CN117193288A - Industrial safety intelligent inspection robot dog based on AI algorithm and inspection method - Google Patents
Industrial safety intelligent inspection robot dog based on AI algorithm and inspection method Download PDFInfo
- Publication number
- CN117193288A CN117193288A CN202310936318.XA CN202310936318A CN117193288A CN 117193288 A CN117193288 A CN 117193288A CN 202310936318 A CN202310936318 A CN 202310936318A CN 117193288 A CN117193288 A CN 117193288A
- Authority
- CN
- China
- Prior art keywords
- recognition
- inspection
- algorithm
- representing
- industrial
- 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.)
- Pending
Links
- 238000007689 inspection Methods 0.000 title claims abstract description 107
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 53
- 241000282472 Canis lupus familiaris Species 0.000 claims description 50
- 230000006870 function Effects 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 13
- 210000000689 upper leg Anatomy 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 11
- 230000004888 barrier function Effects 0.000 claims description 8
- 238000007418 data mining Methods 0.000 claims description 5
- 238000003058 natural language processing Methods 0.000 claims description 5
- 230000001681 protective effect Effects 0.000 claims description 3
- 239000000779 smoke Substances 0.000 claims description 3
- 230000000391 smoking effect Effects 0.000 claims description 3
- 244000309466 calf Species 0.000 claims description 2
- 238000004088 simulation Methods 0.000 description 19
- 238000004458 analytical method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 101150050759 outI gene Proteins 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000002922 simulated annealing Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The application discloses an industrial safety intelligent inspection machine dog and an inspection method based on an AI algorithm. The application not only improves the objectivity of the enterprise safety inspection result, but also improves the execution efficiency of the inspection route.
Description
Technical Field
The application relates to an industrial safety intelligent inspection machine dog and an inspection method based on an AI algorithm, and belongs to the technical field of path planning.
Background
The common inspection mode of industrial enterprises is to observe the running state of equipment by inspection personnel according to daily working experience so as to judge whether the equipment is safe, and the inspection mode completely depends on the working experience of the inspection personnel and has no objectivity. The existing large enterprises also adopt the unmanned inspection mode of the inspection machine dog to replace the manual inspection, but in the actual process, the inspection scene is single in identification due to inaccurate calculation of the inspection route, and the inspection machine dog can not normally finish the inspection route work.
The routing includes global path planning and local path planning. The global path planning mainly considers the rough path between the initial node and the target node, and ignores factors such as path direction, roadblock, speed and the like. The existing modeling method comprises the following steps: free Space (Free Space), visual Graph (Visibility Graph), voronoi diagram, topology, and grid (Cell Decomposition); the global path planning method comprises the following steps: dijkstra algorithm, ant colony algorithm, neural network algorithm, fuzzy control algorithm, etc.
The local path planning is performed by expanding under the condition of incompletely knowing the external condition and combining the real-time environment information and the self condition. In the planning, the external conditions are received and fed back in real time according to the sensors assembled with the user, so that an optimal path is planned. The local planning method comprises the following steps: artificial potential field method, genetic algorithm, etc., simulated annealing algorithm, particle swarm algorithm, etc.
Disclosure of Invention
The application aims to: in order to solve the problems that the routing computation of an intelligent routing inspection machine is inaccurate, the routing inspection intelligent recognition algorithm is single, and the routing inspection operation cannot be completed normally, the application provides an industrial safety intelligent routing inspection machine dog and a routing inspection method based on an AI algorithm.
The technical scheme is as follows: in order to achieve the above purpose, the application adopts the following technical scheme:
an industrial safety intelligent inspection method for inspecting dogs based on an AI algorithm comprises the following steps:
step 1, a patrol path planning kinematic modeling is carried out, a method of combining kinematic forward solution and kinematic inverse solution is adopted to obtain a motion space range of a dog foot end of an industrial patrol machine, so that a patrol path of the industrial patrol machine dog is planned, and places with stairs and slopes are planned into the patrol path.
And 2, modeling a routing inspection path planning environment, namely performing routing inspection path planning environment modeling by adopting a grid method, knowing routing inspection information and barrier information of the industrial routing inspection machine dog through a small grid, and directly corresponding grid information with actual environment information. The motion on the grid defines 8 motion directions, the definition value of each cell is 0 or 1,1 represents the obstacle area, and 0 represents the free area.
Step 3, optimizing routing inspection path planning algorithm
Establishing a path finding heuristic function, and solving by adopting diagonal distanceValue and pair ofAnd (3) performing weighted optimization:
;
wherein,representing the overall priority of node n,representing the cost of the node n from the start point,representing the cost of the node n from the endpoint,is the weighting coefficient of h (n).
Weighting the heuristic functions:
;
in the method, in the process of the application,representing the cost of the node n from the endpoint,representing the optimized weighting coefficients.
Preferably: the kinematic positive solution process in step 1:
write-firstIs the equation expression of (2):
;
rewriting outIs the equation expression of (2):
;
wherein,the coordinates of the joint are represented,representing the length of the lower leg,the rotation amount is represented, the rotation angle of the shank steering engine is represented,the coordinates of the foot end of the robot are represented,indicating the length of the thigh of the person,the rotation amount is represented, and the rotation angle of the thigh steering engine is represented.
Preferably: the kinematic inverse solution process in step 1:
solving the angle of the lower leg:
;
according to the cosine law:
;
;
obtaining the steering angle of the shank steering engine。
Solving the angle of the thigh:
according to the cosine law:
;
;
;
;
;
;
obtain the steering wheel corner of the thigh。
Wherein,represents the projection distance of the connecting line of the steering engine and the foot end on the xOy plane,represents the rotation angle of the shank steering engine,the rotation quantity between the connecting line of the foot end of the steering engine and the lower leg is represented as the intermediate quantity.
Preferably: in step 3The calculation method of (2) is as follows:
;
;
;
;
;
;
in the method, in the process of the application,representing Greedy Best First Search a heuristic, the value of h (n) satisfies a true consumption value from "square" n to "target square",the distance of the diagonal is indicated as such,representing the manhattan distance of the person,indicating the number of steps movable along the diagonal line,represents the euclidean origin x coordinate value,representing the euclidean endpoint x coordinate value,represents the euclidean origin y coordinate value,represents the euclidean origin and destination y coordinate values,representing the diagonal heuristic optimal path estimate,representing the cost of the diagonal movement,representing the cost of moving one step horizontally or vertically in the map, (x) A , y A ）、(x B , y B ) A, B, respectively.
The industrial safety intelligent inspection machine dog based on the AI algorithm comprises an inspection path planning kinematic modeling unit, an inspection path planning environment modeling unit and an inspection path planning algorithm optimizing unit, wherein the inspection method comprises the following steps of:
the inspection path planning kinematic modeling unit is used for solving the motion space range of the foot end of the industrial inspection machine dog by adopting a method of combining kinematic forward solution and kinematic inverse solution, so that the inspection path of the industrial inspection machine dog is planned, and places with stairs and slopes are planned into the inspection path.
The inspection path planning environment modeling unit is used for carrying out inspection path planning environment modeling by adopting a grid method, and the inspection information and the obstacle information of the industrial inspection robot dog are known through a small grid, so that the grid information is directly corresponding to the actual environment information.
The routing inspection path planning algorithm optimizing unit is used for establishing a routing finding heuristic function, weighting the heuristic function and obtaining an optimized path according to the weighted heuristic function.
Preferably: the intelligent algorithm package unit is used for image recognition, face recognition, natural language processing, data mining, knowledge graph modeling and service.
Preferably: the intelligent recognition system comprises a recognition scene unit, wherein the recognition scene unit is used for one or more of flame recognition, smoke recognition, violation intrusion recognition, personnel off duty recognition, personnel falling recognition, safety helmet wearing recognition, protective clothing wearing recognition, mask wearing recognition, smoking recognition, calling recognition, vehicle license plate recognition, channel occupation recognition, personnel gathering recognition, duty personnel sleep duty recognition and meter reading recognition.
Compared with the prior art, the application has the following beneficial effects:
according to the intelligent inspection system, an inspection route planning algorithm is optimized, an AI recognition algorithm package is implanted in an industrial inspection machine, an industrial inspection machine dog is adopted to replace manual safety inspection in an enterprise factory workshop, industrial safety intelligent inspection is realized, and the defect part of intelligent inspection in the industry is complemented. The objectivity of the enterprise safety inspection result is improved, and the inspection route execution efficiency is improved.
Drawings
Fig. 1 is a kinematic positive solution map of the foot end range of motion of a machine dog.
Fig. 2 is a diagram of an inverse solution to the kinematics of the foot end range of motion of a machine dog.
Fig. 3 is a schematic view of a patrol area, where o represents a start node and a target node, and o represents an obstacle in the environment.
Fig. 4 is a schematic diagram of diagonal distance.
Fig. 5 is a schematic diagram of a A, B two-point movement.
Fig. 6 is a conventional AI algorithm path search.
Fig. 7 is a weighted optimized AI algorithm path search.
Fig. 8 is an experimental simulation of 22 conventional AI algorithms for obstacles.
Fig. 9 is an experimental simulation of the conventional AI algorithm for 34 obstacles.
Fig. 10 is an experimental simulation of 30 conventional AI algorithms for obstacles.
Fig. 11 is an experimental simulation of 22 obstacle-optimizing AI algorithms.
Fig. 12 is an experimental simulation of 34 obstacle-optimizing AI algorithms.
Fig. 13 is an experimental simulation of 30 obstacle-optimizing AI algorithms.
Detailed Description
The present application is further illustrated in the accompanying drawings and detailed description which are to be understood as being merely illustrative of the application and not limiting of its scope, and various equivalent modifications to the application will fall within the scope of the application as defined in the appended claims after reading the application.
An industrial safety intelligent inspection method for inspecting dogs based on an AI algorithm comprises the following steps:
step 1, a patrol path planning kinematic modeling is carried out, a method of combining kinematic forward solution and kinematic inverse solution is adopted to obtain a motion space range of a dog foot end of an industrial patrol machine, so that a patrol path of the industrial patrol machine dog is planned, and places with stairs and slopes are planned into the patrol path.
As shown in fig. 1, the kinematic forward process:
write-firstIs the equation expression of (2):
; (1)
rewriting outIs the equation expression of (2):
; (2)
wherein,the coordinates of the joint are represented,representing the length of the lower leg,the rotation amount is represented, the rotation angle of the shank steering engine is represented,the coordinates of the foot end of the robot are represented,indicating the length of the thigh of the person,the rotation amount is represented, and the rotation angle of the thigh steering engine is represented.
And (3) obtaining the foot end coordinates of the machine dog according to the kinematic forward solution, and conversely, when the foot end coordinates of the machine dog are known, obtaining the angles of the thighs and the calves of the machine dog by using the kinematic reverse solution.
As shown in fig. 2, the inverse kinematics solution process:
solving the angle of the lower leg:
; (3)
according to the cosine law:
; (4)
; (5)
obtaining the steering angle of the shank steering engine。
Wherein,represents the projection distance of the connecting line of the steering engine and the foot end on the xOy plane,and the steering angle of the shank steering engine is indicated.
Solving the angle of the thigh:
according to the cosine law:
; (6)
; (7)
; (8)
; (9)
; (10)
; (11)
obtain the steering wheel corner of the thigh。
Wherein,the rotation quantity between the connecting line of the foot end of the steering engine and the lower leg is represented as the intermediate quantity.
And 2, modeling a routing inspection path planning environment, namely performing routing inspection path planning environment modeling by adopting a grid method, knowing routing inspection information and barrier information of the industrial routing inspection machine dog through a small grid, and directly corresponding grid information with actual environment information. The motion on the grid defines 8 motion directions, the definition value of each cell is 0 or 1,1 represents the obstacle area, and 0 represents the free area.
The simulation environment is simulated by Matlab writing operation under Windows 10 environment, the size of a patrol area set by the simulation environment is 50 multiplied by 50, an included angle in the simulation environment is taken as a coordinate origin, the abscissa represents the length of the simulation environment, the ordinate represents the width of the simulation environment, in FIG. 3, O represents a starting node and a target node, O represents an obstacle in the environment, and 10 target nodes are arranged in the intercepted patrol area. The specific inspection area is shown in fig. 3.
Step 3, optimizing routing inspection path planning algorithm
Establishing a path finding heuristic function, and solving by adopting diagonal distanceValue and pair ofAnd (3) performing weighted optimization:
; (12)
wherein,representing the overall priority of node n,representing the cost of the node n from the start point,representing the cost of the node n from the endpoint,a weighting coefficient of h (n) in the range of 0,1]. When h (n) =0, h (n) has no effect on f (n), but shortest path planning can still be achieved. When h (n) =1, the value of h (n) becomes smaller, the expanded "square" increases, the calculation speed decreases, but the distance of the travel path is shortest. The h (n) value becomes larger, the expanded "square" is reduced, the calculation speed is increased, but the distance of the travel path is not necessarily shortest.
The value of h (n) is smaller than or equal to the actual consumption value from the square grid n to the target square grid, and the AI algorithm can be realized when the value of h (n) is close to the actual value or even the value of h (n) is equal to the actual value.
The calculation formula of (a) is shown as a formula (13) -a formula (15):
; (13)
; (14)
; (15)
in the method, in the process of the application,representing Greedy Best First Search heuristic, the value of h (n) satisfies a value less than or equal to the true consumption value from "square" n to "target square，The distance of the diagonal is indicated as such,representing the manhattan distance of the person,indicating the number of steps movable along the diagonal line,represents the euclidean origin x coordinate value,representing the euclidean endpoint x coordinate value,represents the euclidean origin y coordinate value,indicating the euclidean origin and destination y coordinate values. Where h_diagonal (n) represents the number of steps movable along the diagonal line; h_bar (n) represents the manhattan distance; the sum of all steps movable along the diagonal and all straight-line steps (i.e. manhattan distance steps minus 2 times diagonal) is added to the diagonal distance, and the path planned is shown in fig. 4.
As shown in fig. 5, a is a start point and B is an end point.
; (16)
; (17)
; (18)
In the method, in the process of the application,representing the diagonal heuristic optimal path estimate,representing the cost of the diagonal movement,representing the cost of moving one step horizontally or vertically in the map, (x) A , y A ）、(x B , y B ) A, B, respectively. Equations (16) - (18) represent the costs of moving diagonally, i.e., a cost estimate of the diagonal distance can be calculated. The results of the run in Matlab are shown in figure 6.
As shown in fig. 2-6, the optimization algorithm needs too many expansion points to be calculated, and in the calculation of path search, where many invalid child nodes around a starting point and an end point participate in the calculation of path search, the calculation time of path search is increased, the search efficiency is greatly reduced, and the algorithm is further optimized: the heuristic is weighted.
Weighting the heuristic functions:
; (19)
where h (n) represents the cost of node n from the endpoint, e h(n) Representing the optimized weighting coefficients, the estimated consumption gradually tends to 0 and the weighting coefficients gradually tend to 1 as the current node gradually approaches the target point. Fig. 7 shows the path search of the AI algorithm after the weighting optimization, and compared with fig. 6, the number of extension points is greatly reduced, the operation time is reduced, and the operation efficiency is improved.
The industrial safety intelligent inspection machine dog based on the AI algorithm comprises an inspection path planning kinematic modeling unit, an inspection path planning environment modeling unit and an inspection path planning algorithm optimizing unit, wherein the inspection method comprises the following steps of:
the inspection path planning kinematic modeling unit is used for solving the motion space range of the foot end of the industrial inspection machine dog by adopting a method of combining kinematic forward solution and kinematic inverse solution, so that the inspection path of the industrial inspection machine dog is planned, and places with stairs and slopes are planned into the inspection path.
The inspection path planning environment modeling unit is used for carrying out inspection path planning environment modeling by adopting a grid method, and the inspection information and the obstacle information of the industrial inspection robot dog are known through a small grid, so that the grid information is directly corresponding to the actual environment information.
The routing inspection path planning algorithm optimizing unit is used for establishing a routing finding heuristic function, weighting the heuristic function and obtaining an optimized path according to the weighted heuristic function.
The intelligent algorithm package unit is used for image recognition, face recognition, natural language processing, data mining, knowledge graph modeling and service. The intelligent algorithm package unit is used for integrating NLP, image, data mining and other technologies to complete image recognition, face recognition, natural language processing, data mining, knowledge graph and other field models and algorithm services to form an intelligent algorithm package. The intelligent algorithm package is deployed on a main thread of a machine dog in a dock containerization deployment mode, abnormal events are classified, filtered and screened through a camera connected with a main control panel body, corresponding characteristic objects are obtained from a real-time video picture, and analysis, processing and analysis of a front-end video are realized. The system is decoupled from the model and brand of the front-end camera, and supports standard video coding and decoding protocols.
The intelligent recognition system comprises a recognition scene unit, wherein the recognition scene unit is used for one or more of flame recognition, smoke recognition, violation intrusion recognition, personnel off duty recognition, personnel falling recognition, safety helmet wearing recognition, protective clothing wearing recognition, mask wearing recognition, smoking recognition, calling recognition, vehicle license plate recognition, channel occupation recognition, personnel gathering recognition, duty personnel sleep duty recognition and meter reading recognition.
The control program of the designed industrial inspection machine dog adopts Py-Apple Dynamics, which mainly comprises padog. Py, PA_ATTITUDE. Py, PA_AVGFLY. Py, PA_IK. Py, PA_IMU. Py, PA_SERVO. Py, PA_STABLIZE. Py and PA_TROT. Py, and the functions of all library files are shown in the following table 2.
Table 2 library file function table
Py-Apple Dynamics is used as a core control program of the machine dog, and specific function implementation is shown in Table 3.
TABLE 3 Py-Apple Dynamics functionality
The industrial inspection machine dog is internally provided with WiFi, double control of a mobile terminal and a webpage terminal can be realized through a WiFi connection mode, and the control interface of the industrial inspection machine dog of the embodiment is written by using mPython software according to a program of a Boston dog open source.
The machine dog patrol path optimization algorithm can complete path optimization by adopting Dijkstra algorithm. Except that Dijkstra's algorithm focuses on the shortest path and AI's algorithm focuses more on the optimal path.
The intelligent analysis deployment of the machine dog video can also adopt the deployment of a server side, and the centralized analysis mode of a central side can realize scene recognition and alarm.
Simulation verification of the inspection path:
under the simulation environment, the method runs in the environment of 10 multiplied by 10 of the map grid, sequentially patrols and examines two target nodes in the simulation map, equipment is in a normal state in a default patrol process, and an industrial patrol machine dog patrol complete path.
Setting the same initial node, two same target nodes and 22 barriers in the first (1) group simulation area;
setting the same initial node, two same target nodes and 34 barriers in the second (2) group simulation area;
setting the same initial node, two same target nodes and 30 barriers in the third (3) group simulation area;
under the condition of the same environment map, the number of initial nodes, target nodes and barriers is changed for 3 times, the path running time and inflection point number of the traditional AI algorithm and the optimized AI algorithm are compared, and specific data are shown in table 1.
Table 1 experimental simulation data
8-10, 11-13 and Table 1 show that under the condition of the same environment map, the inflection point of the optimized AI algorithm is minimum, the running time is shorter, and compared with the traditional AI algorithm, the optimized AI algorithm has obvious optimization. Compared with the traditional AI algorithm, the number of inflection points of the (1) group optimization AI algorithm in the simulation data is reduced by 2, and the calculation time is shortened by 0.91s. (2) The number of inflection points of the group optimization AI algorithm is reduced by 3 compared with that of the traditional AI algorithm, and the calculation time is shortened by 1.168s. (3) The number of inflection points of the group optimization AI algorithm is reduced by 3 compared with that of the traditional AI algorithm, and the calculation time is shortened by 1.145s. The optimized AI algorithm is superior to the conventional AI algorithm in terms of the number of inflection points and calculation time.
The foregoing is only a preferred embodiment of the application, it being noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present application, and such modifications and adaptations are intended to be comprehended within the scope of the application.
Claims (7)
1. An industrial safety intelligent inspection method for inspecting dogs based on an AI algorithm is characterized by comprising the following steps:
step 1, a patrol path planning kinematic modeling is carried out, a method of combining kinematic forward solution and kinematic inverse solution is adopted to obtain a motion space range of a dog foot end of an industrial patrol machine, so that a patrol path of the industrial patrol machine dog is planned, and places with stairs and slopes are planned into the patrol path;
step 2, modeling a routing inspection path planning environment, namely performing routing inspection path planning environment modeling by adopting a grid method, knowing routing inspection information and barrier information of an industrial routing inspection machine dog through a small grid, and directly corresponding grid information with actual environment information; defining 8 movement directions by movement on the grid, wherein the definition value of each cell is 0 or 1,1 represents an obstacle region, and 0 represents a free region;
step 3, optimizing routing inspection path planning algorithm
Establishing a path finding heuristic function, and solving by adopting diagonal distanceValue and pair->And (3) performing weighted optimization:
;
wherein,representing the comprehensive priority of node n, +.>Cost representing node n from the origin, +.>Cost representing node n from endpoint, +.>A weighting coefficient of h (n);
weighting the heuristic functions:
;
in the method, in the process of the application,cost representing node n from endpoint, +.>Representing the optimized weighting coefficients.
2. The method for inspecting the industrial safety intelligent inspection machine dog based on the AI algorithm, which is characterized in that: the kinematic positive solution process in step 1: write-firstEquation expression of (2)>Re-write outEquation expression of (2)>Wherein, the->Representing joint coordinates +.>Representing calf length, & lt>Indicating the rotation amount, the rotation angle of the shank steering engine, < ->Representing robot foot coordinates +.>Indicating thigh length, < >>The rotation amount is represented, and the rotation angle of the thigh steering engine is represented.
3. The method for inspecting the industrial safety intelligent inspection machine dog based on the AI algorithm as claimed in claim 2, wherein the method is characterized in that: the kinematic inverse solution process in step 1:
solving the angle of the lower leg:according to the cosine law: />;/>Obtaining the corner of the shank steering engine>The method comprises the steps of carrying out a first treatment on the surface of the Solving the angle of the thigh: according to the cosine law: />;/>;
;/>;/>;/>Obtaining the corner of the thigh steering engine>The method comprises the steps of carrying out a first treatment on the surface of the Wherein (1)>Representing the intermediate quantity, the projection distance of the connecting line of the steering engine and the foot end on the xOy plane, < >>Represents the steering wheel angle of the shank->The rotation quantity between the connecting line of the foot end of the steering engine and the lower leg is represented as the intermediate quantity.
4. The method for inspecting the industrial safety intelligent inspection machine dog based on the AI algorithm as claimed in claim 3, wherein the method comprises the following steps: in step 3The calculation method of (2) is as follows: />;;/>;;/>;/>In>Representing Greedy Best First Search heuristic, the value of h (n) satisfies the true that is less than or equal to "square" n to "target squareReal consumption value->Represents diagonal distance>Representing the manhattan distance of the person,indicates the number of steps movable along the diagonal line, < >>X-coordinate values representing Euclidean origin, < >>X-coordinate values representing euclidean endpoints, +.>Y-coordinate representing Euclidean origin, < >>Y-coordinate value representing Euclidean origin and destination, < >>Representing the optimal path estimate of the diagonal heuristic, < +.>Representing the cost of diagonal movement, +.>Representing the cost of moving one step horizontally or vertically in the map, (x) A , y A ）、(x B , y B ) A, B, respectively.
5. An industrial safety intelligent inspection machine dog based on an AI algorithm is characterized in that: the inspection method for the industrial safety intelligent inspection robot dog based on the AI algorithm, which comprises an inspection path planning kinematic modeling unit, an inspection path planning environment modeling unit and an inspection path planning algorithm optimizing unit, wherein:
the inspection path planning kinematic modeling unit is used for solving the motion space range of the foot end of the industrial inspection machine dog by adopting a method of combining kinematic forward solution and kinematic inverse solution, so that the inspection path of the industrial inspection machine dog is planned, and places with stairs and slopes are planned into the inspection path;
the inspection path planning environment modeling unit is used for carrying out inspection path planning environment modeling by adopting a grid method, and knowing inspection information and barrier information of the industrial inspection robot dog through a small grid, and directly corresponding grid information with actual environment information;
the routing inspection path planning algorithm optimizing unit is used for establishing a routing finding heuristic function, weighting the heuristic function and obtaining an optimized path according to the weighted heuristic function.
6. The AI-algorithm-based industrial safety intelligent patrol robot dog of claim 5, wherein: the intelligent algorithm package unit is used for image recognition, face recognition, natural language processing, data mining, knowledge graph modeling and service.
7. The AI-algorithm-based industrial safety intelligent patrol robot dog of claim 6, wherein: the intelligent recognition system comprises a recognition scene unit, wherein the recognition scene unit is used for one or more of flame recognition, smoke recognition, violation intrusion recognition, personnel off duty recognition, personnel falling recognition, safety helmet wearing recognition, protective clothing wearing recognition, mask wearing recognition, smoking recognition, calling recognition, vehicle license plate recognition, channel occupation recognition, personnel gathering recognition, duty personnel sleep duty recognition and meter reading recognition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310936318.XA CN117193288A (en) | 2023-07-28 | 2023-07-28 | Industrial safety intelligent inspection robot dog based on AI algorithm and inspection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310936318.XA CN117193288A (en) | 2023-07-28 | 2023-07-28 | Industrial safety intelligent inspection robot dog based on AI algorithm and inspection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117193288A true CN117193288A (en) | 2023-12-08 |
Family
ID=89004243
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310936318.XA Pending CN117193288A (en) | 2023-07-28 | 2023-07-28 | Industrial safety intelligent inspection robot dog based on AI algorithm and inspection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117193288A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111504325A (en) * | 2020-04-29 | 2020-08-07 | 南京大学 | Global path planning method based on weighted A-algorithm for expanding search neighborhood |
| CN112325897A (en) * | 2020-11-19 | 2021-02-05 | 东北大学 | Path planning method based on heuristic deep reinforcement learning |
| CN115329922A (en) * | 2022-09-22 | 2022-11-11 | 太原理工大学 | Coal mine underground multifunctional four-legged robot digital twin system and operation method thereof |
| WO2023024279A1 (en) * | 2021-08-24 | 2023-03-02 | 深圳市优必选科技股份有限公司 | Dynamic obstacle avoidance method and apparatus for robot, and robot |
-
2023
- 2023-07-28 CN CN202310936318.XA patent/CN117193288A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111504325A (en) * | 2020-04-29 | 2020-08-07 | 南京大学 | Global path planning method based on weighted A-algorithm for expanding search neighborhood |
| CN112325897A (en) * | 2020-11-19 | 2021-02-05 | 东北大学 | Path planning method based on heuristic deep reinforcement learning |
| WO2023024279A1 (en) * | 2021-08-24 | 2023-03-02 | 深圳市优必选科技股份有限公司 | Dynamic obstacle avoidance method and apparatus for robot, and robot |
| CN115329922A (en) * | 2022-09-22 | 2022-11-11 | 太原理工大学 | Coal mine underground multifunctional four-legged robot digital twin system and operation method thereof |
Non-Patent Citations (3)
| Title |
|---|
| ZJGUILAI: "机器人路径规划总结", pages 2 - 4, Retrieved from the Internet <URL:httpsblog.csdn.netzjguilaiarticledetails89478928> * |
| 勾文浩;袁立鹏;宫赤坤;高彤;: "基于足端轨迹的仿生四足机器人运动学分析与步态规划", 现代制造工程, no. 07, 18 July 2017 (2017-07-18) * |
| 派大星先生C: "四足机器人运动学逆解", pages 2, Retrieved from the Internet <URL:httpsblog.csdn.netm0_58585940articledetails121358455ops_request_misc=&request_id=&biz_id=102&utm_term=%E5%9B%9B%E8%B6%B3%E6%9C%BA%E5%99%A8%E4%BA%BA%E8%BF%90%E5%8A%A8%E5%AD%A6%E9%80%86%E8%A7%A3&utm_medium=distrib> * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108256577B (en) | Obstacle clustering method based on multi-line laser radar | |
| CN108564202A (en) | A kind of unmanned boat Route optimization method based on environmental forecasting information | |
| CN103716324B (en) | Virtual mine risk-taking behavior implementation system and method based on multiple agents | |
| Liu et al. | Map-based deep imitation learning for obstacle avoidance | |
| WO2008014831A2 (en) | Method for observation of a person in an industrial environment | |
| CN116540726A (en) | Intelligent obstacle avoidance method, system and medium for patrol robot | |
| Mokhtari et al. | Safe deep q-network for autonomous vehicles at unsignalized intersection | |
| WO2023082492A1 (en) | Path expansion and passage method for manned robot in terminal building | |
| CN117193288A (en) | Industrial safety intelligent inspection robot dog based on AI algorithm and inspection method | |
| Hu et al. | Toward intelligent workplace: Prediction-enabled proactive planning for human-robot coexistence on unstructured construction sites | |
| CN104097205A (en) | Task space based self-collision avoidance control method for real-time movements of robot | |
| Fan et al. | A fire protection robot system based on SLAM localization and fire source identification | |
| CN112817318B (en) | Multi-unmanned-boat collaborative search control method and system | |
| CN111538343B (en) | System, method and storage medium for robot to set traffic rules | |
| CN113311709A (en) | Intelligent wheelchair compound control system and method based on brain-computer interface | |
| Kaiser et al. | Affordance-based multi-contact whole-body pose sequence planning for humanoid robots in unknown environments | |
| Xiang et al. | An extension of BIM using AI: A multi working-machines pathfinding solution | |
| Mokhtari et al. | Pedestrian collision avoidance for autonomous vehicles at unsignalized intersection using deep q-network | |
| Tang | From agent to avatar | |
| Chen et al. | Connect-based subgoal discovery for options in hierarchical reinforcement learning | |
| Bai et al. | Toward cooperative multi-robot control for detecting and tracking an expanding flood area | |
| Minaeian et al. | Human-in-the-loop agent-based simulation for improved autonomous surveillance using unmanned vehicles | |
| Joshi et al. | Real time motion tracking algorithm for search and rescue robots | |
| Chen et al. | A vision-based dual-axis positioning system with yolov4 and improved genetic algorithms | |
| Freire et al. | Mobile robot navigation based on the fusion of control signals from different controllers |
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 |