CN116774603A - Multi-AGV cooperative scheduling simulation platform and simulation method - Google Patents
Multi-AGV cooperative scheduling simulation platform and simulation method Download PDFInfo
- Publication number
- CN116774603A CN116774603A CN202310413805.8A CN202310413805A CN116774603A CN 116774603 A CN116774603 A CN 116774603A CN 202310413805 A CN202310413805 A CN 202310413805A CN 116774603 A CN116774603 A CN 116774603A
- Authority
- CN
- China
- Prior art keywords
- agv
- task
- path
- information
- data
- 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
- 238000004088 simulation Methods 0.000 title claims abstract description 84
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000012545 processing Methods 0.000 claims abstract description 50
- 238000004891 communication Methods 0.000 claims abstract description 43
- 230000005540 biological transmission Effects 0.000 claims abstract description 21
- 238000003860 storage Methods 0.000 claims abstract description 19
- 230000006870 function Effects 0.000 claims abstract description 17
- 238000013439 planning Methods 0.000 claims description 30
- 238000013507 mapping Methods 0.000 claims description 15
- 230000001133 acceleration Effects 0.000 claims description 14
- 230000033001 locomotion Effects 0.000 claims description 14
- 238000010586 diagram Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 230000004888 barrier function Effects 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 6
- 238000007499 fusion processing Methods 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 6
- 230000006855 networking Effects 0.000 claims description 6
- 238000013468 resource allocation Methods 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 claims description 5
- 240000007651 Rubus glaucus Species 0.000 claims description 3
- 235000011034 Rubus glaucus Nutrition 0.000 claims description 3
- 235000009122 Rubus idaeus Nutrition 0.000 claims description 3
- 230000009471 action Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 description 9
- 230000003993 interaction Effects 0.000 description 7
- 239000000463 material Substances 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000000903 blocking effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 239000012770 industrial material Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B17/00—Systems involving the use of models or simulators of said systems
- G05B17/02—Systems involving the use of models or simulators of said systems electric
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a multi-AGV cooperative scheduling simulation platform, which comprises a service layer, a communication layer and a physical layer, wherein the service layer is used for receiving a message from a user; the service layer intensively deploys software, deploys programs needing to be calculated on the edge cloud system, and simulates the storage environment; the communication layer provides an information transmission channel, has the functions of network access, data transmission, signal processing and resource scheduling, and supports a plurality of devices to access the Internet through wireless communication; the physical layer implements the emulation function. The invention also discloses a simulation method of the multi-AGV collaborative scheduling simulation platform. The invention has good simulation effect and higher communication reliability and stability.
Description
Technical Field
The invention belongs to the technical field of intelligent control, and particularly relates to a multi-AGV cooperative scheduling simulation platform and a simulation method.
Background
With the continuous development of the manufacturing industry around the world, china starts to enter the era of intelligent manufacturing, the industrial field of China is facing a transformation challenge from the traditional manufacturing industry to the intelligent manufacturing industry, although the industry is one of the major countries in the global manufacturing industry field, a certain gap exists in the high-end manufacturing field compared with other manufacturing industry strong countries, in the current manufacturing industry field, the transportation and distribution of materials occupy more than 95% of time, and the processing and assembly operations occupy less than 5% of time, so that the improvement of the efficiency of material transportation and distribution becomes a big research hot spot in the intelligent manufacturing field.
The automatic navigation trolley (Automated Guided Vehicle, AGV) has become one of key equipment for carrying materials in many factories, and the equipment which is flexible to operate, has strong transportation capability and can be automatically guided can automatically complete logistics tasks under the management of automatic software or control programs, integrates various technologies including computers, embedded sensors and the like, and becomes an important component in the field of logistics systems. AGVs are widely used to replace manual labor for industrial material handling, so that labor cost and corresponding enterprise management cost are greatly reduced, and high competitiveness and high income are brought to enterprises. However, in the use process of the AGV, planning a collision-free path for the AGV is one of the most important and challenging tasks in the multi-AGV system, and when the collision-free path is planned for the AGV, the change of the position of the AGV and the real-time change of the warehouse environment cause a static path planning to have a great limitation, however, the real-time path planning needs to know the warehouse environment information and the position information of the AGV, and meanwhile, a reliable and high-speed communication system is needed for transmitting the information data.
When the AGVs jointly execute the cooperative transport tasks, synchronization or coordination among the AGVs needs to be carried out through information communication, such as coordinate determination, path planning, deadlock prevention, collision avoidance and the like. In order to maintain coordination and cooperation among multiple AGVs, the environment of executing tasks by the multiple AGVs is more comprehensively known, and given tasks are better completed through interaction among the AGVs, the interaction modes among the AGVs can be divided into the following three main types: (1) Through communication interaction, namely by means of the existing network technology, each AGV can transmit information to interact, the interaction mode is similar to computer network communication, however, the communication mode of the multi-AGV system is greatly different from the computer network communication, and when the number of AGVs in the system is multiplied or the topology of the communication network is changed frequently, the operation efficiency of the multi-AGV system is obviously reduced due to the burden of the communication network; (2) Through sensing interaction, the AGVs can interact through various sensors rather than through external communication, and the interaction mode requires the AGVs to distinguish other AGVs in the group or objects in the environment; (3) Through environmental interactions, each AGV in the communication system is the simplest but most limited way to have no direct connection between AGVs and still be able to sense the location of nearby team members through the environment. In addition, most of the current multi-AGV collaborative scheduling simulation platforms are ideal environment tests for researching scheduling algorithms, and the main research is the difference of algorithm performances. Little consideration is given to the motion model of the AGV and the vehicle motion. The deployment of the simulation platform in the field scenario requires specific transformations and even changes to the algorithms studied.
The above communication methods related to multi-AGV cooperative scheduling have certain problems, and technical researches related to multi-AGV cooperative scheduling simulation have certain limitations.
Disclosure of Invention
The invention aims to provide a multi-AGV collaborative scheduling simulation platform with good simulation effect and improved communication reliability and stability.
The second purpose of the invention is to provide a simulation method based on the multi-AGV collaborative scheduling simulation platform.
The multi-AGV collaborative scheduling simulation platform provided by the invention comprises a service layer, a communication layer and a physical layer; the service layer is used for deploying software in a centralized way, deploying programs to be calculated on the edge cloud system, and simulating the storage environment; the communication layer is used for providing an information transmission channel, has the functions of network access, data transmission, signal processing and resource scheduling, and supports a plurality of devices to access the Internet through wireless communication; the physical layer is used for realizing the simulation function.
The service layer comprises an Edge Cloud system or an Edge Cloud server, and the Edge Cloud realizes data transmission and processing by deploying the server at a network Edge within a set range of a data source to be processed;
the communication layer supports a plurality of devices to access the Internet through wireless communication by network access, data transmission, signal processing and resource scheduling functions provided by the experimental base station; the experimental base station adopts SDR technology, and uses software to realize the function of the radio station; constructing an experimental base station through computing equipment and a radio frequency board card; the experimental base station supports a non-independent networking NSA mode and an independent networking SA mode, and supports centralized deployment and distributed deployment; the experimental base station realizes data communication between the service layer and the physical layer by accessing the service layer and the physical layer;
the physical layer is used as a controlled main body and comprises a plurality of AGVs with differential driving structures, a laser radar is arranged for collecting surrounding environment information of the AGVs, an inertial sensor is used for collecting speed and acceleration information of the AGVs, raspberry Pi is arranged as a communication tool for sending and receiving AGV data, and the Raspberry Pi transmits control information to a main controller board opencr through wires to control wheel movements of the AGVs; when no real object exists, a Gazebo system is used for simulating the AGV to execute a carrying task process, a model is built according to a storage environment and an AGV model, the model is imported into the Gazebo system, and the system truly restores the collision action of the object through a physical engine; the Gazebo system realizes SLAM mapping and measuring acceleration and angular velocity of the AGV by adding a laser radar and an inertial measurement unit sensor;
the service layer simulates the operation flow of the warehouse environment, defines the software deployed in the service layer as a central dispatching system, and the central dispatching system comprises a simulation platform, a dispatching system and an execution system;
the simulation platform is used for receiving point cloud data and odometer data scanned by the laser radar of the AGV, determining surrounding environment information according to the point cloud data, determining the self position and the self orientation of the AGV when the AGV is relative to the initial position according to the odometer data, carrying out fusion processing on information acquired by the AGV in a topological structure in the simulation platform, and determining overall environment information; the scheduling system performs task allocation and path planning according to the task instruction, the task allocation follows a nearby principle, the path planning obtains k shortest paths from the AGV position to the task point, and a path which cannot be deadlocked or collided in the current environment state is selected; the execution system is used for converting the path planned by the scheduling system and converting the path into a speed instruction and an angular speed instruction according to the motion rule of the AVG;
when the path planning acquires k shortest paths from the AGV position to the task point, the existing moving distance and the future minimum moving distance of the path are considered, and the cost function is described by adopting the following formula:
f(x)＝g(x)+h(x)
wherein g (x) is the moving distance from the starting point of the AGV to the set point position, and h (x) is the future moving distance from the set point position to the target point position; selecting the minimum moving distance from the starting point to the end point through traversal, and acquiring a shortest path; when searching the next node by traversing at each node, considering the future moving distance, and selecting the node with the minimum f (x) when selecting the next node;
when the path planning is used for selecting a path which cannot be subjected to deadlock and collision in the current environment state, taking an aisle through which the AGV can pass as a reusable resource, and judging whether aisle resources occupied by the planned path collide to give the path, whether collision and deadlock occur or not; if each AGV can finish the access to the task end point by using the unoccupied aisle resources and the occupied aisle resources, the path planned to the AGV is considered to be safe; through using a data structure and checking, initializing the nodes which can be reached by the AGVs in the vertical library information, wherein the edge nodes occupied by the AGVs are not counted into the checking set, naming all the nodes which can be driven or stopped by the AGVs on the map, and pointing the father node to the father node; combining two nodes in the vertical library, which can allow the AGVs to pass through the nodes of the aisle, so that the father node of one node points to the father node of the other node, and the same nodes of the father nodes are communicated; judging whether the state of the AGV is safe or not by judging whether the father node of the task starting point of the AGV is consistent with the father node reaching the end point; if the father node of the task starting point of the AGV is consistent with the father node reaching the end point, the state of the AGV is safe, and all the AGVs can finish the task within a set time without deadlock or collision; whether the state of the AGV is safe or not is defined as that the AGV runs according to a planned path and is collided or deadlocked, if not, the state is unsafe, otherwise, the state is safe.
The invention also provides a simulation method of the multi-AGV cooperative scheduling simulation platform, which comprises the following steps:
s1, generating a task instruction by adopting the multi-AGV cooperative scheduling simulation platform, and acquiring transmission data of the AGVs;
s2, acquiring virtual maps through SLAM algorithm by adopting the AGV transmission data acquired in the step S1;
s3, the scheduling system acquires map information and AGV information provided by the simulation platform, and performs task allocation processing on the idle AGVs according to the acquired information;
s4, the scheduling system performs path planning processing on the AGVs subjected to task allocation processing in the step S3, and the shortest path from the current position to the task point of the AGVs is obtained;
s5, carrying out deadlock avoidance algorithm test processing on the shortest path obtained in the step S4, determining a collision-free task path, and sending the collision-free task path to an execution platform through a scheduling system;
s6, the execution platform performs conversion processing on the collision-free task path sent in the step S5, and sends a converted instruction to the AGV;
s7, the AGV completes the target task according to the instruction sent by the step S6;
s8, when the AGV returns to the point cloud data in the step S7 and the task path with the obstacle appears, repeating the steps S3-S7, and carrying out path planning processing again; and the target task is completed until the AGV safely executes.
The step S1 of generating a task instruction by adopting the multi-AGV cooperative scheduling simulation platform and acquiring the sending data of the AGV specifically comprises the following steps:
the central scheduling system in the multi-AGV collaborative scheduling simulation platform models according to a storage environment, abstracts the storage environment, considers the topological relation of elements, defines the position where the AGV needs to arrive as a node, abstracts the whole map into a plurality of nodes, connects two nodes through edges and indicates that a passable channel exists, and sets weights on the edges to indicate the distance between the two nodes; the whole map is obtained by abstracting the storage environment;
the multi-AGV collaborative scheduling simulation platform randomly picks the preset executable task point positions by reading the preset storage environment topology structure diagram information or acquires the task point positions by reading the file storing the task information so as to generate a task instruction;
the simulation platform receives point cloud data and odometer data sent by the AGV, the point cloud data comprise information of surrounding maps acquired by the AGV through the laser radar and orientation information of the AGV, the odometer data comprise position and speed information estimated by the AGV in a space, and errors of the AGV in a simulation environment are reduced by combining the point cloud data and the odometer data.
The step S2 of acquiring the virtual map by using the AGV transmitting data acquired in the step S1 through the SLAM algorithm specifically comprises the following steps:
(1) Performing independent mapping treatment on all AGVs; acquiring a map of the surrounding environment of each AGV, and determining the position and the posture of the independent AGV;
(2) Carrying out data fusion processing on the point cloud data, and carrying out joint mapping to obtain information of the whole environment; comparing the overall environment map with the single AGV sub map obtained in the step (1) to obtain the position information and the movement direction of the single AGV in the overall environment;
(3) Comparing the overall environment map obtained in the step (2) with a preset topological structure diagram, and confirming the position of the AGV in the topological structure diagram so as to facilitate subsequent task planning processing;
the mapping process described in the step (1) specifically includes: carrying out SLAM mapping positioning processing by using a Cartographer algorithm, wherein the Cartographer algorithm reads point cloud data collected by an AGV through a laser radar, odometer data of the AGV, speed and acceleration data collected by an inertial sensor, and uses the latest point cloud data to construct a map; when the used point cloud data reaches the set quantity, the sub map stops acquiring information;
setting a probability at each position in the sub map, indicating that there is a barrier when the probability is greater than a set value, indicating that the position is free of the barrier when the probability is less than the set value, and indicating that the position is unknown between the two probabilities; the Cartographer algorithm realizes scene map building by creating sub-maps and combining and splicing the sub-maps;
the accuracy of the sub map in the set time is considered to be reliable, but accumulated errors exist after the set time is exceeded, and the pose of all the sub maps is optimized through loop detection; the loop detection is carried out by matching the newly generated point cloud data with the sub map which is most in line with the point cloud data area, the point cloud data is updated into the sub map information, and the error is reduced;
the position of the AGV is calculated by adopting the odometer data of the AGV and the speed and acceleration data collected by the inertial sensor, and the position error generated by the data is reduced by comparing the calculated position with the position of the AGV in the point cloud data.
The scheduling system in step S3 obtains map information and AGV information provided by the simulation platform, and performs task allocation processing for an idle AGV according to the obtained information, specifically including:
the scheduling system acquires map information and AGV information provided by the simulation platform, and defines the position of the AGV in the map as x AGV 、y AGV The position information of the task point is x target 、y target The Euclidean distance between each AGV and all task points is described by adopting the following formula:
and performing task allocation on the idle AGVs by selecting the shortest Euclidean distance.
The scheduling system in step S4 performs path planning processing for the AGV after performing task allocation processing in step S3, and obtains the shortest path from the current position to the task point of the AGV, which specifically includes:
and (3) the scheduling system performs path planning on the position and the target position of the AGV subjected to task allocation processing in the step (S3), and acquires the shortest paths with the set number by searching the shortest feasible paths of the target position and the AGV starting position in the topological map.
And (5) performing deadlock avoidance algorithm test processing on the shortest path acquired in the step (4), determining a collision-free task path, and sending the collision-free task path to an execution platform through a scheduling system, wherein the method specifically comprises the following steps:
carrying out a deadlock avoidance test, and judging whether the AGV after resource allocation can cause the system to enter a deadlock state through simulation before resource allocation; if developing allocated resources can result in deadlock, then deallocating resources;
k shortest paths obtained in the step S4 are reserved, whether the system is deadlocked or not is judged by traversing the k shortest paths of all AGVs, and paths which are not deadlocked or collided are selected; if deadlock or collision exists, traversing a target AGV with the least increase of the next path length among a plurality of AGVs with path collision, modifying the path of the target AGV into the path with the least increase of the selected path length, and judging whether the system is deadlock or not by using an AGV Banker' algorithm to determine the shortest collision-free task path.
The executing platform in step S6 performs conversion processing on the collision-free task path sent in step S5, and sends a converted instruction to the AGV, which specifically includes:
the differential driving robot Differential Drive robot is selected as an AGV used in a simulation process and practical application, and consists of two driving wheels and an additional supporting wheel, wherein the midpoint of a wheel connecting line is set as the position p of the AGV, the point of contact between a non-driving wheel and a floor is set as q, and the relationship between the position and the direction of the AGV and the angular speed of the driving wheel is described by adopting the following formula:
wherein x is the vehicleThe abscissa in the coordinate system, y is the ordinate of the vehicle in the coordinate system, θ is the direction of the vehicle, r is the radius of the driving wheel, 2b is the distance between the two wheels, w L For the angular velocity of the left wheel of the AGV, w R The angular speed of the right wheel of the AGV;
the relationship between the speed, angular velocity, and angular velocity of the drive wheel of the AGV is expressed using the following formula:
where v is the speed of the AGV,is the angular velocity of the AGV;
considering that the AGV only has two choices of direct travel and steering, the motion parameters of the driving wheels of the AGV are obtained in reverse by setting the speed of the AGV when the AGV moves, and the following formula is adopted for description:
through the processing, the task path is converted into basic driving wheel driving information, the execution system sends the driving information to the AGV through the communication module, and the AGV does not need to perform path conversion through the processor and directly operates.
According to the multi-AGV collaborative scheduling simulation platform and the simulation method, the AGVs and the algorithm are decoupled, so that the requirements of the AGVs on hardware can be reduced, the multi-AGV collaborative scheduling simulation platform can be used only by being provided with a driving device, a communication device and a sensor, and the hardware cost is reduced; when the algorithm is deployed, all AGVs are not deployed, and only the edge cloud system is required to be deployed, so that the maintenance and the update of software are convenient; in addition, the invention sends the sensor information of the AGV to the service layer through 5G communication, and the real-time control of the AGV can be realized by relying on the characteristic of low delay of the 5G communication, and the efficiency is effectively improved through reclassification in the face of emergency situations such as goods blocking in a channel, other AGV faults and the like; therefore, the invention has good simulation effect and higher communication reliability and stability.
Drawings
FIG. 1 is a schematic diagram of a simulation platform module of the method of the present invention.
FIG. 2 is a schematic diagram of the structure of a central dispatching system in the method of the present invention.
FIG. 3 is a schematic flow chart of the method of the present invention.
Detailed Description
As shown in FIG. 1, a schematic diagram of a simulation platform module of the method of the invention is shown, and the multi-AGV collaborative scheduling simulation platform provided by the invention comprises a service layer, a communication layer and a physical layer; the service layer is used for deploying software in a centralized way, deploying programs to be calculated on the edge cloud system, and simulating the storage environment; the communication layer is used for providing an information transmission channel, has the functions of network access, data transmission, signal processing and resource scheduling, and supports a plurality of devices to access the Internet through wireless communication; the physical layer is used for realizing the simulation function;
the service layer comprises an Edge Cloud system or an Edge Cloud server, and the Edge Cloud realizes data transmission and processing by deploying the server at a network Edge within a set range of a data source to be processed; the server used in the invention is a server deployed in a laboratory, is provided with a processor of i9-13900K and an Nvidia RTX4090 display card, is provided with a memory of 64G, is provided with a 3T disk capacity Ubuntu system, and is deployed near a 5G experimental base station in the laboratory;
the communication layer supports a plurality of devices to access the Internet through wireless communication by network access, data transmission, signal processing and resource scheduling functions provided by the experimental base station; the experimental base station adopts SDR technology, and uses software to realize the function of the radio station; constructing an experimental base station through computing equipment and a radio frequency board card; the experimental base station supports a non-independent networking NSA mode and an independent networking SA mode, and supports centralized deployment and distributed deployment; the experimental base station realizes data communication between the service layer and the physical layer by accessing the service layer and the physical layer;
the physical layer is used as a controlled main body and comprises a plurality of AGVs with differential driving structures, a laser radar is arranged for collecting surrounding environment information of the AGVs, an inertial sensor is used for collecting speed and acceleration information of the AGVs, raspberry Pi is arranged as a communication tool for sending and receiving AGV data, and the Raspberry Pi transmits control information to a main controller board opencr through wires to control wheel movements of the AGVs; when no real object exists, a Gazebo system is used for simulating and simulating the AGV to execute a carrying task process, a model is built according to a storage environment and an AGV model, the model is imported into the Gazebo system, and the system truly restores the collision action of the object through a physical engine; the Gazebo system realizes SLAM mapping and measuring acceleration and angular velocity of the AGV by adding a laser radar and an inertial measurement unit sensor; the AGV with the differential driving structure is provided with a laser radar RPLIDAR-A2, and 360-degree omnibearing laser scanning with the scanning frequency of 10Hz is carried out on an area with the radius of 16m, so that the surrounding environment information of the AGV is acquired; the inertial sensor MPU9250 is used for collecting speed and acceleration information of the AGV, the MPU9250 consists of a 3-axis gyroscope, a 3-axis accelerometer and a 3-axis magnetometer, and the angular rate measurement range of the gyroscope is +/-2000 degrees/s; the measurement range of the accelerometer is +/-16 g (g is gravity acceleration); the magnetometer adopts a high-flexibility Hall type sensor to collect data, and the magnetic induction intensity measuring range is +/-4800 mu T, so that the magnetometer can be used for auxiliary measurement of yaw angle;
the service layer simulates the operation flow of the warehouse environment, defines the software deployed in the service layer as a central dispatching system, and the central dispatching system comprises a simulation platform, a dispatching system and an execution system;
the simulation platform is used for receiving point cloud data and odometer data scanned by the laser radar of the AGV, determining surrounding environment information according to the point cloud data, determining the self position and the self orientation of the AGV when the AGV is relative to the initial position according to the odometer data, carrying out fusion processing on information acquired by the AGV in a topological structure in the simulation platform, and determining overall environment information; the scheduling system performs task allocation and path planning according to the task instruction, the task allocation follows a nearby principle, the path planning obtains k shortest paths from the AGV position to the task point, and a path which cannot be deadlocked or collided in the current environment state is selected; the execution system is used for converting the path planned by the scheduling system and converting the path into a speed instruction and an angular speed instruction according to the motion rule of the AVG;
when the path planning acquires k shortest paths from the AGV position to the task point, the existing moving distance and the future minimum moving distance of the path are considered, and the cost function is described by adopting the following formula:
f(x)＝g(x)+h(x)
wherein g (x) is the moving distance from the starting point of the AGV to the set point, and h (x) is the future moving distance from the set point to the target point; selecting the minimum moving distance from the starting point to the end point through traversal, and acquiring a shortest path; when searching the next node by traversing at each node, considering the future moving distance, and selecting the node with the minimum f (x) when selecting the next node;
when the path planning is used for selecting a path which cannot be subjected to deadlock and collision in the current environment state, taking an aisle through which the AGV can pass as a reusable resource, and judging whether aisle resources occupied by the planned path collide to give the path, whether collision and deadlock occur or not; if each AGV can finish the access to the task end point by using the unoccupied aisle resources and the occupied aisle resources, the path planned to the AGV is considered to be safe; through using a data structure and checking, initializing the nodes which can be reached by the AGVs in the vertical library information, wherein the edge nodes occupied by the AGVs are not counted in the checking set, naming all the nodes which can be driven or stopped (i.e. reached) by the AGVs on the map, and pointing the father node to the father node; combining two nodes in the vertical library, which can allow the AGVs to pass through the nodes of the aisle, so that the father node of one node points to the father node of the other node, and the same nodes of the father nodes are communicated; judging whether the state of the AGV is safe or not by judging whether the father node of the task starting point of the AGV is consistent with the father node reaching the end point; if the father node of the task starting point of the AGV is consistent with the father node reaching the end point, the state of the AGV is safe, and all the AGVs can finish the task within a set time without deadlock or collision; whether the state of the AGV is safe or not, specifically, if the AGV runs according to a planned path, collision or deadlock occurs, the state is unsafe, otherwise, the state is safe;
the invention also provides a simulation method comprising the multi-AGV cooperative scheduling simulation platform, which comprises the following steps:
s1, generating a task instruction by adopting a multi-AGV cooperative scheduling simulation platform, and acquiring transmission data of the AGVs; the method specifically comprises the following steps:
the central scheduling system in the multi-AGV collaborative scheduling simulation platform models according to a storage environment, abstracts the storage environment, considers the topological relation of elements, defines the arrived position required by the AGV as nodes, abstracts the whole map as a set number of nodes, connects two nodes through edges and indicates that a passable channel exists, and sets weights on the edges to indicate the distance between the two nodes; the whole map is obtained by abstracting the storage environment;
the multi-AGV collaborative scheduling simulation platform randomly picks the preset executable task point positions by reading the preset storage environment topology structure diagram information or acquires the task point positions by reading the file storing the task information so as to generate a task instruction;
the simulation platform receives point cloud data and odometer data sent by the AGV, wherein the point cloud data comprise information of a surrounding map acquired by the AGV through a laser radar and orientation information of the AGV, the odometer data comprise position and speed information estimated by the AGV in a space, and the error of the AGV in a simulation environment is reduced by combining the point cloud data and the odometer data;
s2, acquiring virtual maps through SLAM algorithm by adopting the AGV transmission data acquired in the step S1; the method specifically comprises the following steps:
(1) Performing independent mapping treatment on all AGVs; acquiring a map of the surrounding environment of each AGV, and determining the position and the posture of the independent AGV;
(2) Carrying out data fusion processing on the point cloud data, and carrying out joint mapping to obtain information of the whole environment; comparing the overall environment map with the single AGV sub map obtained in the step (1) to obtain the position information and the movement direction of the single AGV in the overall environment;
(3) Comparing the overall environment map obtained in the step (2) with a preset topological structure diagram, and confirming the position of the AGV in the topological structure diagram so as to facilitate subsequent task planning processing;
the mapping process described in the step (1) specifically includes: carrying out SLAM mapping positioning processing by using a Cartographer algorithm, wherein the Cartographer algorithm reads point cloud data collected by an AGV through a laser radar, odometer data of the AGV, speed and acceleration data collected by an inertial sensor, and uses the latest point cloud data to construct a map; when the used point cloud data reaches the set quantity, the sub map stops acquiring information;
setting a probability at each position in the sub map, indicating that there is a barrier when the probability is greater than a set value, indicating that the position is free of the barrier when the probability is less than the set value, and indicating that the position is unknown between the two probabilities; the Cartographer algorithm realizes scene map building by creating a large number of sub-maps and combining and splicing the sub-maps;
the accuracy of the sub map in the set time is considered to be reliable, but accumulated errors exist after the set time is exceeded, and the pose of all the sub maps is optimized through loop detection; the loop detection is carried out by matching the newly generated point cloud data with the sub map which is most in line with the point cloud data area, the point cloud data is updated into the sub map information, and the error is reduced;
calculating the position of the AGV by adopting odometer data of the AGV and speed and acceleration data collected by an inertial sensor, and reducing position errors generated by the data by comparing the calculated position with the position of the AGV in the point cloud data;
s3, the scheduling system acquires map information and AGV information provided by the simulation platform, and performs task allocation processing on the idle AGVs according to the acquired information; the method specifically comprises the following steps:
the scheduling system acquires map information and AGV information provided by the simulation platform, and defines the position of the AGV in the map as x AGV 、y AGV Position information of task pointIs x target 、y target The Euclidean distance between each AGV and all task points is described by adopting the following formula:
task allocation is carried out on the idle AGVs by selecting the shortest Euclidean distance;
s4, the scheduling system performs path planning processing on the AGVs subjected to task allocation processing in the step S3, and the shortest path from the current position to the task point of the AGVs is obtained; the method specifically comprises the following steps:
the scheduling system performs path planning on the position and the target position of the AGV after task allocation processing in the step S3, and acquires the shortest paths with set quantity by searching the shortest feasible paths of the target position and the AGV initial position in the topological map;
s5, carrying out deadlock avoidance algorithm test processing on the shortest path obtained in the step S4, determining a collision-free task path, and sending the collision-free task path to an execution platform through a scheduling system; the method specifically comprises the following steps:
carrying out a deadlock avoidance test, and judging whether the AGV after resource allocation can cause the system to enter a deadlock state through simulation before resource allocation; if developing allocated resources can result in deadlock, then deallocating resources;
k shortest paths obtained in the step S4 are reserved, whether the system is deadlocked or not is judged by traversing the k shortest paths of all AGVs, and paths which are not deadlocked or collided are selected; if deadlock or collision exists, traversing a target AGV with the least increase of the next path length among a plurality of AGVs with path collision, modifying the path of the target AGV into the path with the least increase of the selected path length, judging whether the system is deadlock or not by using an AGV Banker' algorithm, and determining the shortest collision-free task path;
s6, the execution platform performs conversion processing on the collision-free task path sent in the step S5, and sends a converted instruction to the AGV; the method specifically comprises the following steps:
the differential driving robot Differential Drive robot is selected as an AGV used in a simulation process and practical application, and consists of two driving wheels and an additional supporting wheel, wherein the midpoint of a wheel connecting line is set as the position p of the AGV, the point of contact between a non-driving wheel and a floor is set as q, and the relationship between the position and the direction of the AGV and the angular speed of the driving wheel is described by adopting the following formula:
wherein x is the abscissa of the vehicle in the coordinate system, y is the ordinate of the vehicle in the coordinate system, θ is the direction of the vehicle, r is the radius of the driving wheel, 2b is the distance between the two wheels, w L For the angular velocity of the left wheel of the AGV, w R The angular speed of the right wheel of the AGV;
the relationship between the speed, angular velocity, and angular velocity of the drive wheel of the AGV is expressed using the following formula:
where v is the speed of the AGV,is the angular velocity of the AGV;
considering that the AGV only has two choices of direct travel and steering, the motion parameters of the driving wheels of the AGV are obtained in reverse by setting the speed of the AGV when the AGV moves, and the following formula is adopted for description:
through the processing, the task path is converted into basic driving wheel driving information, the driving information is sent to the AGV by the execution system, and the AGV can directly run without path conversion by the processor;
s7, the AGV completes the target task according to the instruction sent by the step S6;
s8, when the AGV returns to the point cloud data in the step S7 and the task path with the obstacle appears, repeating the steps S3-S7, and carrying out path planning processing again; and the target task is completed until the AGV safely executes.
Claims (10)
1. A multi-AGV collaborative scheduling simulation platform comprises a service layer, a communication layer and a physical layer; the service layer is used for deploying software in a centralized way, deploying programs to be calculated on the edge cloud system, and simulating the storage environment; the communication layer is used for providing an information transmission channel, has the functions of network access, data transmission, signal processing and resource scheduling, and supports a plurality of devices to access the Internet through wireless communication; the physical layer is used for realizing the simulation function.
2. The multi-AGV collaborative scheduling simulation platform according to claim 1, wherein the service layer comprises an Edge Cloud system or an Edge Cloud server, and the Edge Cloud realizes data transmission and processing by deploying the server at a network Edge within a set range of a data source to be processed;
the communication layer supports a plurality of devices to access the Internet through wireless communication by network access, data transmission, signal processing and resource scheduling functions provided by the experimental base station; the experimental base station adopts SDR technology, and uses software to realize the function of the radio station; constructing an experimental base station through computing equipment and a radio frequency board card; the experimental base station supports a non-independent networking NSA mode and an independent networking SA mode, and supports centralized deployment and distributed deployment; the experimental base station realizes data communication between the service layer and the physical layer by accessing the service layer and the physical layer;
the physical layer is used as a controlled main body and comprises a plurality of AGVs with differential driving structures, a laser radar is arranged for collecting surrounding environment information of the AGVs, an inertial sensor is used for collecting speed and acceleration information of the AGVs, raspberry Pi is arranged as a communication tool for sending and receiving AGV data, and the Raspberry Pi transmits control information to a main controller board opencr through wires to control wheel movements of the AGVs; when no real object exists, a Gazebo system is used for simulating the AGV to execute a carrying task process, a model is built according to a storage environment and an AGV model, the model is imported into the Gazebo system, and the system truly restores the collision action of the object through a physical engine; the Gazebo system realizes SLAM mapping and measuring acceleration and angular velocity of the AGV by adding a laser radar and an inertial measurement unit sensor.
3. The multi-AGV collaborative scheduling simulation platform according to claim 2, wherein the service layer simulates a job flow of a warehouse environment, defines software deployed in the service layer as a central scheduling system, and the central scheduling system comprises a simulation platform, a scheduling system and an execution system;
the simulation platform is used for receiving point cloud data and odometer data scanned by the laser radar of the AGV, determining surrounding environment information according to the point cloud data, determining the self position and the self orientation of the AGV when the AGV is relative to the initial position according to the odometer data, carrying out fusion processing on information acquired by the AGV in a topological structure in the simulation platform, and determining overall environment information; the scheduling system performs task allocation and path planning according to the task instruction, the task allocation follows a nearby principle, the path planning obtains k shortest paths from the AGV position to the task point, and a path which cannot be deadlocked or collided in the current environment state is selected; the execution system is used for converting the path planned by the scheduling system and converting the path into a speed instruction and an angular speed instruction according to the motion rule of the AVG;
when the path planning acquires k shortest paths from the AGV position to the task point, the existing moving distance and the future minimum moving distance of the path are considered, and the cost function is described by adopting the following formula:
f(x)＝g(x)+h(x)
wherein g (x) is the moving distance from the starting point of the AGV to the set point position, and h (x) is the future moving distance from the set point position to the target point position; selecting the minimum moving distance from the starting point to the end point through traversal, and acquiring a shortest path; when searching the next node by traversing at each node, considering the future moving distance, and selecting the node with the minimum f (x) when selecting the next node;
when the path planning is used for selecting a path which cannot be subjected to deadlock and collision in the current environment state, taking an aisle through which the AGV can pass as a reusable resource, and judging whether aisle resources occupied by the planned path collide to give the path, whether collision and deadlock occur or not; if each AGV can finish the access to the task end point by using the unoccupied aisle resources and the occupied aisle resources, the path planned to the AGV is considered to be safe; through using a data structure and checking, initializing the nodes which can be reached by the AGVs in the vertical library information, wherein the edge nodes occupied by the AGVs are not counted into the checking set, naming all the nodes which can be driven or stopped by the AGVs on the map, and pointing the father node to the father node; combining two nodes in the vertical library, which can allow the AGVs to pass through the nodes of the aisle, so that the father node of one node points to the father node of the other node, and the same nodes of the father nodes are communicated; judging whether the state of the AGV is safe or not by judging whether the father node of the task starting point of the AGV is consistent with the father node reaching the end point; if the father node of the task starting point of the AGV is consistent with the father node reaching the end point, the state of the AGV is safe, and all the AGVs can finish the task within a set time without deadlock or collision; whether the state of the AGV is safe or not is defined as that the AGV runs according to a planned path and is collided or deadlocked, if not, the state is unsafe, otherwise, the state is safe.
4. A simulation method of the multi-AGV co-scheduling simulation platform according to any one of claims 1 to 3, comprising the steps of:
s1, generating a task instruction by adopting the multi-AGV cooperative scheduling simulation platform, and acquiring transmission data of the AGVs;
s2, acquiring virtual maps through SLAM algorithm by adopting the AGV transmission data acquired in the step S1;
s3, the scheduling system acquires map information and AGV information provided by the simulation platform, and performs task allocation processing on the idle AGVs according to the acquired information;
s4, the scheduling system performs path planning processing on the AGVs subjected to task allocation processing in the step S3, and the shortest path from the current position to the task point of the AGVs is obtained;
s5, carrying out deadlock avoidance algorithm test processing on the shortest path obtained in the step S4, determining a collision-free task path, and sending the collision-free task path to an execution platform through a scheduling system;
s6, the execution platform performs conversion processing on the collision-free task path sent in the step S5, and sends a converted instruction to the AGV;
s7, the AGV completes the target task according to the instruction sent by the step S6;
s8, when the AGV returns to the point cloud data in the step S7 and the task path with the obstacle appears, repeating the steps S3-S7, and carrying out path planning processing again; and the target task is completed until the AGV safely executes.
5. The simulation method of claim 4, wherein the step S1 of generating a task instruction by using the multi-AGV co-scheduling simulation platform and acquiring the transmission data of the AGV specifically includes:
the central scheduling system in the multi-AGV collaborative scheduling simulation platform models according to a storage environment, abstracts the storage environment, considers the topological relation of elements, defines the position where the AGV needs to arrive as a node, abstracts the whole map into a plurality of nodes, connects two nodes through edges and indicates that a passable channel exists, and sets weights on the edges to indicate the distance between the two nodes; the whole map is obtained by abstracting the storage environment;
the multi-AGV collaborative scheduling simulation platform randomly picks the preset executable task point positions by reading the preset storage environment topology structure diagram information or acquires the task point positions by reading the file storing the task information so as to generate a task instruction;
the simulation platform receives point cloud data and odometer data sent by the AGV, the point cloud data comprise information of surrounding maps acquired by the AGV through the laser radar and orientation information of the AGV, the odometer data comprise position and speed information estimated by the AGV in a space, and errors of the AGV in a simulation environment are reduced by combining the point cloud data and the odometer data.
6. The simulation method according to claim 5, wherein the step S2 is characterized in that the step S1 is adopted to obtain the data sent by the AGV, and the step S2 is adopted to obtain the virtual map through a SLAM algorithm, and the method specifically comprises the steps of:
(1) Performing independent mapping treatment on all AGVs; acquiring a map of the surrounding environment of each AGV, and determining the position and the posture of the independent AGV;
(2) Carrying out data fusion processing on the point cloud data, and carrying out joint mapping to obtain information of the whole environment; comparing the overall environment map with the single AGV sub map obtained in the step (1) to obtain the position information and the movement direction of the single AGV in the overall environment;
(3) Comparing the overall environment map obtained in the step (2) with a preset topological structure diagram, and confirming the position of the AGV in the topological structure diagram so as to facilitate subsequent task planning processing;
the mapping process described in the step (1) specifically includes: carrying out SLAM mapping positioning processing by using a Cartographer algorithm, wherein the Cartographer algorithm reads point cloud data collected by an AGV through a laser radar, odometer data of the AGV, speed and acceleration data collected by an inertial sensor, and uses the latest point cloud data to construct a map; when the used point cloud data reaches the set quantity, the sub map stops acquiring information;
setting a probability at each position in the sub map, indicating that there is a barrier when the probability is greater than a set value, indicating that the position is free of the barrier when the probability is less than the set value, and indicating that the position is unknown between the two probabilities; the Cartographer algorithm realizes scene map building by creating sub-maps and combining and splicing the sub-maps;
the accuracy of the sub map in the set time is considered to be reliable, but accumulated errors exist after the set time is exceeded, and the pose of all the sub maps is optimized through loop detection; the loop detection is carried out by matching the newly generated point cloud data with the sub map which is most in line with the point cloud data area, the point cloud data is updated into the sub map information, and the error is reduced;
the position of the AGV is calculated by adopting the odometer data of the AGV and the speed and acceleration data collected by the inertial sensor, and the position error generated by the data is reduced by comparing the calculated position with the position of the AGV in the point cloud data.
7. The simulation method according to claim 6, wherein the scheduling system in step S3 obtains map information and AGV information provided by the simulation platform, and performs task allocation processing for an idle AGV according to the obtained information, and specifically includes:
the scheduling system acquires map information and AGV information provided by the simulation platform, and defines the position of the AGV in the map as x AGV 、y AGV The position information of the task point is x target 、y target The Euclidean distance between each AGV and all task points is described by adopting the following formula:
and performing task allocation on the idle AGVs by selecting the shortest Euclidean distance.
8. The simulation method of claim 7, wherein the scheduling system of step S4 performs path planning processing for the AGV after performing the task allocation processing in step S3, and obtains a shortest path from the current position to the task point of the AGV, and specifically includes:
and (3) the scheduling system performs path planning on the position and the target position of the AGV subjected to task allocation processing in the step (S3), and acquires the shortest paths with the set number by searching the shortest feasible paths of the target position and the AGV starting position in the topological map.
9. The simulation method according to claim 8, wherein the shortest path obtained in step S4 in step S5 is subjected to a deadlock avoidance algorithm test process, a collision-free task path is determined, and the determined collision-free task path is sent to an execution platform through a scheduling system, and the simulation method specifically includes:
carrying out a deadlock avoidance test, and judging whether the AGV after resource allocation can cause the system to enter a deadlock state through simulation before resource allocation; if developing allocated resources can result in deadlock, then deallocating resources;
k shortest paths obtained in the step S4 are reserved, whether the system is deadlocked or not is judged by traversing the k shortest paths of all AGVs, and paths which are not deadlocked or collided are selected; if deadlock or collision exists, traversing a target AGV with the least increase of the next path length among a plurality of AGVs with path collision, modifying the path of the target AGV into the path with the least increase of the selected path length, and judging whether the system is deadlock or not by using an AGV Banker' algorithm to determine the shortest collision-free task path.
10. The simulation method of claim 9, wherein the executing platform in step S6 performs conversion processing on the collision-free task path sent in step S5, and sends the converted instruction to the AGV, and specifically includes:
the differential driving robot Differential Drive robot is selected as an AGV used in a simulation process and practical application, and consists of two driving wheels and an additional supporting wheel, wherein the midpoint of a wheel connecting line is set as the position p of the AGV, the point of contact between a non-driving wheel and a floor is set as q, and the relationship between the position and the direction of the AGV and the angular speed of the driving wheel is described by adopting the following formula:
wherein x is the abscissa of the vehicle in the coordinate system, y is the ordinate of the vehicle in the coordinate system, θ is the direction of the vehicle, r is the radius of the driving wheel, 2b is the distance between the two wheels, w L For the angular velocity of the left wheel of the AGV, w R The angular speed of the right wheel of the AGV;
the relationship between the speed, angular velocity, and angular velocity of the drive wheel of the AGV is expressed using the following formula:
where v is the speed of the AGV,is the angular velocity of the AGV;
considering that the AGV only has two choices of direct travel and steering, the motion parameters of the driving wheels of the AGV are obtained in reverse by setting the speed of the AGV when the AGV moves, and the following formula is adopted for description:
through the processing, the task path is converted into basic driving wheel driving information, the execution system sends the driving information to the AGV through the communication module, and the AGV does not need to perform path conversion through the processor and directly operates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310413805.8A CN116774603B (en) | 2023-04-18 | 2023-04-18 | Multi-AGV cooperative scheduling simulation platform and simulation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310413805.8A CN116774603B (en) | 2023-04-18 | 2023-04-18 | Multi-AGV cooperative scheduling simulation platform and simulation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116774603A true CN116774603A (en) | 2023-09-19 |
| CN116774603B CN116774603B (en) | 2024-01-30 |
Family
ID=88007003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310413805.8A Active CN116774603B (en) | 2023-04-18 | 2023-04-18 | Multi-AGV cooperative scheduling simulation platform and simulation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116774603B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB0803718D0 (en) * | 2008-02-28 | 2008-04-09 | Advanced Transp Systems Ltd | Method and system for resolving deadlocks |
| US20080216077A1 (en) * | 2007-03-02 | 2008-09-04 | Applied Materials, Inc. | Software sequencer for integrated substrate processing system |
| CN111885550A (en) * | 2020-06-06 | 2020-11-03 | 浙江中力机械有限公司 | Distributed autonomous mobile robot scheduling system |
| EP3838502A1 (en) * | 2018-08-16 | 2021-06-23 | Ju, Hehua | Method for dynamics modeling of multi-axis robot based on axis invariant |
| WO2021254415A1 (en) * | 2020-06-18 | 2021-12-23 | 北京卫星制造厂有限公司 | Time window-based agv intelligent scheduling method |
| CN114115248A (en) * | 2021-11-12 | 2022-03-01 | 湖南大学 | Multi-robot cooperative transportation method and system in storage environment |
| CN114489062A (en) * | 2022-01-18 | 2022-05-13 | 北京理工大学 | Workshop logistics-oriented multi-automatic guided vehicle distributed dynamic path planning method |
| CN115237135A (en) * | 2022-08-02 | 2022-10-25 | 山东大学深圳研究院 | Mobile robot path planning method and system based on conflict |
| CN115293394A (en) * | 2022-04-22 | 2022-11-04 | 湖南大学 | Multi-robot joint scheduling optimization method based on dynamic space-time diagram |
| CN115840454A (en) * | 2023-02-20 | 2023-03-24 | 江苏集萃清联智控科技有限公司 | Multi-vehicle track collaborative planning method and device for unstructured road conflict area |
-
2023
- 2023-04-18 CN CN202310413805.8A patent/CN116774603B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080216077A1 (en) * | 2007-03-02 | 2008-09-04 | Applied Materials, Inc. | Software sequencer for integrated substrate processing system |
| GB0803718D0 (en) * | 2008-02-28 | 2008-04-09 | Advanced Transp Systems Ltd | Method and system for resolving deadlocks |
| EP3838502A1 (en) * | 2018-08-16 | 2021-06-23 | Ju, Hehua | Method for dynamics modeling of multi-axis robot based on axis invariant |
| CN111885550A (en) * | 2020-06-06 | 2020-11-03 | 浙江中力机械有限公司 | Distributed autonomous mobile robot scheduling system |
| WO2021254415A1 (en) * | 2020-06-18 | 2021-12-23 | 北京卫星制造厂有限公司 | Time window-based agv intelligent scheduling method |
| CN114115248A (en) * | 2021-11-12 | 2022-03-01 | 湖南大学 | Multi-robot cooperative transportation method and system in storage environment |
| CN114489062A (en) * | 2022-01-18 | 2022-05-13 | 北京理工大学 | Workshop logistics-oriented multi-automatic guided vehicle distributed dynamic path planning method |
| CN115293394A (en) * | 2022-04-22 | 2022-11-04 | 湖南大学 | Multi-robot joint scheduling optimization method based on dynamic space-time diagram |
| CN115237135A (en) * | 2022-08-02 | 2022-10-25 | 山东大学深圳研究院 | Mobile robot path planning method and system based on conflict |
| CN115840454A (en) * | 2023-02-20 | 2023-03-24 | 江苏集萃清联智控科技有限公司 | Multi-vehicle track collaborative planning method and device for unstructured road conflict area |
Non-Patent Citations (4)
| Title |
|---|
| DAIYU HE 等: "Deadlock Avoidance in Closed Guide-Path Based MultiAGV Systems", 《IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING》, vol. 20, no. 3, pages 2088 - 2098 * |
| LUKA KALINOVCIC 等: "Modified Banker’s algorithm for scheduling in multi-AGV systems", 《2011 IEEE INTERNATIONAL CONFERENCE ON AUTOMATION SCIENCE AND ENGINEERING》, pages 351 - 356 * |
| 叶强: "多AGV系统避障问题与调度效率优化研究", 《万方数据》, pages 7 - 38 * |
| 黄怀崧: "仓储自动化多AGV控制系统与调度算法的研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, no. 7, pages 140 - 432 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116774603B (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qin et al. | Autonomous exploration and mapping system using heterogeneous UAVs and UGVs in GPS-denied environments | |
| Ergezer et al. | 3D path planning for multiple UAVs for maximum information collection | |
| KR20190119237A (en) | Topological map generation apparatus for traveling robot and method thereof | |
| Vincent et al. | Distributed multirobot exploration, mapping, and task allocation | |
| Chen et al. | H-DrunkWalk: Collaborative and adaptive navigation for heterogeneous MAV swarm | |
| CN114355981B (en) | Method and system for autonomous exploration and mapping of four-rotor unmanned aerial vehicle | |
| Azpúrua et al. | A Survey on the autonomous exploration of confined subterranean spaces: Perspectives from real-word and industrial robotic deployments | |
| Garzón et al. | A multirobot system for distributed area coverage and signal searching in large outdoor scenarios | |
| CN112650244A (en) | Multi-floor automatic mapping method for mobile robot in building based on feature point matching | |
| Dharmasiri et al. | Novel implementation of multiple automated ground vehicles traffic real time control algorithm for warehouse operations: djikstra approach | |
| Mannucci et al. | Autonomous 3D exploration of large areas: A cooperative frontier-based approach | |
| Almadhoun et al. | Coverage path planning for complex structures inspection using unmanned aerial vehicle (UAV) | |
| Choi et al. | Cellular Communication-Based Autonomous UAV Navigation with Obstacle Avoidance for Unknown Indoor Environments. | |
| Lian et al. | Improved coding landmark-based visual sensor position measurement and planning strategy for multiwarehouse automated guided vehicle | |
| Sinnemann et al. | Systematic literature review of applications and usage potentials for the combination of unmanned aerial vehicles and mobile robot manipulators in production systems | |
| CN111830995B (en) | Group intelligent cooperation method and system based on hybrid architecture | |
| Wang et al. | ROS-base multi-sensor fusion for accuracy positioning and SLAM system | |
| Mantha et al. | Robots in indoor and outdoor environments | |
| CN116774603B (en) | Multi-AGV cooperative scheduling simulation platform and simulation method | |
| Mantha et al. | Investigating the fiducial marker network characteristics for autonomous mobile indoor robot navigation using ROS and Gazebo | |
| US20220300002A1 (en) | Methods and systems for path planning in a known environment | |
| Legovich et al. | Integration of modern technologies for solving territory patroling problems with the use of heterogeneous autonomous robotic systems | |
| Lebedeva et al. | Method for distributed mapping of terrain by a heterogeneous group of robots based on google cartographer | |
| Wei et al. | Research on the slam of mobile robot based on particle filter method | |
| Dosoftei et al. | Real-Time Motion Control of an Electric Driven OMR using a ROS to Matlab Bridged Approach |
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 |