Intelligent robot sampling path optimization and field reconstruction method for environmental monitoring
By introducing probability weight and information density algorithms and greedy algorithms into the robot sampling path, combining the Kernel series and dynamic grid algorithms, the sampling path and field reconstruction are optimized, and the problem of long sampling time and low efficiency is solved, and sampling efficiency and accuracy are improved rate, reduce costs.
Patent Information
- Application Number
- CN202410961318.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing robot sampling methods generally have problems such as long sampling time, low sampling efficiency and excessive redundancy in sampling point settings, resulting in inaccurate environmental parameter field reconstruction results.
Using the intelligent robot sampling path optimization method, by dividing the sampling points into initial sampling points and additional sampling points, using the algorithm of probability weight and information density to select additional sampling points, combining the greedy algorithm to optimize the sampling path, and combining the Kernel series algorithm Field reconstruction with dynamic grid improvement algorithm.
The sampling efficiency and accuracy of the robot's environmental parameter field reconstruction system are improved, the cost is reduced, and the robot's learning ability to the environment is enhanced.
Smart Images

Figure CN119002332A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of environmental safety and sensor technology, and in particular relates to an intelligent robot sampling path optimization and field reconstruction method for environmental monitoring. Background Art
[0002] Health threats and property losses caused by deteriorating indoor environmental quality are emerging one after another. The reason is that people cannot effectively perceive the "invisible" dangerous substances diffused in the indoor environment, and cannot consciously find that they are in or about to enter a dangerous indoor environment. Environmental parameter field reconstruction technology can effectively make up for this physiological limitation, replace people to perceive and accurately quantify the fine distribution of sensitive environmental parameters, which is of great significance for ensuring indoor air quality, personnel health and safety.
[0003] Environmental parameter field reconstruction technology mainly consists of two stages: one is to obtain parameter data of some points in the environment, and the other is to use the field reconstruction algorithm to deduce and calculate based on these data. In theory, the environmental parameter information of any point in the room can be obtained through this technology. Using a robot platform with various environmental parameter sensor groups to sample a limited number of points indoors is the main technical means to obtain environmental parameter data. This technology has the following advantages: 1. In the event of a sudden leak of toxic and harmful gases, it can achieve rapid deployment of sensors; 2. The robot can move flexibly and can collect parameters of multiple points in the environment with a small number of sensors, which can avoid tedious sensor calibration; 3. It has strong real-time adaptability to the environment and can enter dangerous areas that people cannot reach.
[0004] However, in practical applications, most existing robot field reconstruction systems use a traversal scanning sampling method, that is, the robot moves along a fixed path and samples at preset points. This method cannot accurately restore a larger environmental parameter field when fewer sampling points are set, so more sampling points need to be set, which in turn leads to low robot sampling efficiency. In some dynamic environments, too long sampling time will also lead to distortion of field reconstruction results. How to ensure high accuracy of field reconstruction results while ensuring high efficiency of robot sampling is a difficult problem that needs to be solved for environmental parameter field reconstruction systems based on robots. Summary of the invention
[0005] The technical problem to be solved by the embodiments of the present invention is that the existing sampling methods based on robot platforms are generally scanning or traversing, which leads to long sampling time, low sampling efficiency and excessive redundancy of sampling point settings. A method for optimizing the sampling path and reconstructing a field of an intelligent robot for environmental monitoring is provided. In this method, the sampling points are divided into initial sampling points and additional sampling points. An algorithm based on probability weight and information density is used as an additional sampling point selection algorithm, a greedy algorithm is used as a sampling path optimization algorithm for sampling points, and a Kernel series algorithm combined with an improved algorithm of a dynamic grid is used as a field reconstruction algorithm. The sampling efficiency and accuracy of the robot environmental parameter field reconstruction system are improved and the cost is reduced.
[0006] An embodiment of the present invention provides an intelligent sampling path optimization and field reconstruction method applicable to a mobile robot platform, the method comprising the following steps:
[0007] Step S1: Initialize the position and number of iterations of the field reconstruction mobile robot; set the initial information of the sampling environment, sampling point selection algorithm parameters and field reconstruction algorithm parameters; set the sampling time t and the maximum number of iterations T max ; Set up empty arrays A, B and C;
[0008] Step S2: The robot is put into the environment and the algorithm is executed. Since there is no historical data reference, the initial sampling point of the robot is determined according to the size of the sampling area and the maximum step length of the robot's movement, and the sampling path is optimized by the greedy algorithm. The number of initial sampling points is set to N1, and the coordinates of all initial sampling points are stored in array A;
[0009] Step S3: The robot moves in the order of the coordinates in the array A. When the robot reaches the target position, the sensor group carried by the robot starts to collect the environmental parameter data of the position and calculates the average value within the collection time t. Then, the robot moves to the next target position until the sampling of all the coordinates stored in the array A is completed and the position information is updated.
[0010] Step S4: After completing the sampling, the robot stays at the last sampling point and uses the collected data to perform field reconstruction calculations. After that, the number of iterations is increased by 1, and the grid size parameter in the field reconstruction algorithm parameters is adjusted;
[0011] Step S5: fill missing values and normalize the data of the current field reconstruction, store the processing results in array B, call the sampling point selection algorithm, determine the coordinates of the next additional sampling points, and set the number of additional sampling points to N2;
[0012] Step S6: Determine whether array B and array C are both not empty. If yes, proceed to step S11; if no, proceed to step S7;
[0013] Step S7: Call the greedy algorithm to N 2个 Append sampling points for sorting, and append the sorted additional sampling points to the end of array A;
[0014] Step S8: The robot moves in the order of the newly added coordinates in A. When the robot reaches the target position, the sensor group carried by the robot starts to collect environmental parameter data at the position and calculates the average value within the collection time. Then, the robot moves to the next target position until the sampling of all coordinates stored in A is completed and the position information is updated.
[0015] Step S9: After completing the sampling, the robot stays at the last sampling point and uses all the collected data to perform field reconstruction calculations. Then, the number of iterations is increased by 1, and the grid size parameter in the field reconstruction algorithm parameters is adjusted;
[0016] Step S10: perform missing value filling and normalization processing on the field reconstruction data of this time, transfer the previous field reconstruction processing result stored in array B to array C, and then update the stored data of array B to the current processing result, and transfer to step S6;
[0017] Step S11: calling the stop condition algorithm to calculate the similarity index of the two arrays, and then initializing the array C;
[0018] Step S12: determine whether the stop condition is met; if yes, end the sampling; if no, proceed to step S5;
[0019] Further, the steps S2 and S7 include:
[0020] The calculation principle of the greedy algorithm is to calculate the Euclidean distance between the remaining coordinate points to be sampled and the current position of the robot, sort the coordinate points to be sampled according to the distance, and the point with the closest distance is the next sampling point;
[0021] Furthermore, the step S3 comprises:
[0022] The robot follows the formula Determine the average concentration during the acquisition time;
[0023] Among them, C n is the sensor reading; t is the sampling time; C t is the instantaneous concentration detected by the sensor.
[0024] Furthermore, the step S4 and step S8 include:
[0025] The field reconstruction algorithm used is an improved algorithm of the Kernel series algorithm combined with dynamic grid. The Kernel series algorithm can be divided into Kernel DM+V algorithm and Kernel DM+V / W algorithm according to whether there is wind in the sampling environment;
[0026] The Kernel DM+V algorithm only needs to use the sampling point coordinates and their corresponding sensors to collect data, and is suitable for light wind or no wind environments. The Kernel DM+V / W algorithm requires wind speed and direction information in addition to coordinates and sensor data, and is suitable for mechanical ventilation or natural wind environments.
[0027] The dynamic grid size adjustment strategy of the improved algorithm is: set the maximum grid size L max and the minimum mesh size L min , the initial grid size is L max , which is halved as the number of iterations increases until the minimum grid size L min .
[0028] Furthermore, the steps S5 and S10 include:
[0029] The missing value processing method of the field reconstruction result is the nearest neighbor interpolation method. The sampling point selection algorithm consists of a probability weight setting algorithm based on concentration data and a weight adjustment algorithm based on information density. Among them, the principle of the probability weight setting algorithm based on concentration data is as follows: set the concentration weight P for each grid point according to the data of array B, and then randomly select the points. The larger the concentration weight, the higher the probability of the grid point being selected.
[0030] The principle of the weight adjustment algorithm based on information density is as follows: To avoid repeated sampling, first, the weights of the sampled points, that is, the coordinate points that already exist in array A, are set to 0; secondly, according to the formula Adjust the weights of the remaining grid points;
[0031] Where P is the adjusted grid point weight, P0 is the original weight of the grid point calculated by the probability weight algorithm, α is the weight adjustment factor, and its value range is (0, 1), N is the number of sampled points in the circular influence range with the grid point as the center and the radius r. The optimal weight adjustment factor α and the optimal influence range radius r are obtained through simulation and actual experiments.
[0032] Furthermore, the step S11 includes:
[0033] The stopping condition algorithm consists of a grid adjustment algorithm and a root mean square error indicator algorithm. The grid adjustment algorithm is used to process the field reconstruction results of two grids with different sizes to the same size. The root mean square error indicator is used to measure the similarity of the two field reconstruction results. The root mean square error indicator is calculated according to the formula Calculate and obtain;
[0034] Among them, RMSE is the root mean square error index, n is the number of grid points, B ij is the calculation result of the last field reconstruction at the grid point (i, j), C ij is the calculation result of the current field reconstruction at the grid point (i, j);
[0035] Furthermore, the step S12 includes:
[0036] When any of the following conditions is met, the stop condition is considered met and sampling stops:
[0037] The first condition is to determine whether the maximum number of iterations T has been reached. max ;
[0038] The second condition is to determine whether the calculation result of the RMSE index reaches the stopping requirement, where the value of the stopping requirement is determined manually;
[0039] The beneficial effects of the above technical solution of the present invention are as follows:
[0040] In the above scheme, by combining the sampling point selection and path optimization method that combines the probability weight selection algorithm, the information density-based weight adjustment algorithm and the greedy algorithm with the robot platform, the robot field reconstruction system can be equipped with the ability to learn the environment, thereby improving the sampling efficiency and the accuracy of field reconstruction. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the technical solution of the present invention, the technical solution of the present invention is briefly introduced with reference to the accompanying drawings.
[0042] Figure 1 A schematic diagram of a method flow of an embodiment of the present invention;
[0043] Figure 2 This is a schematic diagram of a weight adjustment algorithm according to an embodiment of the present invention;
[0044] Figure 3 A schematic diagram of a field reconstruction algorithm according to an embodiment of the present invention;
[0045] Figure 4 A schematic diagram of a grid size adjustment strategy according to an embodiment of the present invention; DETAILED DESCRIPTION
[0046] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.
[0047] The technical problem to be solved by the present invention is that the existing sampling methods based on robot platforms are generally scanning or traversing, which leads to long sampling time, low sampling efficiency and excessive redundancy of sampling point settings. A method for optimizing the sampling path and field reconstruction of an intelligent robot for environmental monitoring is provided. In this method, the sampling points are divided into initial sampling points and additional sampling points. The algorithm based on probability weight and information density is used as the additional sampling point selection algorithm, the greedy algorithm is used as the sampling path optimization algorithm for the sampling points, and the improved algorithm of the Kernel series algorithm combined with the dynamic grid is used as the field reconstruction algorithm. The sampling efficiency and accuracy of the robot environmental parameter field reconstruction system are improved and the cost is reduced. The environmental parameters described in the present invention include but are not limited to indoor temperature, humidity, CO2, possible toxic and harmful gases and particulate matter of different particle sizes.
[0048] The technical solution of the present invention is further described in detail below through specific embodiments.
[0049] The embodiment of the present invention provides an intelligent sampling path optimization and field reconstruction method applicable to a mobile robot platform. Figure 1 FIG. 4 is a schematic diagram of the process of the method described in this embodiment. Figure 1 As shown, the method comprises the following steps:
[0050] Step S1: Initialize the position and number of iterations of the field reconstruction mobile robot; set the initial information of the sampling environment, sampling point selection algorithm parameters and field reconstruction algorithm parameters; set the sampling time t and the maximum number of iterations T max ; Set up empty arrays A, B and C;
[0051] Step S2: The robot is put into the environment and the algorithm is executed. Since there is no historical data reference, the initial sampling point of the robot is determined according to the size of the sampling area and the maximum step length of the robot's movement, and the sampling path is optimized by the greedy algorithm. The number of initial sampling points is set to N1, and the coordinates of all initial sampling points are stored in array A;
[0052] The calculation principle of the greedy algorithm described in this step and step S7 is to calculate the Euclidean distance between the remaining coordinate points to be sampled and the current position of the robot, sort the coordinate points to be sampled according to the distance, and the point with the closest distance is the next sampling point;
[0053] Step S3: The robot moves in the order of the coordinates in the array A. When the robot reaches the target position, the sensor group carried by the robot starts to collect the environmental parameter data of the position and calculates the average value within the collection time t. Then, the robot moves to the next target position until the sampling of all the coordinates stored in the array A is completed and the position information is updated.
[0054] In this step, the robot uses the formula Determine the average concentration during the acquisition time; where C n is the sensor reading; t is the sampling time; C t is the instantaneous concentration detected by the sensor.
[0055] Step S4: After completing the sampling, the robot stays at the last sampling point and uses the collected data to perform field reconstruction calculations. After that, the number of iterations is increased by 1, and the grid size parameter in the field reconstruction algorithm parameters is adjusted;
[0056] In this step and step S9, the calculation principle of the Kernel series field reconstruction algorithm is as follows: Figure 3 As shown in the figure, the algorithm can be divided into Kernel DM+V algorithm and Kernel DM+V / W algorithm according to whether there is wind in the sampling environment.
[0057] The Kernel DM+V algorithm only uses the sampling point coordinates and their corresponding sensors to collect data, and is suitable for light wind or no wind environment. The principle is as follows Figure 3 (A) shows; Kernel DM+V / W algorithm requires wind speed and direction information in addition to coordinates and sensor data, and is suitable for mechanical ventilation or natural wind environments. The principle is as follows Figure 3 (B)
[0058] In this step and step S9, the adjustment strategy of the grid size parameters is as follows: Figure 4 As shown, the specific strategy is: set the maximum grid size L max and the minimum mesh size L min , the initial grid size is L max , which is halved as the number of iterations increases until the minimum grid size L is reached. min .
[0059] Step S5: fill missing values and normalize the data of the current field reconstruction, store the processing results in array B, call the sampling point selection algorithm, determine the coordinates of the next additional sampling points, and set the number of additional sampling points to N2;
[0060] In this step, the missing value processing method of the field reconstruction result is the nearest neighbor interpolation method; the sampling point selection algorithm consists of a probability weight setting algorithm based on concentration data and a weight adjustment algorithm based on information density;
[0061] The principle of the probability weight setting algorithm based on concentration data is as follows: set the concentration weight P for each grid point according to the data in array B, and then randomly select the points. The grid points with larger concentration weights have higher probability of being selected.
[0062] The principle of the weight adjustment algorithm based on information density is as follows Figure 2As shown, first, the weights of the sampled points, that is, the coordinate points that already exist in array A, are set to 0; secondly, according to the formula Adjust the weights of the remaining grid points;
[0063] Where P is the adjusted grid point weight, P0 is the original weight of the grid point calculated by the probability weight algorithm, α is the weight adjustment factor, and its value range is (0, 1), N is the number of sampled points in the circular influence range with the grid point as the center and the radius r. The optimal weight adjustment factor α and the optimal influence range radius r are obtained through simulation and actual experiments.
[0064] Step S6: Determine whether array B and array C are both not empty. If yes, proceed to step S11; if no, proceed to step S7;
[0065] Step S7: Call the greedy algorithm to N 2个 Append sampling points for sorting, and append the sorted additional sampling points to the end of array A;
[0066] Step S8: The robot moves in the order of the newly added coordinates in A. When the robot reaches the target position, the sensor group carried by the robot starts to collect environmental parameter data at the position and calculates the average value within the collection time. Then, the robot moves to the next target position until the sampling of all coordinates stored in A is completed and the position information is updated.
[0067] Step S9: After completing the sampling, the robot stays at the last sampling point and uses all the collected data to perform field reconstruction calculations. Then, the number of iterations is increased by 1, and the grid size parameter in the field reconstruction algorithm parameters is adjusted;
[0068] Step S10: perform missing value filling and normalization processing on the field reconstruction data of this time, transfer the previous field reconstruction processing result stored in array B to array C, and then update the stored data of array B to the current processing result, and transfer to step S6;
[0069] Step S11: calling the stop condition algorithm to calculate the similarity index of the two arrays, and then initializing the array C;
[0070] In this step, the stop condition algorithm consists of a grid adjustment algorithm and a root mean square error index algorithm. The grid adjustment algorithm is used to process the field reconstruction results of two grids with different sizes to the same size. The root mean square error index is used to measure the similarity of the two field reconstruction results. The root mean square error index is calculated according to the formula Calculate and obtain;
[0071] Among them, RMSE is the root mean square error index, n is the number of grid points, B ijis the calculation result of the last field reconstruction at the grid point (i, j), C ij is the calculation result of the current field reconstruction at the grid point (i, j);
[0072] Step S12: determine whether the stop condition is met; if yes, end the sampling; if no, proceed to step S5;
[0073] In this step, whether the condition is satisfied is further determined by determining whether any of the following conditions is satisfied:
[0074] The first condition is to determine whether the maximum number of iterations T has been reached. max ;
[0075] The second condition is to determine whether the calculation result of the RMSE index reaches the stopping requirement, where the value of the stopping requirement is determined manually;
[0076] The effective effects of the present invention are as follows:
[0077] (1) The present invention designs an intelligent robot sampling path optimization and field reconstruction method for environmental monitoring. It provides a solution to the problems of long sampling time, low sampling efficiency and excessive redundancy of sampling point settings caused by the existing sampling methods based on robot platforms, which are generally scanning or traversing.
[0078] (2) The present invention designs a sampling point selection algorithm based on a probability weight setting algorithm based on concentration data and a weight adjustment algorithm based on information density, and combines the algorithm with a greedy algorithm for sampling path planning, which significantly reduces the number of sampling points required and the time consumption in the robot field reconstruction sampling process, performs well in terms of sampling efficiency and field reconstruction accuracy, and effectively improves the performance and flexibility of the robot field reconstruction system.
[0079] The above is a specific description of an embodiment of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for intelligent robot sampling path optimization and field reconstruction, It is characterized in that The method consists of the following steps: Step S1: Initialize the position and number of iterations of the field reconstruction mobile robot; set the initial information of the sampling environment, sampling point selection algorithm parameters and field reconstruction algorithm parameters; set the sampling time t and the maximum number of iterations T max ; Set up empty arrays A, B and C; Step S2: The robot is put into the environment and the algorithm is executed. Since there is no historical data reference, the initial sampling point of the robot is determined according to the size of the sampling area and the maximum step length of the robot's movement, and the sampling path is optimized by the greedy algorithm. The number of initial sampling points is set to N 1 , all initial sampling point coordinates are stored in array A; Step S3: The robot moves in the order of the coordinates in the array A. When the robot reaches the target position, the sensor group carried by the robot starts to collect the environmental parameter data of the position and calculates the average value within the collection time t. Then, the robot moves to the next target position until the sampling of all the coordinates stored in the array A is completed and the position information is updated. Step S4: After completing the sampling, the robot stays at the last sampling point and uses the collected data to perform field reconstruction calculations. After that, the number of iterations is increased by 1, and the grid size parameter in the field reconstruction algorithm parameters is adjusted; Step S5: Fill missing values and normalize the data for this field reconstruction, and store the processing results in array B. Call the sampling point selection algorithm to determine the coordinates of the next additional sampling points, and set the number of additional sampling points to N. 2 ; Step S6: Determine whether array B and array C are both not empty. If yes, proceed to step S11; if no, proceed to step S7; Step S7: Call the greedy algorithm to N 2个 Append sampling points for sorting, and append the sorted additional sampling points to the end of array A; Step S8: The robot moves in the order of the newly added coordinates in A. When the robot reaches the target position, the sensor group carried by the robot starts to collect environmental parameter data at the position and calculates the average value within the collection time. Then, the robot moves to the next target position until the sampling of all coordinates stored in A is completed and the position information is updated. Step S9: After completing the sampling, the robot stays at the last sampling point and uses all the collected data to perform field reconstruction calculations. Then, the number of iterations is increased by 1, and the grid size parameter in the field reconstruction algorithm parameters is adjusted; Step S10: perform missing value filling and normalization processing on the field reconstruction data of this time, transfer the previous field reconstruction processing result stored in array B to array C, and then update the stored data of array B to the current processing result, and transfer to step S6; Step S11: calling the stop condition algorithm to calculate the similarity index of the two arrays, and then initializing the array C; Step S12: Determine whether the stop condition is met; if yes, end the sampling; if no, proceed to step S5.
2. The sampling path optimization and field reconstruction method according to claim 1, It is characterized in that The steps S2 and S7 further include: The calculation principle of the greedy algorithm is to calculate the Euclidean distance between the remaining coordinate points to be sampled and the current position of the robot, sort the coordinate points to be sampled according to the distance, and the point with the closest distance is the next sampling point.
3. The sampling path optimization and field reconstruction method according to claim 1, It is characterized in that The step S3 further comprises: The robot follows the formula Determine the average concentration during the acquisition time; Among them, C n is the sensor reading; t is the sampling time; C t is the instantaneous concentration detected by the sensor.
4. The sampling path optimization and field reconstruction method according to claim 1, It is characterized in that The steps S4 and S8 further include: The field reconstruction algorithm used is an improved algorithm of the Kernel series algorithm combined with dynamic grid. The Kernel series algorithm can be divided into Kernel DM+V algorithm and Kernel DM+V / W algorithm according to whether there is wind in the sampling environment; The Kernel DM+V algorithm only needs to use the sampling point coordinates and their corresponding sensors to collect data, and is suitable for light wind or no wind environments. The Kernel DM+V / W algorithm requires wind speed and direction information in addition to coordinates and sensor data, and is suitable for mechanical ventilation or natural wind environments. The dynamic grid size adjustment strategy of the improved algorithm is: set the maximum grid size L max and the minimum mesh size L min , the initial grid size is L max , which is halved as the number of iterations increases until the minimum grid size L min .
5. The sampling path optimization and field reconstruction method according to claim 1, It is characterized in that The steps S5 and S10 further include: The missing value processing method of the field reconstruction result is the nearest neighbor interpolation method. The sampling point selection algorithm consists of a probability weight setting algorithm based on concentration data and a weight adjustment algorithm based on information density. Among them, the principle of the probability weight setting algorithm based on concentration data is as follows: according to the data of array B, the concentration weight P is set for each grid point, and then the point is randomly selected. The larger the concentration weight, the higher the probability of the grid point being selected. The principle of the weight adjustment algorithm based on information density is as follows: To avoid repeated sampling, first, the weights of the sampled points, that is, the coordinate points that already exist in array A, are set to 0; secondly, according to the formula Adjust the weights of the remaining grid points; Among them, P is the adjusted grid point weight, P 0 is the original weight of the grid point calculated by the probability weight algorithm, α is the weight adjustment factor, and its value range is (0, 1). N is the number of sampled points in the circular influence range with radius r and the grid point as the center. The optimal weight adjustment factor α and the optimal influence range radius r are obtained through simulation and actual experiments.
6. The sampling path optimization and field reconstruction method according to claim 1, It is characterized in that The step S11 further comprises: The stopping condition algorithm consists of a grid adjustment algorithm and a root mean square error indicator algorithm. The grid adjustment algorithm is used to process the field reconstruction results of two grids with different sizes to the same size. The root mean square error indicator is used to measure the similarity of the two field reconstruction results. The root mean square error indicator is calculated according to the formula Calculate and obtain; Among them, RMSE is the root mean square error index, n is the number of grid points, B ij is the calculation result of the last field reconstruction at the grid point (i, j), C ij is the calculation result of the current field reconstruction at the grid point (i, j).
7. The sampling path optimization and field reconstruction method according to claim 1, It is characterized in that The step S12 further comprises: When any of the following conditions is met, the stop condition is considered met and sampling stops: The first condition is to determine whether the maximum number of iterations T has been reached. max ; The second condition is to determine whether the calculation result of the RMSE indicator reaches the stopping requirement, wherein the value of the stopping requirement is determined manually.
Citation Information
Patent Citations
RRT mechanical arm trajectory planning method based on non-obstacle space probability potential field sampling
CN116117822A
Robot path planning algorithm based on cylinder heuristic sampling
CN116382309A
Cited By
Air detection method and system
CN121164550A
An air detection method and system
CN121164550B
