate system;
[0010] Step 2: Analyze and calculate the odometer data and navigation data represented in the global coordinate system to obtain analysis data;
[0011] Step 3: Get the robot's positioning information based on the analysis data.
[0012] The step 1 includes the following sub-steps:
[0013] 1-1: Set the starting point (x0, y0, z0) as the origin of the global coordinate system, and use the right-hand rule to determine the positive directions of the x-axis, y-axis, and z-axis;
[0014] 1-2: The robot moves on a plane. The odometer on the robot starts to calculate the robot's motion increment information from the starting position (x0, y0, z0), including (Δx, Δy, Δz), where: Δx records the distance the robot has moved forward, Δz records the robot's rotation angle, and converts the data information to the global coordinate system; the coordinate system conversion library TF library of the robot operating system ROS platform is needed to convert the odometer data to the global coordinate system.
[0015] 1-3: The navigation system continuously outputs the robot's current state information, including the rotation angle w and acceleration a x and a y , convert the data information to the global coordinate system. Use the coordinate system conversion library TF library of the robot operating system ROS platform to convert the navigation data to the global coordinate system.
[0016] In the step 2, the odometer data and navigation data represented in the global coordinate system are analyzed and calculated by a data coupling prediction algorithm to obtain analysis data.
[0017] The step 2 includes the following sub-steps:
[0018] 2-1: Set state variables;
[0019] 2-2: Using the odometer data iteration data as the motion update state quantity corresponding to the state variable;
[0020] 2-3: Using the secondary time integration data of the navigation data as the corresponding correction variable;
[0021] 2-4: According to the data coupling prediction algorithm, the motion update state quantity and the correction variable are superimposed to obtain the analyzed data.
[0022] In step 2-1, the state variable is (x, y, z).
[0023] In step 2-2, the odometer data (Δx, Δy, Δz) is used as the motion update state quantity. After the odometer data (Δx, Δy, Δz) is iteratively calculated at the starting position (x0, y0, z0), it corresponds to the new motion update state quantity (x+Δx, y+Δy, z+Δz).
[0024] In step 2-3, the navigation data acceleration a x and a y The second time integration can obtain the motion information in the x and y directions (∫∫a x ,∫∫a y ), and the rotation direction angle w of the navigation data is the corresponding variable to be corrected.
[0025] In steps 2-4, the calculation steps of the data coupling prediction algorithm are as follows:
[0026] 2-4-1: Predict the motion state X(K) of K at the current moment. The calculation formula is as follows:
[0027] X(K)={(x0+Δx), (y0+Δy), (z0+Δz)};
[0028] k is the current time;
[0029] 2-4-2: Predict the correction amount Y(K) of K at the current moment. The calculation formula is as follows:
[0030] Y(K)={a x ,a y ,w};
[0031]
[0032] 2-4-3: First-order linearized state equation, solve the state transfer matrix Φ(k), the calculation formula is as follows:
[0033]
[0034] 2-4-4: First-order linearized observation equation, solve the observation matrix H(K), the calculation formula is as follows:
[0035]
[0036] 2-4-5: Find the covariance matrix prediction P(k|k-1), the calculation formula is as follows:
[0037] P(k|k-1)=Φ(k)P(k-1|k-1)Φ T (k);
[0038] 2-4-6: Noise removal processing W(k), the calculation formula is as follows:
[0039] W(k)=P(k|k-1)H T (k)(H(k)P(k|k-1)H T (k))
[0040] 2-4-7: Find the state update X(k), the calculation formula is as follows:
[0041] X(k)=X(k|k-1)+K(Y(k)-Y(k|k-1))
[0042] 2-4-8: Covariance update P(k), the calculation formula is as follows:
[0043] P(k)=(I n -K(k)H(k))P(k|k-1), where I n is a 3D vector matrix. The value of n is determined by the dimensions of vectors X and Y. Through the displacement correction obtained by the above data coupling analysis algorithm, the robot positioning is more accurate and is not affected by the interference and influence of the indoor and outdoor environment.
[0044] Compared with the prior art, the present invention has the following beneficial effects:
[0045] The present invention provides a mobile robot positioning method based on multi-data analysis, which solves the problems of large errors and inaccurate positioning in the prior art that rely solely on navigation positioning by performing comprehensive analysis and calculation on the robot's movement data recorded by an odometer and the robot's movement data recorded by a navigation system, thereby improving the accuracy of robot positioning. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 It is a flow chart of the present invention.
[0047] Figure 2 It is a data analysis calculation flow chart of the present invention. DETAILED DESCRIPTION
[0048] The embodiments of the present invention are further described below in conjunction with the accompanying drawings:
[0049] Example 1
[0050] like Figure 1 to Figure 2 As shown, the robot is provided with an odometer and a navigation system, wherein the odometer is used to record the robot's forward distance and the robot's rotation angle, and the navigation system is used to record the robot's current rotation direction angle and acceleration;
[0051] The robot's forward distance and the robot's rotation angle recorded by the odometer are recorded as odometer data;
[0052] The robot's current rotation angle and acceleration recorded by the navigation system are recorded as navigation data;
[0053] The following steps are involved:
[0054] Step 1: Convert the odometer data and navigation data to the global coordinate system to obtain the representation of the relevant mobile data information in the global coordinate system;
[0055] Step 2: Analyze and calculate the odometer data and navigation data represented in the global coordinate system to obtain analysis data; in the step 2, the odometer data and navigation data represented in the global coordinate system are analyzed and calculated by a data coupling prediction algorithm to obtain analysis data.
[0056] The step 2 includes the following sub-steps:
[0057] 2-1: Set state variables;
[0058] 2-2: Using the odometer data iteration data as the motion update state quantity corresponding to the state variable;
[0059] 2-3: Using the secondary time integration data of the navigation data as the corresponding correction variable;
[0060] 2-4: superimposing the motion update state quantity and the correction variable according to the data coupling prediction algorithm to obtain analyzed data;
[0061] The data coupling prediction algorithm is as follows:
[0062] The motion state prediction X(K) is calculated as follows:
[0063] k is the current time;
[0064] The correction amount prediction Y(K) is calculated as follows:
[0065]
[0066] The first-order linearized state equation is used to solve the state transfer matrix Φ(k). The calculation formula is:
[0067]
[0068] The first-order linearized observation equation is used to solve the observation matrix H(K). The calculation formula is:
[0069]
[0070] The formula for calculating the covariance matrix prediction P(k|k-1) is:
[0071] P(k|k-1)=Φ(k)P(k-1|k-1)Φ T (k);
[0072] De-noising processing W(k), the calculation formula is:
[0073] W(k)=P(k|k-1)H T (k)(H(k)P(k|k-1)H T (k));
[0074] Find the state update X(k), the calculation formula is:
[0075] X(k)=X(k|k-1)+K(Y(k)-Y(k|k-1));
[0076] Covariance update P(k) is calculated as:
[0077] P(k)=(I n -K(k)H(k))P(k|k-1), where I n is an n-dimensional vector matrix, the value of n is determined by the dimensions of vectors X and Y. The above calculation steps are a calculation cycle, and this cycle is the data processing process at each moment.
[0078] Step 3: Get the robot's positioning information based on the analysis data.
[0079] Example 2
[0080] Reference Figure 1 and Figure 2 , based on Example 1, the step 1 includes the following sub-steps:
[0081] 1-1: Set the starting point (x0, y0, z0) as the origin of the global coordinate system, and use the right-hand rule to determine the positive directions of the x-axis, y-axis, and z-axis;
[0082] 1-2: The robot moves on a plane. The odometer on the robot starts to calculate the robot's motion increment information from the starting position (x0, y0, z0), including (Δx, Δy, Δz), where: Δx records the distance the robot has moved forward, Δz records the robot's rotation angle, and converts the data information to the global coordinate system; the coordinate system conversion library TF library of the robot operating system ROS platform is needed to convert the odometer data to the global coordinate system.
[0083] 1-3: The navigation system continuously outputs the robot's current state information, including the rotation angle w and acceleration a x and a y, convert the data information to the global coordinate system. Use the coordinate system conversion library TF library of the robot operating system ROS platform to convert the navigation data to the global coordinate system.
[0084] In step 2-1, the state variable is (x, y, z).
[0085] In step 2-2, the odometer data (Δx, Δy, Δz) is used as the motion update state quantity. After the odometer data (Δx, Δy, Δz) is iteratively calculated at the starting position (x0, y0, z0), it corresponds to the new motion update state quantity (x+Δx, y+Δy, z+Δz).
[0086] In step 2-3, the navigation data acceleration a x and a y The second time integration can obtain the motion information in the x and y directions (∫∫a x ,∫∫a y ), and the rotation direction angle w of the navigation data is the corresponding variable to be corrected.
[0087] In steps 2-4, the calculation steps of the data coupling prediction algorithm are as follows:
[0088] 2-4-1: Predict the motion state X(K) of K at the current moment. The calculation formula is as follows:
[0089] X(K)={(x0+Δx), (y0+Δy), (z0+Δz)};
[0090] k is the current time;
[0091] 2-4-2: Predict the correction amount Y(K) of K at the current moment. The calculation formula is as follows:
[0092] Y(K)={a x ,a y ,w};
[0093]
[0094] 2-4-3: First-order linearized state equation, solve the state transfer matrix Φ(k), the calculation formula is as follows:
[0095]
[0096] 2-4-4: First-order linearized observation equation, solve the observation matrix H(K), the calculation formula is as follows:
[0097]
[0098] 2-4-5: Find the covariance matrix prediction P(k|k-1), the calculation formula is as follows:
[0099] P(k|k-1)=Φ(k)P(k-1|k-1)Φ T (k);
[0100] 2-4-6: Noise removal processing W(k), the calculation formula is as follows:
[0101] W(k)=P(k|k-1)H T (k)(H(k)P(k|k-1)H T (k))
[0102] 2-4-7: Find the state update X(k), the calculation formula is as follows:
[0103] X(k)=X(k|k-1)+K(Y(k)-Y(k|k-1))
[0104] 2-4-8: Covariance update P(k), the calculation formula is as follows:
[0105] P(k)=(I n -K(k)H(k))P(k|k-1), where I n is a 3D vector matrix. The value of n is determined by the dimensions of vectors X and Y. The above 8 steps are a calculation cycle, and this cycle is the data processing process at each moment. Through the displacement correction obtained by the above data coupling analysis algorithm, the robot positioning is more accurate and is not affected by the interference and influence of indoor and outdoor environments.
[0106] In step 3, the processed data is output, and the output data is standardized to the standard data of the odometer under the robot operating system ROS platform nav_msgs / Odometry; the topic / odom_fusion is published on the robot operating system ROS platform.
Claims
1. A mobile robot positioning method based on multi-data analysis, characterized in that: The robot is provided with an odometer and a navigation system. The odometer is used to record the robot's forward distance and the robot's rotation angle, and the navigation system is used to record the robot's current rotation direction angle and acceleration; The robot's forward distance and the robot's rotation angle recorded by the odometer are recorded as odometer data; The robot's current rotation angle and acceleration recorded by the navigation system are recorded as navigation data; The following steps are involved: Step 1: Convert the odometer data and navigation data to the global coordinate system to obtain the representation of the relevant mobile data information in the global coordinate system; Step 2: Analyze and calculate the odometer data and navigation data represented in the global coordinate system to obtain analysis data; In the step 2, the odometer data and navigation data represented in the global coordinate system are analyzed and calculated by a data coupling prediction algorithm to obtain analysis data; The step 2 includes the following sub-steps: 2-1: Set state variables; in step 2-1, the state variables are (x, y, z); 2-2: Using the iterative data of the odometer data as the motion update state quantity corresponding to the state variable; In step 2-2, the odometer data (Δx, Δy, Δz) is used as the motion update state quantity, and the odometer data (Δx, Δy, Δz) corresponds to the new motion update state quantity (x+Δx, y+Δy, z+Δz) after iterative calculation at the starting position (x0, y0, z0); 2-3: The second time integration data of the navigation data is used as the corresponding correction variable; in step 2-3, the second time integration of the accelerations ax and ay of the navigation data can obtain the motion information in the x and y directions (∫∫ax, ∫∫ay), and the rotation direction angle w of the navigation data is the corresponding variable z to be corrected; 2-4: superimposing the motion update state quantity and the correction variable according to the data coupling prediction algorithm to obtain analyzed data; In steps 2-4, the calculation steps of the data coupling prediction algorithm are as follows: 2-4-1: Predict the motion state X(K) of K at the current moment. The calculation formula is as follows: X(K)={(x0+Δx), (y0+Δy), (z0+Δz)}; k is the current time; 2-4-2: Predict the correction amount Y(K) of K at the current moment. The calculation formula is as follows: Y(K)={ax,ay,w}; 2-4-3: First-order linearized state equation, solve the state transfer matrix Φ(k), the calculation formula is as follows: 2-4-4: First-order linearized observation equation, solve the observation matrix H(K), the calculation formula is as follows: 2-4-5: Find the covariance matrix prediction p(k), the calculation formula is as follows: P(k|k-1)=Φ(k)P(k-1|k-1)Φ T (k); 2-4-6: Noise reduction processing, the calculation formula is as follows: W(k)=P(k|k-1)H T (k)(H(k)P(k|k-1)H T (k)) 2-4-7: Calculate the status update using the following formula: X(k)=X(k|k-1)+K(Y(k)-Y(k|k-1)) 2-4-8: Covariance update, the calculation formula is as follows: P(k)=(I n -K(k)H(k))P(k|k-1), where I n is a 3D vector matrix; Step 3: Get the robot's positioning information based on the analysis data.
2. The mobile robot positioning method based on multi-data analysis according to claim 1 is characterized in that: The step 1 includes the following sub-steps: 1-1: Set the starting point (x0, y0, z0) as the origin of the global coordinate system, and use the right-hand rule to determine the positive directions of the x-axis, y-axis, and z-axis; 1-2: The robot moves on a plane. The odometer starts to calculate the robot's motion increment information from the starting position (x0, y0, z0), including (Δx, Δy, Δz), where: Δx records the distance the robot has moved forward, and Δz records the robot's rotation angle, and the data information is converted to the global coordinate system; 1-3: The navigation system outputs the current state information of the robot, including the rotation angle w and acceleration a x and a y , convert the data information to the global coordinate system.
