Outdoor Robot Control System Based on Integrated Positioning Method

Abstract

The present invention discloses an outdoor robot control system based on a fusion positioning method, which includes a positioning and navigation module, an interaction module, and an execution module. The positioning and navigation module respectively establishes data connections with the interaction module and the execution module, and the interaction module establishes a data connection with the execution module. By fusing RTK, UWB, and IMU units and utilizing the collaborative work of each unit, the present invention solves the problem of continuous precise positioning of traffic isolation robots in complex urban road environments. The present invention proposes to judge the road scene where the robot is located through a positioning signal attenuation model and realize the accurate switching of corresponding optimal positioning methods, achieving high-precision positioning of traffic isolation robots. By fusing the positioning signal attenuation map and the obstacle map into a two-factor environment map, the present invention generates an optimal path that not only avoids obstacles but also considers the positioning signal attenuation factor and conducts navigation, thereby ensuring the positioning accuracy and the accuracy of navigation.

Description

Technical Field

[0001] The present invention relates to the technical field of robots, specifically an outdoor robot control system based on a fusion positioning method. Background Technique

[0002] With the development of technology, mobile robots are increasingly widely used in the field of intelligent transportation. In a complex traffic environment, urban roads show the characteristics of complexity and three-dimensionality, involving the switching of multiple scenarios, which poses high requirements for the positioning and navigation of mobile robots.

[0003] Currently known related technologies, such as the self-moving traffic isolation robot, system and its lane isolation method proposed in the patent with the application number CN116356739A. The traffic isolation robot involved in this patent uses the longitude and latitude information calculated by an inertial navigation module to achieve positioning in a road environment, and compares the position information with the starting position or the target end point to perform path tracking and navigation. This positioning method has the following defects: its inertial navigation module calculates the position change by measuring acceleration and angular velocity, and its positioning error will continuously accumulate with the increase of time and driving distance, resulting in a gradual decrease in positioning accuracy. Moreover, when passing through speed bumps and potholed roads, it will cause deviations in the measurement data of the inertial navigation module, thereby affecting the positioning accuracy. Currently known related technologies, such as the indoor and outdoor positioning system and method based on RTK, UWB and INS proposed in the patent with the application number CN118089739A. This patent proposes to select different positioning methods by judging whether it is indoors or outdoors through the RTK satellite observation number judging unit. Because the RTK satellite observation number is affected by the change of satellite geometric distribution and signal propagation delay, for a complex urban traffic application environment, the lag and incorrect judgment of this single judgment mechanism are likely to lead to untimely switching and incorrect selection of positioning methods, and there is a risk of missing positioning signals and a significant decrease in accuracy. And this positioning system is only limited to obtaining positioning information, and no solution is proposed for fusing the positioning information of mobile robots with subsequent path planning and navigation. Summary of the Invention

[0004] The purpose of the present invention is to provide an outdoor robot control system based on a fusion positioning method to solve the problems raised in the above background technique.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An outdoor robot control system based on a fusion positioning method, including a positioning and navigation module, an interaction module and an execution module. The positioning and navigation module establishes data connections with the interaction module and the execution module respectively, and the interaction module establishes a data connection with the execution module.

[0006] Preferably, the positioning and navigation module includes a positioning sub-module, a map construction sub-module and a decision-making and navigation sub-module.

[0007] Preferably, the positioning sub-module includes an RTK unit, a UWB unit and an IMU unit. The RTK unit is connected to a fixed base station through a wireless network, receives differential signals, and uses a built-in high-precision algorithm to calculate and output a centimeter-level positioning result. The UWB unit is used as a positioning tag and communicates with multiple UWB base stations to obtain accurate distance information. The IMU unit integrates multiple sensors and provides real-time measurements of the robot's acceleration, angular velocity and magnetometer. When the RTK signal is interfered with or missing, it performs combined positioning with RTK and UWB; in the positioning sub-module, there is an IMU data processing formula, which is based on the acceleration, angular velocity and magnetometer measurements obtained by the IMU unit and processes through the trapezoidal integration method to obtain the position information of the robot. The calculation formula is as follows:

[0008] Integrate acceleration to obtain velocity. Using the trapezoidal integration method, the velocity update formula is:

[0009]

[0010] Where is the measured value of the acceleration in the direction at time , the sampling time interval is , assuming measurements are made at discrete time ;

[0011] Using the trapezoidal integration method, the displacement update formula is:

[0012]

[0013] Thus, the relative displacement data is obtained. It is necessary to fuse the magnetometer and gyroscope data to estimate the heading and use Kalman filtering for fusion, and finally obtain the heading angle ;

[0014]

[0015] State transition equation:

[0016]

[0017] Observation equation:

[0018]

[0019] Where, the angle between the projection of the magnetic field on the horizontal plane and the axis is set as , the heading angle obtained by integrating the gyroscope is , and the state variable is set in the fusion through Kalman filtering, where is the fused heading angle, and is the rate of change of the heading angle, is the process noise, is the measurement noise.

[0020] Preferably, the map construction sub-module includes a multi-line lidar unit and a binocular vision unit. The multi-line lidar unit is used to help the robot perceive the surrounding environment, and the binocular vision unit is used to capture and process stereo image information, providing key data for constructing the positioning signal attenuation model map and optimizing the active switching strategy of the positioning method. The two cooperate to achieve environmental perception, obstacle recognition, and map construction. In the map construction sub-module, the data set of the environmental perception part of the positioning signal attenuation model is:

[0021]

[0022] where, is the image frame provided by the binocular vision unit, used to detect obstacles and estimate their positions, is the point cloud data provided by the multi-line lidar, used to accurately measure the shape and height of obstacles, is the acceleration and angular velocity data provided by the IMU, used to estimate the motion state of the robot;

[0023] When the map construction sub-module classifies obstacles, it calculates the weight affecting the positioning signal strength through the characteristics of obstacles and signal attenuation. The basic formula is:

[0024]

[0025] where, is the obstacle classification weight, indicates the obstacle type, represents the weight calculation function, indicates the signal type, is the signal type corresponding weight coefficient, indicates the obstacle type for the signal type attenuation degree; Considering three common obstacle types in the urban road environment, trees T, buildings B, and metal obstacles M, the following formula can be used to calculate the weight:

[0026]

[0027] Thus, a positioning signal attenuation model is constructed

[0028]

[0029] where, is the transmitted signal strength, is the predicted received signal strength at position ; is the predicted attenuation function at position which depends on the obstacle classification weight and the signal propagation distance at that position; ;

[0030] The map construction sub-module determines a reasonable threshold according to the test environment and multiple tests to judge the road scene where the machine is located and select the corresponding optimal positioning method. Among them, in outdoor positioning judgment, when the outdoor received signal strength is within the threshold, the robot uses RTK / INS loose combination for positioning:

[0031]

[0032] In indoor positioning judgment, when the indoor received signal strength is within the threshold, the robot uses UWB / INS tight combination for positioning:

[0033]

[0034] In the indoor-outdoor transition area judgment, when neither the indoor nor the outdoor received signal is within the threshold, the robot uses a combination of IMU and lidar for positioning.

[0035] Preferably, the decision-making and navigation sub-module includes a ROS main control unit and an STM32 main control unit. The ROS main control unit serves as the upper-layer decision-making center, which is used to integrate multi-source sensor data, evaluate the positioning signal strength on different paths through the positioning signal energy formula, perform path planning according to the current environment and the target position, and generate navigation instructions. The STM32 main control unit serves as the underlying execution core, which is used to receive the navigation instructions from the ROS main control unit, convert them into motor control signals, drive the robot to move forward along the planned path, and monitor the motion state of the robot in real time, and feedback to the ROS main control unit for adjustment and optimization; in the decision-making and navigation sub-module, there are RTK / INS loose-coupling integrated navigation system and UWB / INS tight-coupling integrated navigation system. The state vector of the RTK / INS loose-coupling integrated navigation system is:

[0036]

[0037] wherein, is the position error, is the velocity error, is the attitude error, which can be specifically represented by one of Euler angle error or quaternion error, is the gyroscope bias, is the first-order Markov error of the gyroscope, and is the accelerometer error; The observation equation describes the relationship between the observed data and the system state, and the observation equation can be expressed as:

[0038]

[0039] The system observation vector of the Kalman filter is:

[0040]

[0041] The observation noise vector is a Gaussian white noise vector with zero mean. If the attitude is represented by Euler angles, then can be expressed as:

[0042]

[0043] where, is the position difference, is the velocity difference, is the attitude difference, , and are the noise components in the position difference, velocity difference, and attitude difference respectively;

[0044] is the measurement matrix, and the formula is as follows:

[0045]

[0046] The heating state vector of the UWB / INS tightly coupled integrated navigation system is:

[0047]

[0048] where, is the position error, is the velocity error, is the attitude error, which can be specifically represented by either Euler angle error or quaternion error, is the gyroscope bias, is the accelerometer error;

[0049] The observation equation of UWB is:

[0050]

[0051] The observation vector is:

[0052]

[0053] where the UWB base station position is , the robot's position is , the observation noise vector;

[0054] In the extended Kalman filter processing, the covariance prediction and state transition matrix formulas used are:

[0055]

[0056] where is the estimated covariance matrix at time is the process noise covariance matrix, , which is the Jacobian matrix of the function with respect to the state vector evaluated at

[0057]

[0058] where is the observation Jacobian matrix, is the observation noise covariance matrix, and finally the high-precision positioning information of the robot is obtained , and at the same time, the biases of the accelerometer and gyroscope can be estimated and corrected;

[0059] The positioning signal energy formula is specifically:

[0060]

[0061] where represents the set of signal types, is the weight coefficient of the signal type used to balance the contributions of different signal types to the energy map, is the classification weight of the obstacle type at position for the signal type , is the reference signal strength of the signal type ;

[0062] The decision-making and navigation sub-module overlaps and fuses the road obstacle map and the positioning signal attenuation model map to generate an optimal path that avoids obstacles and takes into account the positioning signal attenuation factor and is used for navigation. The optimization function it constructs is:

[0063]

[0064] where represents the coordinate sequence of the path, and is a weight coefficient, which is important for balancing the importance of path safety and positioning accuracy. The overlapping cost of the path and the obstacle map is the cost of the attenuation of the positioning signal on the path.

[0065] Preferably, the interaction module includes a real-time monitoring sub-module and a voice broadcast communication sub-module. The real-time monitoring sub-module captures and monitors the road environment where the robot is located in real time through a high-definition camera, provides on-site monitoring records, provides real-time image information for the map construction and navigation decision-making of the robot, and assists in positioning and path planning. The voice broadcast communication sub-module is used to broadcast the traffic conditions according to the real-time traffic data or the robot's own perception information, and is also used to emit warning sounds or voice prompts.

[0066] Preferably, the execution module includes a chassis motor sub-module, a telescopic arm sub-module, and a warning sub-module.

[0067] Preferably, the chassis motor sub-module includes a first motor unit, a first motor controller unit, and a first transmission unit. The first motor unit is used to convert electrical energy into mechanical energy to provide power for the movement of the robot. The first motor controller unit is used to receive movement instructions and control the rotation speed, steering, and braking of the first motor unit. The first transmission unit is used to transmit the power of the first motor unit to the drive wheels to make the robot move.

[0068] Preferably, the telescopic arm sub-module includes a second motor unit, a second motor controller unit, a second transmission unit, a telescopic arm body unit, and a safety device unit. The second motor unit is used to convert electrical energy into mechanical energy and provide power for the telescopic arm body unit through the second transmission unit. The telescopic arm body unit is used to form a physical isolation. The second motor controller unit is used to receive movement instructions and control the rotation speed, steering, and braking of the second motor unit.

[0069] Preferably, the warning sub-module includes a warning light unit, a high-decibel alarm unit, and a control circuit unit. The warning light unit and the high-decibel alarm unit emit warning signals in a dual way of vision and hearing under the action of the control circuit unit.

[0070] Compared with the prior art, the beneficial effects of the present invention are as follows: By integrating RTK, UWB, and IMU units and using the collaborative work of each unit, the present invention solves the problem of continuous precise positioning of traffic isolation robots in complex urban road environments; The present invention proposes to judge the road scene where the machine is located through a positioning signal attenuation model and realize the accurate switching of the corresponding optimal positioning method, achieving high-precision positioning of traffic isolation robots; By fusing the positioning signal attenuation map and the obstacle map into a two-factor environment map, the present invention generates an optimal path that avoids obstacles and takes into account the positioning signal attenuation factor and navigates, thereby ensuring the positioning accuracy and the accuracy of navigation. Brief Description of the Drawings

[0071] Figure 1 is the system structure block diagram of the present invention;

[0072] Figure 2 is the structure block diagram of the positioning and navigation module of the present invention;

[0073] Figure 3 is the structure block diagram of the interaction module of the present invention;

[0074] Figure 4 is the structure block diagram of the execution module of the present invention;

[0075] Figure 5 is the method flow chart of the present invention;

[0076] Figure 6 is the flow chart for switching the fusion positioning method;

[0077] Figure 7 is the schematic diagram of road positioning and working environment;

[0078] Figure 8 is the dual-factor fusion and path planning map.

[0079] In the figure: 1. Positioning and navigation module; 11. Positioning sub-module; 111. RTK unit; 112. UWB unit; 113. IMU unit; 12. Map construction sub-module; 121. Multi-line lidar unit; 122. Binocular vision unit; 13. Decision-making and navigation sub-module; 131. ROS main control unit; 132. STM32 main control unit; 2. Interaction module; 21. Real-time monitoring sub-module; 22. Voice broadcast and communication sub-module; 3. Execution module; 31. Chassis motor sub-module; 311. First motor unit; 312. First motor controller unit; 313. First transmission unit; 32. Telescopic arm sub-module; 321. Second motor unit; 322. Second motor controller unit; 323. Second transmission unit; 324. Telescopic arm body unit; 325. Safety device unit; 33. Warning sub-module; 331. Warning light unit; 332. High-decibel alarm unit; 333. Control circuit unit. Detailed Description of the Preferred Embodiment

[0080] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0081] Please refer to the attached Figure 1-Appendix Figure 8 An embodiment provided by the present invention: An outdoor robot control system based on a fusion positioning method, including a positioning and navigation module 1, an interaction module 2, and an execution module 3. The positioning and navigation module 1 respectively establishes data connections with the interaction module 2 and the execution module 3, and the interaction module 2 establishes a data connection with the execution module 3; the positioning and navigation module 1 includes a positioning sub-module 11, a map construction sub-module 12, and a decision-making and navigation sub-module 13; the positioning sub-module 11 includes an RTK unit 111, a UWB unit 112, and an IMU unit 113. The RTK unit 111 is connected to a fixed base station through a wireless network, receives differential signals, and uses a built-in high-precision algorithm to calculate and output a centimeter-level positioning result. The UWB unit 112 serves as a positioning tag and communicates with multiple UWB base stations to obtain accurate distance information. The IMU unit 113 integrates multiple sensors and provides real-time measurements of the robot's acceleration, angular velocity, and magnetometer. When the RTK signal is interfered with or missing, it performs combined positioning with RTK and UWB; in the positioning sub-module 11, an IMU data processing formula is included. The IMU data processing formula is based on the acceleration, angular velocity, and magnetometer measurements obtained by the IMU unit 113 and is processed by the trapezoidal integration method to obtain the position information of the robot. The calculation formula is as follows:

[0082] Integrate acceleration to obtain velocity. Using the trapezoidal integration method, the velocity update formula is:

[0083]

[0084] where is the measured value of acceleration in the direction at time , the sampling time interval is , assuming measurements are made at discrete time ;

[0085] Using the trapezoidal integration method, the displacement update formula is:

[0086]

[0087] Thus, relative displacement data is obtained. It is necessary to fuse the magnetometer and gyroscope data to estimate the heading and use Kalman filtering for fusion, and finally obtain the heading angle ;

[0088]

[0089] State transition equation:

[0090]

[0091] Observation equation:

[0092]

[0093] Among them, the angle between the projection of the magnetic field on the horizontal plane and the axis is set as , the heading angle obtained by integrating the gyroscope is , in the fusion through Kalman filtering, the state variable is set, where is the fused heading angle, and is the rate of change of the heading angle, is the process noise, is the measurement noise; The map construction sub-module 12 includes a multi-line lidar unit 121 and a binocular vision unit 122. The multi-line lidar unit 121 is used to help the robot perceive the surrounding environment, and the binocular vision unit 122 is used to capture and process stereo image information, providing key data for constructing the positioning signal attenuation model map and optimizing the active switching strategy of the positioning method. The two cooperate to achieve environmental perception, obstacle recognition and map construction. In the map construction sub-module 12, the data set of the environmental perception part of the positioning signal attenuation model is:

[0094]

[0095] Among them, is the image frame provided by the binocular vision unit 122, which is used to detect obstacles and estimate their positions, is the point cloud data provided by the multi-line lidar, which is used to accurately measure the shape and height of obstacles, is the acceleration and angular velocity data provided by the IMU, which is used to estimate the motion state of the robot;

[0096] When the map construction sub-module 12 classifies obstacles, it calculates the weight affecting the positioning signal strength through the characteristics of obstacles and signal attenuation. The basic formula is:

[0097]

[0098] Among them, is the obstacle classification weight, indicates the obstacle type, represents the weight calculation function, represents the signal type, is the weight coefficient corresponding to the signal type , indicates the obstacle type for the signal type attenuation degree; Considering three common obstacle types in the urban road environment, trees T, buildings B, and metal obstacles M, the following formula can be used to calculate the weight:

[0099]

[0100] Thus, a positioning signal attenuation model is constructed.

[0101]

[0102] Wherein, is the transmitted signal strength, is the predicted received signal strength at position , is the predicted attenuation function at position , which depends on the obstacle classification weight and the signal propagation distance at this position and the signal propagation distance ;

[0103] The map construction sub-module 12 determines a reasonable threshold according to the test environment and multiple tests to judge the road scene where the machine is located and select the corresponding optimal positioning method. Among them, in outdoor positioning judgment, when the outdoor received signal strength is within the threshold, the robot uses RTK / INS loose combination for positioning:

[0104]

[0105] In indoor positioning judgment, when the indoor received signal strength is within the threshold, the robot uses UWB / INS tight combination for positioning:

[0106]

[0107] In the judgment of the indoor-outdoor transition area, when neither the indoor nor the outdoor received signal is within the threshold, the robot uses a combination of IMU and lidar for positioning; The decision-making and navigation sub-module 13 includes a ROS main control unit 131 and an STM32 main control unit 132. The ROS main control unit 131 serves as the upper-layer decision-making center, which is used to integrate multi-source sensor data, evaluate the positioning signal strength on different paths through the positioning signal energy formula, plan the path according to the current environment and the target position, and generate navigation instructions. The STM32 main control unit 132 serves as the underlying execution core, which is used to receive the navigation instructions of the ROS main control unit 131, convert them into motor control signals, drive the robot to move forward along the planned path, and monitor the motion state of the robot in real time, and feedback it to the ROS main control unit 131 for adjustment and optimization; In the decision-making and navigation sub-module 13, there are RTK / INS loose coupling integrated navigation system and UWB / INS tight coupling integrated navigation system. The state vector of the RTK / INS loose coupling integrated navigation system is:

[0108]

[0109] Among them, is the position error, is the velocity error, is the attitude error, which can be specifically represented by either Euler angle error or quaternion error. is the gyroscope bias, is the first-order Markov error of the gyroscope, is the accelerometer error; The observation equation describes the relationship between the observed data and the system state, and the observation equation can be expressed as:

[0110]

[0111] The system observation vector of the Kalman filter is:

[0112]

[0113] The observation noise vector is a Gaussian white noise vector with zero mean. If the attitude is represented by Euler angles, then can be expressed as:

[0114]

[0115] Among them, is the position difference, is the velocity difference, is the attitude difference, , and are the noise components in the position difference, velocity difference, and attitude difference, respectively;

[0116] is the measurement matrix, and the formula is as follows:

[0117]

[0118] The heating state vector of the UWB / INS tightly coupled integrated navigation system is:

[0119]

[0120] Among them, is the position error, is the velocity error, is the attitude error, which can be specifically represented by either Euler angle error or quaternion error. is the gyroscope bias, is the accelerometer error;

[0121] The observation equation of UWB is:

[0122]

[0123] The observation vector is:

[0124]

[0125] Among them, the UWB base station location is , the robot location is , is the observation noise vector;

[0126] In the extended Kalman filter processing, the covariance prediction and state transition matrix formulas used are:

[0127]

[0128] Among them, is the estimated covariance matrix at time is the process noise covariance matrix, , and it is the Jacobian matrix of the function with respect to the state vector at

[0129]

[0130] Among them, is the observation Jacobian matrix, is the observation noise covariance matrix, and finally the high-precision positioning information of the robot is obtained, and at the same time, the biases of the accelerometer and gyroscope can be estimated and corrected;

[0131] The specific formula for the positioning signal energy is:

[0132]

[0133] Among them, represents the set of signal types, is the weight coefficient of the signal type , which is used to balance the contributions of different signal types to the energy map, is the classification weight of the obstacle type at the position for the signal type , is the reference signal strength of the signal type ;

[0134] The decision-making and navigation sub-module 13 overlaps and fuses the road obstacle map with the positioning signal attenuation model map to generate an optimal path that avoids obstacles and takes into account the positioning signal attenuation factor, and navigates. The optimization function it constructs is:

[0135]

[0136] Among them, represents the coordinate sequence of the path, and are weight coefficients used to balance the importance of path safety and positioning accuracy. The overlapping cost of the path and the obstacle map, is the cost of positioning signal attenuation on the path; The interaction module 2 includes a real-time monitoring sub-module 21 and a voice broadcast communication sub-module 22. The real-time monitoring sub-module 21 captures and monitors the road environment where the robot is located in real time through a high-definition camera, provides on-site monitoring records, provides real-time image information for the map construction and navigation decision-making of the robot, and assists in positioning and path planning. The voice broadcast communication sub-module 22 is used to broadcast the traffic conditions according to real-time traffic data or the robot's own perception information, and is also used to emit warning sounds or voice prompts; The execution module 3 includes a chassis motor sub-module 31, a telescopic arm sub-module 32, and a warning sub-module 33; The chassis motor sub-module 31 includes a first motor unit 311, a first motor controller unit 312, and a first transmission unit 313. The first motor unit 311 is used to convert electrical energy into mechanical energy to provide power for the movement of the robot. The first motor controller unit 312 is used to receive movement instructions and control the rotation speed, steering, and braking of the first motor unit 311. The first transmission unit 313 is used to transmit the power of the first motor unit 311 to the drive wheels to make the robot move; The telescopic arm sub-module 32 includes a second motor unit 321, a second motor controller unit 322, a second transmission unit 323, a telescopic arm body unit 324, and a safety device unit 325. The second motor unit 321 is used to convert electrical energy into mechanical energy and provide power for the telescopic arm body unit 324 through the second transmission unit 323. The telescopic arm body unit 324 is used to form a physical isolation. The second motor controller unit 322 is used to receive movement instructions and control the rotation speed, steering, and braking of the second motor unit 321; The warning sub-module 33 includes a warning light unit 331, a high-decibel alarm unit 332, and a control circuit unit 333. Under the action of the control circuit unit 333, the warning light unit 331 and the high-decibel alarm unit 332 emit warning signals in a dual way of vision and hearing.

[0137] Working principle: When using the present invention to achieve high-precision fusion positioning and navigation of a traffic isolation robot, the following steps are adopted: Step 1, data acquisition and preliminary processing; Step 2, scene judgment and positioning method selection; Step 3, outdoor positioning; Step 4, indoor positioning; Step 5, positioning in the indoor-outdoor transition area; Step 6, path planning and navigation; specifically:

[0138] Step 1, in the positioning sub-module 11, the IMU unit 113 uses the measured values of acceleration in the x, y, and z directions, angular velocity, and magnetometer measurement values obtained. In addition, for the mobile robot, the trapezoidal integration method is used to integrate the acquired acceleration and velocity data successively. First, the velocity change is calculated to obtain the velocities in each direction, and then the displacements in each direction are further deduced to obtain the relative displacement data; and the magnetometer and gyroscope data are fused by Kalman filtering to estimate the heading; finally, the heading angle is combined with the relative displacement vector calculated by the trapezoidal integration method to deduce the motion trajectory and relative position of the robot; the multi-line lidar unit 121 scans the surrounding environment to obtain point cloud data for accurately measuring the shape, height, and position information of obstacles; the binocular vision unit 122 captures the image information of the real-time scene, simulating the human eye vision principle, and provides data support for subsequent environmental perception and the construction of the positioning signal attenuation model map;

[0139] Step 2, the RTK unit 111 and the UWB unit 112 in the positioning sub-module 11 continuously obtain the positioning signal strength data; the map construction sub-module 12 uses the data collected by the multi-line lidar unit 121 and the binocular vision unit 122, combines the robot motion state information provided by the IMU unit 113, and constructs a positioning signal attenuation model map; according to the combination of the detected real-time positioning signal strength and the positioning signal attenuation model map, the road scene where the robot is located is judged, and the ROS main control unit 131 in the decision-making and navigation sub-module 13 receives the data from the positioning sub-module 11 and the map construction sub-module 12, and selects the corresponding optimal positioning method according to the scene judgment result; if in an outdoor environment, RTK / INS loose combination is used for positioning, and enter Step 3; if in an indoor environment, UWB / INS tight combination is used for positioning, and enter Step 4; if in the indoor-outdoor transition area, a combination of IMU and lidar is used for positioning, and enter Step 5;

[0140] Step 3: The RTK unit 111 establishes a connection with a fixed base station using wireless network communication, and receives differential signals containing precise position information and error correction parameters in real time. After being solved by the built-in high-precision algorithm, it outputs a positioning result with centimeter-level accuracy. When the RTK signal is blocked or interfered, the IMU unit 113 uses its own inertial navigation ability to perform loose combined positioning with the RTK. The STM32 main control unit 132 in the decision-making and navigation sub-module 13 takes the differences in position, speed, and attitude measured by the RTK unit 111 and the position, speed, and attitude solved by the INS as observation values and inputs them into the Kalman filter for integrated navigation, and outputs the corrected position information.

[0141] Step 4: The UWB unit 112, as a positioning tag, communicates and interacts with multiple UWB base stations by transmitting and receiving ultra-wideband signals, and measures and obtains the precise distance information from each base station to the tag. The IMU unit 113 provides continuous and high-precision motion state information. The ROS main control unit 131 in the decision-making and navigation sub-module 13 uses a neural network algorithm to perform intelligent error compensation on the UWB ranging data, and then obtains the high-precision positioning information of the robot based on the extended Kalman filter algorithm.

[0142] Step 5: The IMU unit 113 continuously provides attitude, speed, and relative displacement information. The multi-line lidar unit 121 scans the surrounding environment to generate point cloud data for constructing a local map and performing matching for precise positioning. The STM32 main control unit 132 in the decision-making and navigation sub-module 13 combines and processes the two sets of data to achieve precise positioning in the indoor-outdoor transition area.

[0143] Step 6: The ROS main control unit 131 integrates the high-precision positioning information from different positioning methods, and comprehensively considers the obstacle map generated by the map construction sub-module 12 and the energy map of the positioning signal attenuation model. Through a specific path planning algorithm and optimization function, it generates an optimal path that not only avoids obstacles but also takes into account the positioning signal attenuation factor. The STM32 main control unit 132 receives the navigation instructions generated by the ROS main control unit 131, converts them into specific motor control signals, and drives the mobile robot to move forward along the planned path. At the same time, the STM32 main control unit 132 is responsible for monitoring the motion state of the robot, such as speed, steering angle, etc., and feeds back this information to the ROS main control unit 131 in real time for necessary adjustment and optimization.

[0144] Among them, in the interaction module 2, the real-time monitoring sub-module 21 captures and monitors the road environment where the mobile robot is located in real time through a high-definition camera. The monitored images and relevant environmental information are transmitted to the positioning and navigation module 1 to provide real-time data support for map construction, positioning, and path planning. At the same time, it also retains the monitoring records of the accident scene for traffic police officers, facilitating subsequent accident handling. The voice broadcast and communication sub-module 22 conveys the current traffic conditions, such as road closures, diversion information, traffic congestion, etc., to surrounding people according to real-time traffic data, the robot's own perception information, and the operating status of the positioning and navigation module 1. When the robot detects potential dangerous situations, such as pedestrians illegally crossing the isolation area, vehicles approaching the isolation belt, etc., it immediately emits a warning sound or voice prompt to alert relevant personnel to pay attention to safety; in the execution module 3, the first motor unit 311 of the chassis motor sub-module 31 is used to convert electrical energy into mechanical energy to provide power for the movement of the robot. The first motor controller unit 312 is used to receive motion instructions and control the rotation speed, steering, and braking of the first motor unit 311. The first transmission unit 313 is used to transmit the power of the first motor unit 311 to the drive wheels to make the robot move; the second motor unit 321 of the telescopic arm sub-module 32 is used to convert electrical energy into mechanical energy and provide power for the telescopic arm body unit 324 through the second transmission unit 323. The telescopic arm body unit 324 is used to form a physical isolation. The second motor controller unit 322 is used to receive motion instructions and control the rotation speed, steering, and braking of the second motor unit 321; in the warning sub-module 33, the warning light unit 331 and the high-decibel alarm unit 332 emit warning signals in a dual way of vision and hearing under the action of the control circuit unit 333;

[0145] As Figure 7 shown, the traffic isolation robot passes through three different road scenarios from the starting point to the end of the isolation area, namely the outdoor area, the indoor-outdoor transition area, and the indoor area. In the outdoor area, the space above the traffic isolation robot is open and there are no obstacles; in the indoor-outdoor transition area, discontinuous obstacles such as tree shades and pedestrian overpasses will intermittently appear above the robot; in the indoor area, continuous and uninterrupted obstacles such as viaducts or tunnels will appear above the robot; the positioning and navigation are carried out according to the above six steps, and the scene is judged according to the received signal strength detected and predicted in real time:

[0146] Outdoor positioning judgment, when the outdoor received signal strength is within the threshold, the robot uses the RTK / INS loose combination for positioning;

[0147]

[0148] Indoor positioning judgment, when the indoor received signal strength If it is within the threshold, the robot uses UWB / INS tightly coupled for positioning;

[0149]

[0150] Judgment of the indoor-outdoor transition area. When the received signals indoors and outdoors are both outside the threshold, the robot uses the combination of IMU and lidar for positioning;

[0151] During the process of the traffic isolation robot moving from the starting point to the end point of the isolation area, RTK / INS loose integration is first used for positioning. First, it passes through the unobstructed outdoor area. When it is predicted that it is about to enter the indoor-outdoor transition area, the robot needs to switch at the positioning method switching node 1 when the RTK signal has not completely failed but has started to decay, to ensure that there is enough time to complete the smooth transition of the positioning mode; the last obtained RTK coordinates before the switch will be used as the initial coordinates, which is the benchmark for subsequent IMU and lidar combined positioning; the IMU will start calculating the relative motion trajectory and attitude change of the robot based on this initial coordinate and the acceleration and angular velocity information measured by itself. In the specific implementation, it should be noted that during the driving in the indoor-outdoor transition area, the RTK unit 111 carried by the robot still continuously receives the latitude and longitude coordinates, and judges whether the received positioning information is valid according to the number of satellites in view; when the number of satellites in view is greater than 20, it is considered that the positioning information is valid and can be used to correct the IMU and lidar combined positioning; then, passing through the indoor-outdoor transition area, when it is predicted that it is about to enter the indoor area, the robot needs to switch at the positioning method switching node 2 when the RTK signal has not completely failed but has started to decay significantly, to ensure that there is enough time to complete the smooth transition of the positioning mode; finally, entering the indoor area under continuous occlusion of the urban viaduct, at this time, UWB and INS integrated positioning is adopted; Reasonable layout of UWB base stations in the indoor area is the key to achieving high-precision positioning; The layout of UWB base stations needs to follow certain principles. First, it is necessary to ensure full coverage of the indoor area. According to the size, shape and obstacle distribution of the indoor space, the base stations are evenly distributed on the surrounding walls or ceilings. The distance between adjacent base stations should be optimized according to the actual environment, generally between 5-10 meters, to ensure that the robot can receive the signals of at least three base stations at any position in the indoor area to achieve triangulation positioning; At the same time, multiple UWB tags are equipped for the robot. The tags should be installed at stable positions of the robot to avoid being blocked by other components to ensure the stability of signal transmission; The distance between the tag and the base station is measured by sending and receiving ultra-wideband pulse signals; INS continues to play a role in the indoor area. It can provide the acceleration and angular velocity information of the robot in real time and calculate the relative motion of the robot; Integrating UWB and INS for positioning, through the data fusion algorithm, combining the high-precision position information of UWB and the continuous motion information of INS, can effectively make up for the deficiencies of a single sensor; For example, when the UWB signal is temporarily lost due to certain reasons during the movement of the robot, INS can maintain the continuity of positioning, and when the UWB signal is stable, it can correct the accumulated errors of INS in time, so as to achieve high-precision and stable positioning of the robot in the indoor area until it successfully reaches the end point of the isolation area;

[0152] Such as Figure 8Schematic diagram of dual-factor fusion and path planning shown, which contains various information for path planning and navigation of traffic isolation robots; based on environmental perception, through the fusion analysis of multi-sensor data, an obstacle map and a positioning signal attenuation model energy map are constructed, and the obstacle information obtained from environmental perception and obstacle classification, as well as the signal strength and weight information of different regions obtained from the construction of the positioning signal attenuation model, are fused. The dual-factor fusion environment map obtained consists of squares, and each square represents a regional unit; some squares are marked as "Z", and these squares indicate obstacle areas that the robot needs to avoid during travel to ensure safe passage; the squares marked as "2" indicate that the signal attenuation in this area is relatively strong; this means that in these areas, the intensity of positioning signals such as RTK signals and UWB signals is relatively weak, which may have a certain impact on the positioning accuracy of the robot. The squares marked as "1" indicate that the signal attenuation in this area is relatively weak, and the positioning signals are relatively strong in these areas, which is more conducive to the robot obtaining accurate positioning information; after obtaining the dual-factor fusion environment map, path planning needs to be carried out according to the map information to generate an optimal path. This path is the optimal path generated by the system after obtaining high-precision positioning information from different positioning methods such as RTK / INS, UWB / INS, and IMU + lidar combination, considering the obstacle map and the positioning signal attenuation model energy map comprehensively; as Figure 8 The path planning that comprehensively considers obstacles and positioning signal attenuation factors shown is the solid line. This path starts from the starting point. First, the path avoids the nearby obstacle areas, that is, the "Z" squares, and at the same time tries to choose the areas with relatively weak signal attenuation, that is, the "1" squares or relatively weak, that is, the "2" squares, to travel; for example, near the starting point, the path extends towards the upper left, avoiding the obstacles below and on the left and the areas with relatively strong signal attenuation, and then winds along the areas with relatively weak signal attenuation and finally reaches the end point; the planning of this path fully considers the avoidance of obstacles and the attenuation of positioning signals, aiming to ensure that the robot can safely avoid obstacles during travel and obtain relatively good positioning signals, so as to ensure the positioning accuracy and the accuracy of navigation; in contrast, the D algorithm path planning that only considers obstacles, that is, the dotted line, this path is planned only based on the obstacle map using the D algorithm, only considering the factor of avoiding obstacles and not considering the attenuation of positioning signals. This path starts from the starting point and extends along the direction of avoiding obstacles, but it can be seen that it passes through many areas with relatively strong signal attenuation, that is, the "2" squares; for example, in the middle part, the path directly passes through an area with relatively strong signal attenuation, which may lead to a decrease in the positioning accuracy of the robot in this area and increase the uncertainty and risk of navigation.

[0153] It is obvious to those skilled in the art that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, in any regard, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claimed invention.

Claims

1. An outdoor robot control system based on a fusion positioning method, comprising a positioning and navigation module (1), an interaction module (2) and an execution module (3), characterized in that: The positioning and navigation module (1) establishes data connections with the interaction module (2) and the execution module (3) respectively, and the interaction module (2) establishes a data connection with the execution module (3); the positioning and navigation module (1) includes a positioning sub-module (11), a map construction sub-module (12), and a decision-making and navigation sub-module (13); the positioning sub-module (11) includes an RTK unit (111), a UWB unit (112), and an IMU unit (113). The RTK unit (111) is connected to a fixed base station through a wireless network, receives differential signals, and uses a built-in high-precision algorithm to calculate and output a centimeter-level positioning result. The UWB unit (112) serves as a positioning tag and communicates with multiple UWB base stations to obtain accurate distance information. The IMU unit (113) integrates multiple sensors and provides real-time measurements of the robot's acceleration, angular velocity, and magnetometer. When the RTK signal is interfered with or missing, it performs combined positioning with RTK and UWB; in the positioning sub-module (11), an IMU data processing formula is included. The IMU data processing formula is based on the acceleration, angular velocity, and magnetometer measurements obtained by the IMU unit (113) and is processed by the trapezoidal integration method to obtain the position information of the robot; the map construction sub-module (12) includes a multi-line lidar unit (121) and a binocular vision unit (122). The multi-line lidar unit (121) is used to help the robot perceive the surrounding environment, and the binocular vision unit (122) is used to capture and process stereo image information, providing key data for constructing a positioning signal attenuation model map and optimizing the active switching strategy of the positioning method. The two work together to achieve environmental perception, obstacle recognition, and map construction. In the map construction sub-module (12), the dataset of the environmental perception part of the positioning signal attenuation model is: D e = {V t , L t , I t | t ∈ T} Among them, V t is an image frame provided by the binocular vision unit (122) for detecting obstacles and estimating their positions, L t is the point cloud data provided by the multi-line lidar for precisely measuring the shape and height of obstacles, I t is the acceleration and angular velocity data provided by the IMU for estimating the motion state of the robot; When the map construction sub-module (12) classifies obstacles, it calculates the weight of the impact on the positioning signal strength through the characteristics of obstacles and signal attenuation. The basic formula is: w c = f(C, S) w c = ∑ s k s · Attenuation(C, S) Among them, w c is the obstacle classification weight, C represents the obstacle type, f represents the weight calculation function, s represents the signal type, and K s is the weight coefficient corresponding to the signal type S, and Attenuation(C, S) represents the attenuation degree of the obstacle type C to the signal type S; considering three common obstacle types in the urban road environment, trees T, buildings B, and metal obstacles M, the following formula can be used to calculate the weight: w i = a i · s T + k GNSS，T · Attenuation(T, GNSS) + k UWB，T · Attenuation(T, UWB), i = T, B, M Thus, a positioning signal attenuation model is constructed Among them, S t is the transmission signal strength, is the predicted received signal strength at the position (x, y), and Attenuation(w(x, y), d(x, y)) is the predicted attenuation function at the position (x, y), which depends on the obstacle classification weight w(x, y) and the signal propagation distance d(x, y) at that position; The map construction sub-module (12) determines a reasonable threshold according to the test environment and multiple tests to judge the road scene where the machine is located and select the corresponding optimal positioning method. Among them, in outdoor positioning judgment, when the outdoor received signal strength is within the threshold, the robot uses RTK / INS loose combination for positioning: In indoor positioning judgment, when the indoor received signal strength is within the threshold, the robot uses the UWB / INS tightly coupled method for positioning: In the judgment of the indoor-outdoor transition area, when the received signals indoors and outdoors are both outside the threshold, the robot uses the combination of IMU and lidar for positioning; the positioning signal energy formula is specifically: where S represents the set of signal types, and k s is the weight coefficient of signal type S, which is used to balance the contributions of different signal types to the energy map, and w C(x，y)，s is the classification weight of obstacle type C(x, y) for signal type S at position (x, y), and S ref，s is the reference signal strength of signal type S; The decision-making and navigation sub-module (13) overlaps and fuses the road obstacle map and the positioning signal attenuation model map to generate an optimal path that avoids obstacles and takes into account the positioning signal attenuation factor and conducts navigation. The optimization function it constructs: J(P) = λ1·ObCost(P) + λ2·SACost(P) Where P represents the coordinate sequence of the path, λ1 and λ2 are weight coefficients used to balance the importance of path safety and positioning accuracy. ObCost(P) is the overlap cost of the path and the obstacle map, and SACost(P) is the cost of positioning signal attenuation on the path 2. The outdoor robot control system based on the fusion positioning method according to claim 1, wherein: The calculation formula for obtaining the position information of the robot through processing by the trapezoidal integration method is as follows: Integrate acceleration to obtain velocity. Using the trapezoidal integration method, the velocity update formula is: where a i，k is the measured value of the acceleration in the x, y, and z directions at time t k , and the sampling time interval is Δt = t k+1 - t k . It is assumed that the measurements are made at discrete times t k , k = 0, 1, 2, …; Using the trapezoidal integration method, the displacement update formula is: Thus, the relative displacement data (S x，k , S y，k , S z，k ) are obtained. It is necessary to fuse the magnetometer and gyroscope data to estimate the heading, and use Kalman filtering for fusion, and finally obtain the heading angle θ k ; θ m = arctan(m y，k , m x，k ) θ g，k+1 = θ g，k + ω z，k Δt State transition equation: Observation equation: z k = [10]x k + v k Among them, the angle between the projection of the magnetic field on the horizontal plane and the x-axis is set as θ m , and the heading angle obtained by integrating the gyroscope is θ g . When performing fusion through Kalman filtering, the state variable is set as where θ k is the fused heading angle, and is the rate of change of the heading angle, w k is the process noise, and v k is the measurement noise.

3. The outdoor robot control system based on the fusion positioning method according to claim 1, characterized in that: The decision-making and navigation sub-module (13) includes a ROS master unit (131) and an STM32 master unit (132). The ROS master unit (131) serves as the upper-layer decision-making center, which is used to integrate multi-source sensor data, evaluate the positioning signal strength on different paths through the positioning signal energy formula, perform path planning according to the current environment and target location, and generate navigation instructions. The STM32 master unit (132) serves as the lower-layer execution core, which is used to receive the navigation instructions from the ROS master unit (131), convert them into motor control signals, drive the robot to move along the planned path, and monitor the motion state of the robot in real time, and feedback it to the ROS master unit (131) for adjustment and optimization; In the decision-making and navigation sub-module (13), there is an RTK / INS loosely coupled integrated navigation system and a UWB / INS tightly coupled integrated navigation system. The state vector of the RTK / INS loosely coupled integrated navigation system is: Among them, δ p (k) is the position error, δ v (k) is the velocity error, φ(k) is the attitude error, and specifically, it can be represented by one of the Euler angle error or the quaternion error. β g (k) is the gyroscope bias, β a (k) is the first-order Markov error of the gyroscope, is the accelerometer error; the observation equation describes the relationship between the observation data and the system state, and the observation equation can be expressed as: Z(k) = H(k)·X(k)+V(k) The system observation vector Z(k) of the Kalman filter is: The observation noise vector V(k) is a zero-mean Gaussian white noise vector. If the attitude is represented by Euler angles, then V(k) can be expressed as: where, δ pRTK-INS is the position difference, δ vRTK-INS is the speed difference, φ RTK-INS is the attitude difference, v pi (k), v vi (k) and v φi (k), (i = 1, 2, 3) are the noise components in the position difference, speed difference and attitude difference respectively; H(k) is the measurement matrix, and the formula is as follows: The heating state vector of the UWB / INS tightly coupled integrated navigation system is: where, δ p (k) is the position error, δ v (k) is the velocity error, φ(k) is the attitude error, and specifically, it can be represented by either Euler angle error or quaternion error. β g (k) is the gyroscope bias, is the accelerometer error; The observation equation of UWB is: Z i，k = ||p k - b i || + v i，k The observation vector is: Among them, the position of the UWB base station is b i (i = 1, 2, …, n), the position of the robot is p k , V k is the observation noise vector; In the extended Kalman filter processing, the covariance prediction and state transition matrix formulas used are: where, P k-1|k-1 is the estimated covariance matrix at time k-1, Q k is the process noise covariance matrix, which is the Jacobian matrix of the function f(x, u) with respect to the state vector X at the location. P k|k = (I - K k ·H k )P k|k-1 Among them, H k is the observation Jacobian matrix, R k is the observation noise covariance matrix, and finally the high-precision positioning information P k|k-1 of the robot is obtained. At the same time, the biases of the accelerometer and gyroscope can also be estimated and corrected.

4. The outdoor robot control system based on the fusion positioning method according to claim 1, characterized in that: The interaction module (2) includes a real-time monitoring sub-module (21) and a voice broadcast and communication sub-module (22). The real-time monitoring sub-module (21) captures and monitors the road environment where the robot is located in real time through a high-definition camera, provides on-site monitoring records, provides real-time image information for the map construction and navigation decision-making of the robot, and assists in positioning and path planning. The voice broadcast and communication sub-module (22) is used to broadcast the traffic conditions according to the real-time traffic data or the robot's own perception information, and is also used to issue warning sounds or voice prompts.

5. The outdoor robot control system based on the fusion positioning method according to claim 1, characterized in that: The execution module (3) includes a chassis motor sub-module (31), a telescopic arm sub-module (32), and a warning sub-module (33).

6. The outdoor robot control system based on the fusion positioning method according to claim 5, characterized in that: The chassis motor sub-module (31) includes a first motor unit (311), a first motor controller unit (312), and a first transmission unit (313). The first motor unit (311) is used to convert electrical energy into mechanical energy to provide power for the movement of the robot. The first motor controller unit (312) is used to receive motion instructions and control the rotation speed, steering, and braking of the first motor unit (311). The first transmission unit (313) is used to transmit the power of the first motor unit (311) to the drive wheels to make the robot move.

7. The outdoor robot control system based on the fusion positioning method according to claim 5, characterized in that: The telescopic arm sub-module (32) includes a second motor unit (321), a second motor controller unit (322), a second transmission unit (323), a telescopic arm body unit (324), and a safety device unit (325). The second motor unit (321) is used to convert electrical energy into mechanical energy and provide power for the telescopic arm body unit (324) through the second transmission unit (323). The telescopic arm body unit (324) is used to form physical isolation. The second motor controller unit (322) is used to receive motion instructions and control the rotation speed, steering, and braking of the second motor unit (321).

8. The outdoor robot control system based on the fusion positioning method according to claim 5, characterized in that: The warning sub-module (33) includes a warning light unit (331), a high-decibel alarm unit (332), and a control circuit unit (333). Under the action of the control circuit unit (333), the warning light unit (331) and the high-decibel alarm unit (332) emit warning signals in a dual manner of vision and hearing.

Images