CN110909105A - Robot map construction method and system - Google Patents
Robot map construction method and system Download PDFInfo
- Publication number
- CN110909105A CN110909105A CN201911167361.4A CN201911167361A CN110909105A CN 110909105 A CN110909105 A CN 110909105A CN 201911167361 A CN201911167361 A CN 201911167361A CN 110909105 A CN110909105 A CN 110909105A
- Authority
- CN
- China
- Prior art keywords
- generation
- map
- particle
- particles
- processor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
A robot map construction method and system comprises the following steps: initializing particle pose and map data, collecting robot laser data to obtain a first generation of particles, and accordingly obtaining a first generation of sampled particle map; acquiring weighted particle information in the first generation of particles, sampling the first generation of particles and inheriting a first generation of sampled particle map to obtain a second generation of particles and a second generation of sampled particle map; acquiring weighted particle information in the second-generation particles, sampling the second-generation particles and inheriting a second-generation sampled particle map to obtain a third-generation particle map and a third-generation particle map; emptying the first generation of particles and updating the second generation of particles and the third generation of particles to obtain mapping data; judging whether the map building is finished or not; if yes, generating a complete grid map according to the map building data; if not, circularly executing the steps until the map building is completed. The invention improves the efficiency and the precision of robot map building.
Description
Technical Field
The invention relates to the technical field of robot laser scanning image processing, in particular to a robot map construction method and system.
Background
With the continuous development of electronic control and artificial intelligence technology, intelligent robots are widely applied to the fields of intelligent home, industrial manufacturing, electric power, petroleum and petrochemical industry and the like. Intelligent robots of various types are widely used in various scenes of production and life. In the prior art, a robot performs environment sensing operations by using a simultaneous localization and mapping technology, wherein sampling particles are redistributed according to a weight ratio in a resampling stage of map particles, when the number of the particles is large, a large amount of particle information needs to be copied, especially, a map carried by each particle needs to be completely copied from the particles with large values, mapping and localization efficiencies are seriously affected, only a small number of particles need to be used in order to improve mapping speed, and the reduction of the number of the particles adversely affects the mapping and localization accuracy. In the conventional technology, the information carried by each sampled particle needs to be completely copied, and particularly, the copying of the carried grid map takes a lot of time. The larger the number of particles adopted by Slam (simultaneous localization and mapping) during map building, the more particles need to be copied in a resampling stage, the more time is consumed for copying a large number of maps, and the real-time performance of map building is seriously influenced. With the increase of the construction environment map, more particles are theoretically needed to achieve a better mapping effect, and resampling of more particles takes more time, which causes that the mapping real-time performance under a large environment is extremely poor and is limited to the performance of hardware, and the mapping effect of the large environment is not ideal.
In summary, the prior art has the technical problems of low mapping efficiency and poor mapping accuracy.
Disclosure of Invention
In view of the technical problems of low mapping efficiency and poor mapping accuracy in the prior art, the present invention aims to provide a robot map construction method and system, which solve the technical problems of low mapping efficiency and poor mapping accuracy in the prior art.
In an embodiment of the present invention, a method for constructing a robot map includes: initializing particle pose and map data, collecting robot laser data to obtain a first generation of particles, and accordingly obtaining a first generation of sampled particle map; acquiring weighted particle information in the first generation of particles, sampling the first generation of particles and inheriting a first generation of sampled particle map to obtain a second generation of particles and a second generation of sampled particle map; acquiring weighted particle information in the second-generation particles, sampling the second-generation particles and inheriting a second-generation sampled particle map to obtain a third-generation particle map and a third-generation particle map; emptying the first generation of particles and updating the second generation of particles and the third generation of particles to obtain mapping data; judging whether the map building is finished or not; if yes, generating a grid map according to the map building data; if not, circularly executing the steps until the map building is completed.
In an embodiment of the present invention, the step of acquiring and filtering the laser data by the acquisition device to obtain the first-generation particles, and accordingly acquiring the pose data of the first-generation sampled particle map particles includes: the method comprises the steps of obtaining robot laser data through induction, and obtaining particle filtering processing information; processing the robot laser data according to the particle filtering processing information to obtain particle pose data; and initializing and processing the particle pose data into first-generation particles.
In an embodiment of the present invention, the step of obtaining weighted particle information in the first-generation particles, and accordingly sampling the first-generation particles and inheriting the first-generation sampled particle map to obtain the second-generation particles and the second-generation sampled particle map includes: processing the weighted particle information to obtain second-generation resampling data; sampling the first generation particles according to the second generation resampling data to obtain second generation particles; inheriting the first generation of sampled particle map according to the second generation of resampling data to obtain a second generation of sampled particle map; and obtaining second-generation particle map maintenance information according to the second-generation particles so as to maintain the second-generation sampled particle map.
In an embodiment of the present invention, the step of obtaining weighted particle information in the second-generation particles, and accordingly sampling the second-generation particles and inheriting the second-generation sampled particle map to obtain the third-generation particles and the third-generation particle map includes: processing the weighted particle information to obtain third generation resample data; sampling second-generation particles according to the third-generation resampling data pair to obtain third-generation particles; inheriting a second-generation sampled particle map according to the third-generation resampling data to obtain a third-generation particle map; and obtaining the maintenance information of the third-generation particle map according to the third-generation particles so as to maintain the third-generation particle map.
In one embodiment of the present invention, the step of emptying the first generation particles and updating the second generation particles and the third generation particles to obtain the mapping data includes: merging the first generation of sampled particle map and the second generation of sampled particle map; emptying the first generation particles; updating the second generation sampled particle map into a first generation sampled particle map; and updating the third generation particle map into a second generation sampled particle map.
In one embodiment of the present invention, a robot map construction system includes: the laser data processing device is used for initializing the position and the pose of the particles and map data, acquiring the laser data of the robot to obtain first generation particles, and accordingly acquiring a first generation sampled particle map; the second-generation resampling processor is used for acquiring weighted particle information in the first-generation particles, sampling the first-generation particles and inheriting a first-generation sampled particle map to obtain a second-generation particle map and a second-generation sampled particle map, and the second-generation resampling processor is connected with the laser processing device; the third-generation resampling processor is used for acquiring weighted particle information in the second-generation particles, sampling the second-generation particles and inheriting a second-generation sampled particle map so as to obtain a third-generation particle map and a third-generation particle map; the third-generation resampling processor is connected with the second-generation secondary sampling processor; the updating processor is used for emptying the first generation of particles and updating the second generation of particles and the third generation of particles so as to obtain image building data, and the updating processor is connected with the second generation resampling processor and the third generation resampling processor; the image building judgment processor is used for judging whether the image building is finished or not, and is connected with the updating processor; the grid map processor is used for generating a grid map according to the mapping data when mapping is completed and is connected with the mapping judgment processor; and the updating expansion processor is used for circularly executing the steps until the map building is finished when the map building is not finished, and is connected with the map building judgment processor.
In one embodiment of the present invention, a laser data processing apparatus includes: the induction filtering device is used for induction obtaining of laser data of the robot and obtaining of particle filtering processing information; the particle pose device is used for processing the laser data of the robot according to the particle filtering processing information to obtain particle pose data, and the particle pose device is connected with the induction filtering device; and the primary initialization processor is used for initializing and processing the particle pose data into first-generation particles and is connected with the laser data processing device.
In one embodiment of the present invention, a second generation resampling processor comprises: a second generation weighting processor for processing the weighted particle information to obtain second generation resample data; the second-generation particle processor is used for sampling the first-generation particles according to the second-generation resampling data to obtain second-generation particles, and the second-generation particle processor is connected with the first-generation weighting processor; the second-generation map processor is used for inheriting the first-generation sampled particle map according to the second-generation resampling data so as to obtain a second-generation sampled particle map, and the second-generation map processor is connected with the second-generation particle processor; and the second-generation maintenance processor is connected with the second-generation particle processor and the second-generation map processor.
In one embodiment of the present invention, a third generation resampling processor comprises: a second generation weighting processor for processing the weighted particle information to obtain third generation resample data; the third-generation particle processor is used for sampling second-generation particles according to third-generation resampling data to obtain third-generation particles, and the third-generation particle processor is connected with the second-generation weighting processor; the third-generation map processor is used for inheriting the second-generation sampled particle map according to the third-generation resampling data to obtain a third-generation particle map, and the third-generation map processor is connected with the third-generation particle processor; and the third generation maintenance processor is used for acquiring the third generation particle map maintenance information according to the third generation particles so as to maintain the third generation particle map, and is connected with the third generation map processor and the third generation particle processor.
In one embodiment of the present invention, an update processor includes: a merging processor to merge the first generation of sampled particle maps and the second generation of sampled particle maps; a clearing processor for clearing the first generation particles; the second generation updating processor is used for updating the second generation sampled particle map into a first generation sampled particle map and is connected with the merging processor; and the third generation updating processor is used for updating the third generation particle map into a second generation sampled particle map, and is connected with the second generation updating processor.
As described above, the object of the present invention is to provide a method and a system for constructing a robot map, so as to overcome the deficiencies of the prior art, in the present invention, the sampled particles and the information carried by the particles do not need to be completely copied, only the adopted particles need to be inherited, and the remaining particles are removed, thereby avoiding the adverse effects of the copying process of the grid map in the prior art, such as occupying a large amount of system running time and memory, and improving the real-time performance of map construction.
Meanwhile, under the condition that the constructed environment map is enlarged, for example, in order to achieve a better mapping effect, more time is occupied for resampling of multi-generation particles, the mapping instantaneity of the robot in a large environment is improved, the limitation of the mapping function by the performance of hardware is reduced, and the mapping effect of the robot is optimized.
In summary, the invention provides a robot map construction method and system, which solve the technical problems of low map construction efficiency and poor map construction accuracy in the prior art.
Drawings
Fig. 1 is a flowchart illustrating steps of a method for constructing a robot map according to the present invention.
Fig. 2 is a schematic diagram illustrating a specific flow of step S1 in fig. 1 according to an embodiment.
Fig. 3 is a flowchart illustrating a specific flow of step S2 in fig. 1 according to an embodiment.
Fig. 4 is a flowchart illustrating a specific flow of step S3 in fig. 1 according to an embodiment.
Fig. 5 is a flowchart illustrating a specific flow of step S4 in fig. 1 according to an embodiment.
Fig. 6 is a schematic diagram showing components of a robot map building system according to the present invention.
Fig. 7 is a schematic diagram showing the connection of specific components of the laser data processing apparatus 1 in fig. 6 in an embodiment.
Fig. 8 is a schematic diagram illustrating specific component connections of the second-generation resampling processor 2 in fig. 6 in an embodiment.
Fig. 9 is a schematic diagram showing the connection of the third generation resampling processor 3 in fig. 6 in an embodiment.
FIG. 10 is a diagram illustrating the connection of specific components of the update processor 4 of FIG. 6 in one embodiment.
Description of the element reference numerals
1 laser data processing device
2 second generation resampling processor
3 third generation resampling processor
4 updating processor
5 construct picture and judge the processor
6 grid map processor
7 updating extended processor
11 inductive filter device
12 laser data processing device
13 initial generation processor
21 generation weighting processor
22 second generation particle processor
23 second generation map processor
24-generation maintenance processor
31 second generation weighting processor
32 three generations particle processor
33 three-generation map processor
34 third generation maintenance processor
41 merging processor
42 flush processor
43 second generation updating processor
44 three generation updating processor
Description of step designations
Method steps S1-S7
Method steps S11-S13
Method steps S21-S24
Method steps S31-S34
Method steps S41-S44
Detailed Description
The following description of the embodiments of the present invention is provided for illustrative purposes, and other advantages and effects of the present invention will become apparent to those skilled in the art from the present disclosure.
Referring to fig. 1 to 10, it should be understood that the structures shown in the drawings are only used for understanding and reading the present disclosure, and are not used to limit the conditions of the present invention, which can be implemented, so that the present invention has no technical significance, and any structural modification, ratio change or size adjustment should still fall within the scope of the present invention without affecting the function and the achievable object of the present invention. In addition, the terms "upper", "lower", "left", "right", "middle" and "one" used in the present specification are for clarity of description, and are not intended to limit the scope of the present invention, and the relative relationship between the terms and the terms is not to be construed as a scope of the present invention.
Referring to fig. 1, fig. 1 is a flow chart illustrating steps of a method for constructing a robot map according to the present invention, as shown in fig. 1, a method for constructing a robot map includes S1, initializing pose and map data of particles, collecting laser data of the robot to obtain first generation particles, obtaining a first generation sampled particle map according to the first generation particles, in one embodiment, slam (simultaneous localization and mapping) is a technology of mobile robot positioning and navigation, which relates to the storage and management of grid maps, in one embodiment, the acquisition device 1 may include, but is not limited to, a lidar, a cell camera, a dual (multi) view camera, a depth camera, a millimeter wave radar, an ultrasonic radar, and the like, the laser radar has strong pattern-building directivity and high response speed, and the eye-catching camera can estimate the position, depth data and the like of a space point; s2, obtaining weighted particle information in the first-generation particles, and accordingly sampling the first-generation particles and inheriting the first-generation sampled particle map to obtain a second-generation particle and a second-generation sampled particle map, in an embodiment, in a resampling stage of particle data processing, for example, according to a particle weight, the sampled particles may generate second-generation particles, and a map maintained by the first-generation sampled particles may not need to be copied, and in an embodiment, the first-generation particles may be obtained in an inheritance manner, and the newly generated particles initialize a map maintained by themselves; s3, obtaining weighted particle information in the second-generation particles, and accordingly sampling the second-generation particles and inheriting the second-generation sampled particle map to obtain third-generation particles and a third-generation particle map, wherein in one embodiment, when resampling is performed on the second-generation particles, the sampled particles can generate third-generation particles, the map maintained by the second-generation sampled particles can be obtained in an inheriting manner without copying, and meanwhile, the map maintained by the second-generation sampled particles can be initialized by newly generated particles; s4, emptying the first generation particles and updating the second generation particles and the third generation particles to obtain mapping data; s5, judging whether the map building is finished; s6, if yes, generating a grid map according to the map building data; if not, S7, circularly executing the steps until the map building is completed; .
Please refer to fig. 2, which is a schematic diagram showing a specific flow step of step S1 in fig. 1 in an embodiment, as shown in fig. 2, the step S1 of acquiring first-generation sampled particle map particle pose data by acquiring and filtering laser data by an acquisition device to obtain first-generation particles includes S11, sensing to acquire robot laser data and acquiring particle filtering processing information, in an embodiment, when the robot is at an unknown position in, for example, an unknown environment, the robot can estimate, for example, a real-time pose of the robot while moving based on its own sensors including, but not limited to, an encoder, an IMU, a laser, a camera, etc. through, for example, slam technology, and then continuously expand and update the map to gradually construct a complete map of the environment; s12, processing the robot laser data according to the particle filter processing information to obtain the particle pose data, in an embodiment, the initialized particles may be the first generation, in this embodiment, the particle filter laser slam may initialize, for example, a given number of particles and information for maintenance thereof, etc. according to the set number of particles in the initialization stage. The key information maintained by each particle may be, for example, the map constructed and its pose in the map. In this embodiment, the particle pose may be initialized to, for example, (0,0,0), etc., and each cell of the map is initialized to, for example, an unknown state. The initially generated particles may be, for example, first generation particles; s13, initializing particle pose data as first generation particles, in one embodiment, the core idea of the particle filter algorithm is to approximate an integration operation by summation using, for example, a weighted sum of a series of random samples to approximate a posterior probability density function. In one embodiment, the slam of particle filtering approximates the posterior distribution of the robot pose using a large number of particles, each particle representing, for example, a sample, each particle carrying, for example, an environmental map, and in one embodiment, a removable storage medium readable by, for example, particle data, and in one embodiment, a portable storage medium is installed on a robot-owned processing component, such as a microprocessor, and has a map construction program stored thereon, which executes, with the map pose processor 12, a construction method for implementing the robot map, and in this embodiment, all or part of the steps for implementing the method embodiments described above can be performed by hardware associated with a computer program. The aforementioned computer program may be stored in a computer-readable portable storage medium. When executed, the program performs steps comprising the method embodiments described above; the map portable storage medium may include, but is not limited to, various media that can store program codes, such as ROM, RAM, magnetic or optical disk.
Referring to fig. 3, which is a schematic diagram illustrating a specific flow step of step S2 in fig. 1 in an embodiment, as shown in fig. 3, the step S2 of obtaining weighted particle information in a first generation of particles, sampling the first generation of particles accordingly and inheriting the first generation of sampled particle map to obtain a second generation of particles and a second generation of sampled particle map includes S21, processing the weighted particle information to obtain second generation resampled data; s22, sampling the first-generation particles according to the second-generation resampling data to obtain second-generation particles, in an embodiment, in a resampling stage, the laser slam based on, for example, particle filtering redistributes the sampled particles according to, for example, a weight ratio, and a probability that a particle with a larger weight is sampled is larger; s23, inheriting the first generation sampled particle map according to the second generation resampled data to obtain a second generation sampled particle map, in one embodiment, the duplication of the particle map is reduced under the condition that the number of used particles is not changed, and the optimization of the management of the particle map is a key technology for improving the mapping efficiency and precision; and S24, acquiring second-generation particle map maintenance information according to the second-generation particles so as to maintain the second-generation sampled particle map.
Referring to fig. 4, which is a schematic diagram illustrating a specific flow step of step S3 in fig. 1 in an embodiment, as shown in fig. 4, step S3 of obtaining weighted particle information in second-generation particles, sampling the second-generation particles and inheriting a second-generation sampled particle map to obtain third-generation particles and a third-generation particle map includes step S31 of processing the weighted particle information to obtain third-generation resampled data, and in an embodiment, after the third-generation particles are completely generated, merging a first-generation particle map onto a second-generation particle maintained map inheriting the second-generation particle map, and emptying all information maintained by the first-generation particles; s32, sampling the second-generation particles according to the third-generation resampling data to obtain third-generation particles, and S33 inheriting the second-generation sampled particle map according to the third-generation resampling data to obtain a third-generation particle map; s34, obtaining the maintenance information of the third generation particle map according to the third generation particles, and accordingly maintaining the third generation particle map; in one embodiment, the second-generation particles are updated to, for example, first-generation particles, and the third-generation particles are updated to second-generation particles, in this embodiment, after the third-generation particles are completely generated, the maintained particle set is equal to three generations, that is, the particle set includes particles generated in three different sampling stages, a map maintained by the first-generation particles is merged onto a map maintained by the second-generation particles inheriting the map, and information maintained by the first-generation particles is cleared, and the non-sampled particle set is not maintained any more, and information thereof is cleared directly, in one embodiment, the particle set maintained by the present invention includes three generations at most, that is, the particle set includes particles generated in three different sampling stages at most, so that occupation of the memory is reduced, in one embodiment, the update processor 32 may include a micro random access memory (abbreviated as RAM), and may further include a non-volatile memory (RAM), such as at least one portable disk storage. The map merge processor 32 and the update processor 43 can be, for example, general-purpose processors including a Central Processing Unit (CPU), a Network Processor (NP), and the like; the integrated circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, or discrete hardware components.
Referring to fig. 5, which is a schematic diagram illustrating a specific flow step of step S4 in fig. 1 in an embodiment, as shown in fig. 5, the step S4 of processing the map data of multiple stages by the map processor to generate the grid map includes, S41, merging the first-generation sampled particle map and the second-generation sampled particle map; s42, emptying the first generation particles; s43, updating the second generation sampled particle map into the first generation sampled particle map, S44, updating the third generation particle map into the second generation sampled particle map, in an embodiment, as the map is continuously updated and expanded, the re-sampling is continuously performed, and new particles are continuously generated. The maintained particle set contains three generations at most, thereby reducing the occupation of the memory. In this embodiment, after the resampling of the slam is completed, the newly generated map maintained by the particles is continuously updated along with the continuous input of the laser data. The maps maintained by the rest of the particle sets are not inserted with new laser data and are not updated.
Referring to fig. 6, which is a schematic diagram showing components of a robot map construction system of the present invention, as shown in fig. 6, a robot map construction system includes a laser data processing device 1, a second-generation resampling processor 2, a third-generation resampling processor 3, an updating processor 4, a mapping judgment processor 5, a grid map processor 6 and an updating expansion processor 7, the laser data processing device 1 is used to initialize a particle pose and map data, collect robot laser data to obtain a first-generation particle, and obtain a first-generation sampled particle map, in one embodiment, a slice (simultaneous localization and mapping) is called as a simultaneous localization and mapping, which is a mobile robot localization and navigation technology, and relates to grid-based storage and management, in one embodiment, the laser processing device 1 may use, for example, a laser radar, a unit camera, and storage and management, The system comprises a double (multi) view camera, a depth camera, a millimeter wave radar, an ultrasonic radar and the like, wherein the laser radar is strong in drawing establishing directivity and high in response speed, and the position of a space point, depth data and the like can be estimated by the aid of the view capturing camera; the second-generation resampling processor 2 is used for acquiring weighted particle information in the first-generation particles, sampling the first-generation particles and inheriting a first-generation sampled particle map to obtain a second-generation particle map and a second-generation sampled particle map, and the second-generation resampling processor 2 is connected with the laser processing device 1, in an embodiment, in a resampling stage of particle data processing, for example, according to particle weight values, the second-generation particles can be generated from the sampled particles, the map maintained by the first-generation sampled particles does not need to be copied, and in an embodiment, the second-generation resampling processor can acquire the weighted particle information in an inheriting manner, and meanwhile, the newly generated particles initialize the map maintained by the second-generation particles; the third generation resampling processor 3 is used for acquiring weighted particle information in the second generation particles, sampling the second generation particles and inheriting the second generation sampled particle map to obtain a third generation particle map and a third generation particle map; a third generation resampling processor 3 connected with the second generation resampling processor 2, an updating processor 4 for emptying the first generation particles and updating the second generation particles and the third generation particles to obtain mapping data, the updating processor 4 connected with the second generation resampling processor 2 and the third generation resampling processor 3, a mapping judgment processor 5 for judging whether mapping is completed, the mapping judgment processor 5 connected with the updating processor 4, a grid map processor 6 for generating a grid map according to the mapping data when mapping is completed, a grid map processor 6 connected with the mapping judgment processor 5, an updating extension processor 7 for circularly executing the above steps until mapping is completed when mapping is not completed, the updating extension processor 7 connected with the mapping judgment processor 5, in one embodiment, when second generation particles are resampled, the sampled particles can generate third generation particles, the second generation of maps maintained by the sampled particles can be acquired without copying, for example, in an inheritance mode, and meanwhile, the maps maintained by the second generation of maps can be initialized by newly generated particles.
Referring to fig. 7, which is a schematic diagram showing connection of specific components of the laser data processing apparatus 1 in fig. 6 in an embodiment, as shown in fig. 7, the acquisition apparatus 1 includes an inductive filter 11, a laser data processing apparatus 12, and an initialization processor 13, where the inductive filter 11 is configured to acquire laser data of the robot by induction and acquire particle filter processing information; in one embodiment, when the robot is at an unknown position in an unknown environment, for example, the robot can estimate, for example, a real-time pose of the robot while moving based on a sensor (e.g., (encoder, IMU, laser, camera, etc.) carried by the robot through, for example, slam technology, and then continuously expand and update a map to gradually construct a complete map of the environment; the laser data processing device 12 is configured to process the robot laser data according to the particle filter processing information to obtain particle pose data, and the laser data processing device 12 is connected to the inductive filter device 11, in an embodiment, the initialized particles may be a first generation, and in this embodiment, the particle filter laser slam may initialize, for example, a given number of particles and information for maintenance thereof, and the like, according to the set number of particles in an initialization stage. The key information maintained by each particle may be, for example, the map constructed and its pose in the map. In this embodiment, the particle pose may be initialized to, for example, (0,0,0), etc., and each cell of the map is initialized to, for example, an unknown state. The initially generated particles may be, for example, first generation particles; a primary initialization processor 13 for initializing the particle pose data as a first generation particle, the primary initialization processor 13 being connected to the laser data processing device 12, and in an embodiment the core idea of the particle filter algorithm is to approximate the integration operation by summation using, for example, a weighted sum of a series of random samples to approximate the posterior probability density function. In one embodiment, the slam of particle filtering uses a large number of particles to approximate the posterior distribution of the robot pose, each particle representing, for example, a sample, and each particle carrying, for example, an environment map, and in one embodiment, a removable storage medium readable by, for example, particle data, and in one embodiment, a portable storage medium is installed on a processing component of the robot, such as a microprocessor, and a map building program is stored thereon, and in this embodiment, all or part of the steps for implementing the above method embodiments can be performed by hardware associated with a computer program. The aforementioned computer program may be stored in a computer-readable portable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned map portable storage media may include, but are not limited to: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
Referring to fig. 8, which is a schematic diagram illustrating the connection of specific components of the second-generation resampling processor 2 in fig. 6 in an embodiment, as shown in fig. 8, the second-generation resampling processor 2 includes a first-generation weighting processor 21, a second-generation particle processor 22, a second-generation map processor 23, a second-generation maintenance processor 24, and the first-generation weighting processor 21, which is configured to process weighted particle information to obtain second-generation resampling data; a second-generation particle processor 22, configured to sample the first-generation particles according to the second-generation resampling data to obtain second-generation particles, where the second-generation particle processor 22 is connected to the first-generation weighting processor 21, and in an embodiment, in a resampling stage based on, for example, a laser slam of particle filtering, sample particles are redistributed according to, for example, a weight ratio, and a probability that a particle with a larger weight is sampled is higher; the second-generation map processor 23 is configured to inherit the first-generation sampled particle map according to the second-generation resampling data to obtain a second-generation sampled particle map, and the second-generation map processor 23 is connected to the second-generation particle processor 22, in an embodiment, the duplication of the particle map is reduced under the condition that the number of used particles is not changed, and the management of the optimized particle map is a key technology for improving the mapping efficiency and precision, and in an embodiment, the duplication reduction of the particle map can improve the speed and effect of mapping, for example, the particle filter laser slam; the second generation maintenance processor 24 acquires second generation particle map maintenance information according to the second generation particles so as to maintain the second generation sampled particle map, and the second generation maintenance processor 24 is connected with the second generation particle processor 22 and the second generation map processor 23.
Referring to fig. 9, which is a connection diagram illustrating specific components of the third generation resampling processor 3 in fig. 6 in an embodiment, as shown in fig. 9, the third generation resampling processor 3 includes a second generation weighting processor 31, a third generation particle processor 32, a third generation map processor 33, and a third generation maintenance processor 34, the second generation weighting processor 31 is configured to process weighted particle information to obtain third generation resampling data, and in an embodiment, after a third generation particle is completely generated, a map of a first generation particle is merged onto a map maintained by a second generation particle that inherits the map, and all information maintained by the first generation particle is cleared; a third-generation particle processor 32 for sampling the second-generation particles according to the third-generation resampling data to obtain third-generation particles, the third-generation particle processor 32 is connected with the second-generation weighting processor 31, a third-generation map processor 33 for inheriting the second-generation sampled particle map according to the third-generation resampling data to obtain a third-generation particle map, the third-generation map processor 33 is connected with the third-generation particle processor 32, a third-generation maintenance processor 34 for obtaining third-generation particle map maintenance information according to the third-generation particles to maintain the third-generation particle map, the third-generation maintenance processor 34 is connected with the third-generation map processor 32 and the third-generation particle processor 33, in one embodiment, the second-generation particles are updated to be the first-generation particles, the third-generation particles are updated to be the second-generation particles, in this embodiment, after the third-generation particles are completely generated, the maintained particle set, in an embodiment, the particle set maintained by the present invention includes at most three generations, that is, the particle set includes at most three particles generated in three different sampling stages, so as to reduce the occupation of the memory. The processor may be, for example, a general-purpose processor, and may include, but is not limited to, a Central Processing Unit (CPU), a Network Processor (NP), and the like; the integrated circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, or discrete hardware components.
Referring to fig. 10, which is a schematic diagram showing the connection of specific components of the update processor 4 in fig. 6 in an embodiment, as shown in fig. 10, the map creation processor 4 includes a merging processor 41, a clearing processor 42, a second generation update processor 43, a third generation update processor 44, and the merging processor 41 is used to merge the first generation sampled particle map and the second generation sampled particle map; a purge processor 42 to purge the first generation particles; a second generation update processor 43 for updating the second generation sampled particle map to a first generation sampled particle map, the second generation update processor 43 being connected to the merging processor connection 41; a third generation update processor 44 for updating the third generation particle map to a second generation sampled particle map, the third generation update processor 44 being connected to the second generation update processor 43, and in one embodiment, new particles are generated continuously as the map is continuously updated and expanded, and continuously resampled. The maintained particle set contains three generations at most, thereby reducing the occupation of the memory. In this embodiment, after the resampling of the slam is completed, the newly generated map maintained by the particles is continuously updated along with the continuous input of the laser data. The maps maintained by the rest of the particle sets are not inserted with new laser data and are not updated.
In summary, the present invention aims to provide a method and a system for constructing a robot map, so as to overcome the defects of the prior art, the sampled particles and the information carried by the sampled particles do not need to be completely copied, the sampled particle map only needs to be inherited, and the remaining particles are removed, thereby avoiding adverse effects of the grid map copying process in the prior art, such as large system running time and memory occupation, and the like, and improving the real-time performance of map construction.
Meanwhile, under the condition that the constructed environment map is enlarged, for example, in order to achieve a better mapping effect, more time is occupied for resampling of multi-generation particles, the mapping instantaneity of the robot in a large environment is improved, the limitation of the mapping function by the performance of hardware is reduced, and the mapping effect of the robot is optimized.
In conclusion, the invention provides a robot map construction method and system, which solve the technical problems of low map construction efficiency and poor map construction precision in the prior art, and have high commercial value and high practicability.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (10)
1. A robot map construction method is characterized by comprising the following steps:
initializing particle pose and map data, collecting robot laser data to obtain a first generation of particles, and accordingly obtaining a first generation of sampled particle map;
acquiring weighted particle information in the first-generation particles, sampling the first-generation particles and inheriting a first-generation sampled particle map to obtain a second-generation particle map and a second-generation sampled particle map;
acquiring weighted particle information in the second-generation particles, and accordingly sampling the second-generation particles and inheriting a second-generation sampled particle map to obtain a third-generation particle map and a third-generation particle map;
emptying the first generation particles and updating the second generation particles and the third generation particles so as to obtain mapping data;
judging whether the map building is finished or not;
if yes, generating a grid map according to the mapping data;
if not, circularly executing the steps until the map building is completed.
2. The method for constructing a robot map according to claim 1, wherein the step of acquiring and filtering the laser data by the acquisition device to obtain the first-generation particles, and acquiring the first-generation sampled particle map particle pose data comprises:
the method comprises the steps of obtaining robot laser data through induction and obtaining particle filtering processing information;
processing the robot laser data according to the particle filtering processing information to obtain the particle pose data;
and initializing the particle pose data into the first generation particles.
3. The method for constructing a robot map according to claim 1, wherein the step of obtaining weighted particle information in the first-generation particles, so as to sample the first-generation particles and inherit the first-generation sampled particle map, so as to obtain a second-generation particle and a second-generation sampled particle map, comprises:
processing the weighted particle information to obtain second-generation resampling data;
sampling the first generation particles according to the second generation resampling data pair to obtain the second generation particles;
inheriting the first generation sampled particle map according to the second generation resampled data to obtain a second generation sampled particle map;
and obtaining second-generation particle map maintenance information according to the second-generation particles so as to maintain the second-generation sampled particle map.
4. The method for constructing a robot map according to claim 1, wherein the step of obtaining weighted particle information in the second-generation particles, so as to sample the second-generation particles and inherit the second-generation sampled particle map to obtain third-generation particles and a third-generation particle map comprises:
processing the weighted particle information to obtain third generation resample data;
sampling the second-generation particles according to the third-generation resampling data pair to obtain third-generation particles; inheriting the second generation sampled particle map according to third generation resampling data to obtain a third generation particle map;
and obtaining third-generation particle map maintenance information according to the third-generation particles so as to maintain the third-generation particle map.
5. The method for constructing a robot map according to claim 1, wherein the step of emptying the first generation particles and updating the second generation particles and the third generation particles to obtain mapping data comprises:
merging the first generation sampled particle map and the second generation sampled particle map;
emptying the first generation particles;
updating the second generation sampled particle map into a first generation sampled particle map;
and updating the third-generation particle map into a second-generation sampled particle map.
6. A construction system of a robot map, characterized by comprising:
the laser data processing device is used for initializing the position and the pose of the particles and map data, acquiring the laser data of the robot to obtain first generation particles, and accordingly acquiring a first generation sampled particle map;
the second-generation resampling processor is used for acquiring weighted particle information in the first-generation particles, sampling the first-generation particles and inheriting a first-generation sampled particle map so as to obtain a second-generation particle map and a second-generation sampled particle map;
the third-generation resampling processor is used for acquiring weighted particle information in the second-generation particles, sampling the second-generation particles and inheriting a second-generation sampled particle map so as to obtain a third-generation particle map and a third-generation particle map;
the updating processor is used for emptying the first generation particles and updating the second generation particles and the third generation particles so as to obtain mapping data;
the image building judgment processor is used for judging whether the image building is finished or not;
the grid map processor is used for generating a grid map according to the mapping data when mapping is completed;
and the updating expansion processor is used for circularly executing the steps until the map building is finished when the map building is not finished.
7. The robot map building system according to claim 6, wherein the laser data processing device includes:
the induction filtering device is used for induction obtaining of laser data of the robot and obtaining of particle filtering processing information;
a particle pose device for processing the robot laser data according to the particle filter processing information to obtain the particle pose data;
and the primary generation initial processor is used for initializing and processing the particle pose data into the first generation particles.
8. The robotic map building system of claim 6, wherein the second generation resampling processor comprises:
a second generation weighting processor for processing the weighted particle information to obtain second generation resample data;
a second-generation particle processor to sample the first-generation particles according to the second-generation resample data pairs to obtain the second-generation particles;
a second generation map processor for inheriting the first generation sampled particle map according to the second generation resampled data to obtain a second generation sampled particle map;
and the second generation maintenance processor acquires second generation particle map maintenance information according to the second generation particles so as to maintain the second generation sampled particle map.
9. The robot map building system according to claim 6, wherein the third generation resampling processor comprises:
a second generation weighting processor for processing the weighted particle information to obtain third generation resample data;
a third-generation particle processor for sampling the second-generation particles according to the third-generation resample data pairs to obtain the third-generation particles;
the third-generation map processor is used for inheriting the second-generation sampled particle map according to the third-generation resampling data so as to obtain a third-generation particle map;
and the third-generation maintenance processor acquires third-generation particle map maintenance information according to the third-generation particles so as to maintain the third-generation particle map.
10. The robot map building system according to claim 6, wherein the update processor includes: a merging processor to merge the first generation sampled particle map and the second generation sampled particle map;
a clearing processor for clearing the first generation particles;
a second generation update processor to update the second generation sampled particle map to a first generation sampled particle map;
and the third generation updating processor is used for updating the third generation particle map into a second generation sampled particle map.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911167361.4A CN110909105B (en) | 2019-11-25 | 2019-11-25 | Robot map construction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911167361.4A CN110909105B (en) | 2019-11-25 | 2019-11-25 | Robot map construction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110909105A true CN110909105A (en) | 2020-03-24 |
| CN110909105B CN110909105B (en) | 2022-08-19 |
Family
ID=69819434
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911167361.4A Active CN110909105B (en) | 2019-11-25 | 2019-11-25 | Robot map construction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110909105B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288863A (en) * | 2020-12-24 | 2021-01-29 | 之江实验室 | Map construction method in robot synchronous positioning and composition navigation |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011069023A2 (en) * | 2009-12-02 | 2011-06-09 | Qualcomm Incorporated | Fast subspace projection of descriptor patches for image recognition |
| CN103631264A (en) * | 2013-12-04 | 2014-03-12 | 苏州大学张家港工业技术研究院 | Method and device for simultaneous localization and mapping |
| CN107246873A (en) * | 2017-07-03 | 2017-10-13 | 哈尔滨工程大学 | A kind of method of the mobile robot simultaneous localization and mapping based on improved particle filter |
| US20180075643A1 (en) * | 2015-04-10 | 2018-03-15 | The European Atomic Energy Community (Euratom), Represented By The European Commission | Method and device for real-time mapping and localization |
| CN107831765A (en) * | 2017-10-23 | 2018-03-23 | 广州视源电子科技股份有限公司 | Localization method, device, equipment and storage medium |
| CN107991683A (en) * | 2017-11-08 | 2018-05-04 | 华中科技大学 | A kind of robot autonomous localization method based on laser radar |
| CN109470233A (en) * | 2018-09-13 | 2019-03-15 | 北京米文动力科技有限公司 | A kind of localization method and equipment |
| CN109798896A (en) * | 2019-01-21 | 2019-05-24 | 东南大学 | A kind of positioning of Indoor Robot with build drawing method and device |
-
2019
- 2019-11-25 CN CN201911167361.4A patent/CN110909105B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011069023A2 (en) * | 2009-12-02 | 2011-06-09 | Qualcomm Incorporated | Fast subspace projection of descriptor patches for image recognition |
| CN103631264A (en) * | 2013-12-04 | 2014-03-12 | 苏州大学张家港工业技术研究院 | Method and device for simultaneous localization and mapping |
| US20180075643A1 (en) * | 2015-04-10 | 2018-03-15 | The European Atomic Energy Community (Euratom), Represented By The European Commission | Method and device for real-time mapping and localization |
| CN107246873A (en) * | 2017-07-03 | 2017-10-13 | 哈尔滨工程大学 | A kind of method of the mobile robot simultaneous localization and mapping based on improved particle filter |
| CN107831765A (en) * | 2017-10-23 | 2018-03-23 | 广州视源电子科技股份有限公司 | Localization method, device, equipment and storage medium |
| CN107991683A (en) * | 2017-11-08 | 2018-05-04 | 华中科技大学 | A kind of robot autonomous localization method based on laser radar |
| CN109470233A (en) * | 2018-09-13 | 2019-03-15 | 北京米文动力科技有限公司 | A kind of localization method and equipment |
| CN109798896A (en) * | 2019-01-21 | 2019-05-24 | 东南大学 | A kind of positioning of Indoor Robot with build drawing method and device |
Non-Patent Citations (2)
| Title |
|---|
| 刘红林 等: "基于JSD自适应粒子滤波的移动机器人定位算法", 《安徽工程大学学报》 * |
| 王田橙 等: "基于区域粒子群优化和部分高斯重采样的SLAM方法", 《计算机工程》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288863A (en) * | 2020-12-24 | 2021-01-29 | 之江实验室 | Map construction method in robot synchronous positioning and composition navigation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110909105B (en) | 2022-08-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Czerniawski et al. | 6D DBSCAN-based segmentation of building point clouds for planar object classification | |
| JP6902611B2 (en) | Object detection methods, neural network training methods, equipment and electronics | |
| Salinas et al. | Structure‐aware mesh decimation | |
| US7443393B2 (en) | Method, system, and program product for re-meshing of a three-dimensional input model using progressive implicit approximating levels | |
| WO2020077850A1 (en) | Method and apparatus for dividing and identifying indoor region, and terminal device | |
| CN116404760B (en) | Distributed power grid transient steady-state operation method and device based on digital twin map | |
| CN110909105B (en) | Robot map construction method and system | |
| CN109388843B (en) | Visualization system and method of truss antenna based on VTK (virtual terminal K), and terminal | |
| CN115731355A (en) | SuperPoint-NeRF-based three-dimensional building reconstruction method | |
| CN113537324B (en) | House type space matching method and device based on thin geometric plane spline interpolation | |
| CN115619900B (en) | Point cloud map topological structure extraction method based on distance map and probability road map | |
| CN111130569B (en) | Spatial information data self-adaptive fault-tolerant processing method and system | |
| CN117422841A (en) | Three-dimensional reconstruction method and system based on remote detection data | |
| Xiang et al. | Integrating inverse photogrammetry and a deep learning–based point cloud segmentation approach for automated generation of BIM models | |
| CN115540852A (en) | Electronic grid map construction method and device, electronic equipment and storage medium | |
| De Geyter et al. | Automated training data creation for semantic segmentation of 3D point clouds | |
| Tessema et al. | Extraction of indoorgml model from an occupancy grid map constructed using 2d lidar | |
| Shen et al. | A Novel Self‐Positioning Based on Feature Map Creation and Laser Location Method for RBPF‐SLAM | |
| CN113593025B (en) | Geologic body model updating method, device, equipment and storage medium | |
| Tian et al. | Robot autonomous exploration and map building method | |
| Dellinger et al. | Automated waypoint generation with the growing neural gas algorithm | |
| CN113989440B (en) | Point fast matching method suitable for moment method | |
| Lam et al. | A skeletonization technique based on delaunay triangulation and piecewise bezier interpolation | |
| Peter | Modelling of Indoor Environments Using Lindenmayer Systems | |
| Ling et al. | RANSAC-Based Planar Point Cloud Segmentation Enhanced by Normal Vector and Maximum Principal Curvature Clustering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |