CN116757253B - Intelligent automatic learning door opening and closing algorithm for rail transit - Google Patents
Intelligent automatic learning door opening and closing algorithm for rail transit Download PDFInfo
- Publication number
- CN116757253B CN116757253B CN202310817264.5A CN202310817264A CN116757253B CN 116757253 B CN116757253 B CN 116757253B CN 202310817264 A CN202310817264 A CN 202310817264A CN 116757253 B CN116757253 B CN 116757253B
- Authority
- CN
- China
- Prior art keywords
- door
- closing
- door opening
- parameters
- parameter
- 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.)
- Active
Links
- 230000009471 action Effects 0.000 claims abstract description 47
- 230000006870 function Effects 0.000 claims abstract description 14
- 238000000034 method Methods 0.000 claims description 37
- 230000008569 process Effects 0.000 claims description 34
- 238000012545 processing Methods 0.000 claims description 9
- 238000003062 neural network model Methods 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 230000006854 communication Effects 0.000 claims description 5
- 238000004891 communication Methods 0.000 claims description 4
- 238000001125 extrusion Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 238000005299 abrasion Methods 0.000 abstract 1
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000009434 installation Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Power-Operated Mechanisms For Wings (AREA)
Abstract
The invention discloses an intelligent automatic learning door opening and closing algorithm for rail transit, which belongs to the technical field of rail transit vehicle industry and solves the problems that: the vehicle door has different driving parameters each time due to abrasion, assembly errors and the like, and the application range of the same parameter is limited greatly; the key point of the technical scheme is that the real-time parameter K 1 is collected according to the actual action of the door opening and closing, and compared with the predicted parameter K 2, if the result is within the range of algorithm errors, the door opening and closing at the time is considered to be reliable, accurate and safe. And then recording and learning, dynamically generating a new parameter K 3, taking the new parameter K 3 as the input quantity of the next predicted parameter, and performing loop execution in the similar way. Generating new prediction parameters before each door opening and closing, and comparing the new prediction parameters with actual parameters, thereby completing the self-adaptive learning function of each door opening and closing, and achieving the following effects: the motor driving parameters can be automatically learned and adjusted so as to estimate and meet the application requirements in real time.
Description
Technical Field
The invention relates to the technical field of rail transit vehicle industry, in particular to an intelligent automatic learning door opening and closing algorithm for rail transit.
Background
Along with the high-speed development of the rail transit industry, more and more subway vehicles are manufactured and used, the vehicle operation and departure intervals are shorter and shorter, the daily operation time is synchronously prolonged, and the daily opening and closing times of the vehicle doors are gradually increased. In long-time operation and use, various mechanical parts of the vehicle door can be worn to different degrees, or due to assembly errors in installation, parameters such as door opening and closing time, motor moment, motor current and the like when the vehicle door performs door opening and closing actions at different moments can be different. Therefore, false triggering of the anti-extrusion function can occur in the normal door opening and closing process, false reporting of faults is achieved, and serious operation faults such as late train and the like are finally caused.
Disclosure of Invention
In view of the above-mentioned drawbacks of the prior art, the present invention aims to provide an intelligent automatic learning door opening and closing algorithm for rail transit, which can greatly alleviate door opening and closing deviation caused by long-time operation or differences in mechanical structures of different doors. The algorithm takes an automatic learning model as a framework, each door takes parameters (door opening and closing time, motor moment, motor current and the like) during normal operation for a plurality of times before the door is taken as system input, and then the optimal estimated quantity of the next door opening and closing is automatically predicted, so that a dynamic algorithm for door opening and closing action difference caused by factors such as mechanical fatigue, installation errors and the like among different doors is eliminated.
In order to achieve the above object and other related objects, the present invention adopts the following technical scheme: an intelligent automatic learning door opening and closing algorithm for rail transit,
Step 1, data acquisition and comparison: the method comprises the steps of performing driving control on a door motor according to an acquired power-on door closing action instruction, collecting a real-time motor driving parameter K 1 according to the actual action of a door, comparing the real-time parameter of the motor with a predicted parameter K 2, and considering that the door is opened and closed at the time to meet the standard if the result is within an algorithm error range;
step 2, recording and learning, namely dynamically generating a new parameter K 3 according to a historical motor driving parameter K 1 which accords with the standard, wherein the new parameter K 3 can be used as the input quantity of the next predicted parameter and is circularly executed;
And step 3, generating new prediction parameters before opening and closing the door each time, comparing the new prediction parameters with actual parameters, finishing output through a neural network model, and realizing self-learning and control on the door opening action of the motor.
Preferably, after the control board is powered on, the singlechip is initialized to obtain the current software version number, then the current software version is compared with the server to determine whether the current software version is the latest version, if the current software version is not the highest in the linked service or version number, the next door power-on and door-closing action is performed, and otherwise, the upgrading program is started to update and upgrade.
Preferably, when entering the main program, the function process program is started; the functional process program includes:
the data communication bus process is used for receiving and transmitting CAN data;
Interrupt processing process for inputting sampling data;
the door action process is used for providing output parameters to control the motor to perform door opening and closing actions, extrusion preventing functions and non-zero speed door closing actions;
The storage and recording process is used for storing logs, faults and automatically learning parameters through an asynchronous storage function;
And automatically learning a neuron process, wherein the process adopts a three-layer structure according to a neuron network model, and performs self-learning and control on corresponding motor current according to door opening actions and motor codes.
Preferably, the action parameters after the door is correctly opened and closed in the past are read as the input to the neural network model to guide the next actual door opening and closing, the actual parameters of the action are recorded, after judgment, the actual parameters are stored into a storage device if the actual parameters are correct, and the actual parameters are used as a part of parameters when the door is opened and closed in the next time, and the next door opening and closing action is corrected and guided by a method of continuously and circularly learning the correct parameters.
Compared with the background art, the invention has the technical effects that:
design errors, installation errors, fatigue wear and the like of the mechanical mechanism are likely to cause inconsistent motor current when the door is opened and closed at the same coding position. The non-uniform mapping of motor current to force level is embodied as: torque variability. Therefore, even the doors of the same train, the doors of the same carriage and the moment for opening and closing the doors are not consistent. This also results in that in the case of consistent underlying code, each door action may instead be inconsistent. Meanwhile, under the condition that the early warning values are the same, the early warning of some doors can be caused, and the potential safety hazard can be brought to the running of the vehicle under the condition that some doors do not have the early warning. After the bottom layer singlechip is electrified, the action parameters after the door is correctly opened and closed in the past are read firstly to guide the next actual door opening and closing, the actual parameters of the action are recorded, after judgment processing, the actual parameters are stored into a memory if the actual parameters are correct, and are used as a part of parameter basis for the next door opening and closing action, and a method of continuously learning the correct parameters is adopted to correct and guide the next door opening and closing action.
Drawings
The invention will be further described with reference to the drawings and examples.
FIG. 1 shows a software flow diagram of an embodiment of the present invention;
Fig. 2 shows a model of a neural network according to an embodiment of the invention.
Detailed Description
Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be noted that the following embodiments and features in the embodiments may be combined with each other without conflict.
It should be noted that the illustrations provided in the following embodiments merely illustrate the basic concept of the present invention by way of illustration, and only the components related to the present invention are shown in the drawings and are not drawn according to the number, shape and size of the components in actual implementation, and the form, number and proportion of the components in actual implementation may be arbitrarily changed, and the layout of the components may be more complicated.
The following describes the embodiments of the present invention further with reference to the drawings.
Examples:
The intelligent automatic learning door opening and closing algorithm for rail traffic uses automatic learning model as structure, and each door uses its own parameters (door opening and closing time, motor moment, motor current, etc.) as system input, and then automatically predicts the optimal estimated quantity of next door opening and closing, so as to eliminate the door opening and closing action difference caused by mechanical fatigue, installation error, etc. among different doors. Aiming at the requirements of safety, stability and the like of vehicle operation, a prediction parameter of door opening and closing operation is provided, a real-time parameter K 1 is collected according to the actual action of the door opening and closing at this time and is compared with a prediction parameter K 2, and if the result is within the range of algorithm errors, the door opening and closing at this time is considered to be reliable, accurate and safe. And then recording and learning, dynamically generating a new parameter K 3, taking the new parameter K 3 as the input quantity of the next predicted parameter, and performing loop execution in the similar way. And generating new prediction parameters before each door opening and closing, and comparing the new prediction parameters with actual parameters, thereby completing the self-adaptive learning function of each door opening and closing.
Step 1, data acquisition and comparison: the method comprises the steps of performing driving control on a door motor according to an acquired power-on door closing action instruction, collecting a real-time motor driving parameter K 1 according to the actual action of a door, comparing the real-time parameter of the motor with a predicted parameter K 2, and considering that the door is opened and closed at the time to meet the standard if the result is within an algorithm error range;
step 2, recording and learning, namely dynamically generating a new parameter K 3 according to a historical motor driving parameter K 1 which accords with the standard, wherein the new parameter K 3 can be used as the input quantity of the next predicted parameter and is circularly executed;
And step 3, generating new prediction parameters before opening and closing the door each time, comparing the new prediction parameters with actual parameters, finishing output through a neural network model, and realizing self-learning and control on the door opening action of the motor.
Referring to fig. 1 and fig. 2 specifically, after the control board is powered on, the single-chip microcomputer is initialized, including a clock, peripherals, a timer, input/output IO, motor driving, and the like. And acquiring the current software version number, comparing whether the current software version is the latest version or not by communicating with the server, if the current software version is not linked with the service or the version number is the highest, performing the next door power-on and door-closing action, and if not, starting the updating program to update.
When entering the main program, starting a function process program; the functional process program includes:
a data communication bus process (CAN bus receiving and transmitting process) for receiving and transmitting CAN data;
an interrupt processing process (INPUT listening process) for inputting sampled data, wherein the INPUT quantity is changed, and corresponding functions, such as functions of door opening and closing, zero speed, emergency unlocking, service buttons and the like, can be triggered;
a door action process (door action executing function process) for providing output parameters to control the motor to perform door opening and closing actions, anti-extrusion actions, non-zero speed door closing actions and emergency unlocking door opening and closing actions;
The storage and recording process is used for storing logs, faults and automatically learning parameters through an asynchronous storage function;
And automatically learning a neuron process, wherein the process adopts a three-layer structure according to a neuron network model, and performs self-learning and control on corresponding motor current according to door opening actions and motor codes. The method comprises the steps of firstly reading action parameters after a door is correctly opened and closed in the past as input to a neural network model to guide the next actual door opening and closing, recording the actual parameters of the action, storing the actual parameters into a storage after judging and processing, and correcting and guiding the next door opening and closing action by a method of continuously and circularly learning the correct parameters as a part of parameters when the door is correctly opened and closed in the next time. An alarm may be given when the trigger and error range of the comparison are exceeded.
As can be seen in fig. 1, there is also an OUTPUT control process for control of OUTPUT. And the fault alarm control process is used for processing faults and errors. And the UART process is used for controlling the work of the single-port communication process. And the clock process is used for processing the real-time clock. TCPIP process for the processing of network communications.
In connection with fig. 2, the leftmost layer is used for receiving external information, and is called an input layer, and the rightmost layer is the final output after being processed by the neural network and is called an output layer. The layer of all nodes in the middle is used for transformation computation, but no specific computation process is seen, called hidden layer, and their values cannot be observed in the training sample set. It can also be seen that in this scheme there are 3 input units (dimension 3, no offset unit), 3 hidden units and one output unit (generating input). The 3 input units are respectively: current sampling value, motor coding value, input quantity IO value.
Let Lx denote the total number of layers in the network, in this embodiment x=3, and let L1 be the input layer and L3 be the output layer, if we denote layer 1 as L1. For example, the neural network has training parameters (W, b) which may be (W1, b1, W2, b 2), where W1 is the connection parameter between the j-th element of layer 1 and the i-th element of layer 1+1, and the bias element is not input and is processed in the algorithm.
The neural network can obtain highly abstract information and characteristics which cannot be obtained manually in a self-learning mode. The algorithm is a dynamic algorithm, and under the same version of program, each door can generate corresponding dynamic parameters according to the mechanical condition of the door, so that the difference of door opening and closing actions is reduced, the universal adaptability of the program is increased, and a revolutionary breakthrough is achieved in the field of rail traffic control.
The above embodiments are merely illustrative of the principles of the present invention and its effectiveness, and are not intended to limit the invention. Modifications and variations may be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the invention. Accordingly, it is intended that all equivalent modifications and variations of the invention be covered by the claims, which are within the ordinary skill of the art, be within the spirit and scope of the present disclosure.
Claims (1)
1. An intelligent automatic learning door opening and closing algorithm for rail transit is characterized in that:
Step 1, data acquisition and comparison: the method comprises the steps of performing driving control on a door motor according to an acquired power-on door closing action instruction, collecting a real-time motor driving parameter K 1 according to the actual action of a door, comparing the real-time parameter of the motor with a predicted parameter K 2, and considering that the door is opened and closed at the time to meet the standard if the result is within an algorithm error range;
step 2, recording and learning, namely dynamically generating a new parameter K 3 according to a historical motor driving parameter K 1 which accords with the standard, wherein the new parameter K 3 can be used as the input quantity of the next predicted parameter and is circularly executed;
Step 3, generating new prediction parameters before opening and closing the door each time, comparing the new prediction parameters with actual parameters, finishing output through a neural network model, and realizing self-learning and control on the door opening action of the motor; after the control board is electrified, initializing the singlechip to acquire the current software version number, comparing whether the current software version is the latest version or not by communicating with the server, and carrying out the next door electrification and door closing action if the current software version is not linked with the service or the version number is the highest, otherwise, starting the updating and upgrading of the upgrading program; when entering the main program, starting a function process program; the functional process program includes:
the data communication bus process is used for receiving and transmitting CAN data;
Interrupt processing process for inputting sampling data;
the door action process is used for providing output parameters to control the motor to perform door opening and closing actions, extrusion preventing functions and non-zero speed door closing actions;
The storage and recording process is used for storing logs, faults and automatically learning parameters through an asynchronous storage function;
Automatically learning a neuron process, wherein the process adopts a three-layer structure according to a neuron network model, and performs self-learning and control on corresponding motor current according to door opening actions and motor codes; the method comprises the steps of firstly reading action parameters after a door is correctly opened and closed in the past as input to a neural network model to guide the next actual door opening and closing, recording the actual parameters of the action, storing the actual parameters into a storage after judging and processing, and correcting and guiding the next door opening and closing action by a method of continuously and circularly learning the correct parameters as a part of parameters when the door is correctly opened and closed in the next time.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310817264.5A CN116757253B (en) | 2023-07-03 | 2023-07-03 | Intelligent automatic learning door opening and closing algorithm for rail transit |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310817264.5A CN116757253B (en) | 2023-07-03 | 2023-07-03 | Intelligent automatic learning door opening and closing algorithm for rail transit |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116757253A CN116757253A (en) | 2023-09-15 |
CN116757253B true CN116757253B (en) | 2024-06-11 |
Family
ID=87957030
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310817264.5A Active CN116757253B (en) | 2023-07-03 | 2023-07-03 | Intelligent automatic learning door opening and closing algorithm for rail transit |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116757253B (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101226377A (en) * | 2008-02-04 | 2008-07-23 | 南京理工大学 | Robust control method for asphalt mixing plant batching error |
CN107834912A (en) * | 2017-11-08 | 2018-03-23 | 郑州精益达汽车零部件有限公司 | A kind of electronic Passenger door device and control method |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105068515B (en) * | 2015-07-16 | 2017-08-25 | 华南理工大学 | A kind of intelligent home device sound control method based on self-learning algorithm |
CN109033993A (en) * | 2018-06-29 | 2018-12-18 | 南京行者易智能交通科技有限公司 | A kind of method and device of image recognition detection switch door |
CN109179171B (en) * | 2018-10-31 | 2021-11-19 | 苏州汇川控制技术有限公司 | Self-learning method, equipment and storage medium for elevator door mechanical system |
KR102235728B1 (en) * | 2019-05-28 | 2021-04-05 | 한국철도기술연구원 | Fault prediction apparatus and method of electric type side entrance door of electric train |
CN110688982B (en) * | 2019-10-14 | 2022-07-01 | 重庆邮电大学 | Intelligent rail transit time control method based on target detection technology and ACO-BP algorithm |
US11823057B2 (en) * | 2019-12-30 | 2023-11-21 | Dalian University Of Technology | Intelligent control method for dynamic neural network-based variable cycle engine |
WO2022059098A1 (en) * | 2020-09-16 | 2022-03-24 | 三菱電機ビルテクノサービス株式会社 | Elevator control system |
CN113642461A (en) * | 2021-08-13 | 2021-11-12 | 辽宁华盾安全技术有限责任公司 | Elevator door opening and closing state identification method and device based on deep learning |
CN116127839A (en) * | 2023-01-04 | 2023-05-16 | 广东工业大学 | Bus station service interruption prevention and control method based on reinforcement learning |
-
2023
- 2023-07-03 CN CN202310817264.5A patent/CN116757253B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101226377A (en) * | 2008-02-04 | 2008-07-23 | 南京理工大学 | Robust control method for asphalt mixing plant batching error |
CN107834912A (en) * | 2017-11-08 | 2018-03-23 | 郑州精益达汽车零部件有限公司 | A kind of electronic Passenger door device and control method |
Also Published As
Publication number | Publication date |
---|---|
CN116757253A (en) | 2023-09-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Carvalho et al. | Exploiting the use of recurrent neural networks for driver behavior profiling | |
JP3379983B2 (en) | Artificial intelligence traffic modeling and prediction system | |
CN114585983B (en) | Method, device and system for detecting abnormal operation state of equipment | |
KR950005875B1 (en) | Automatic operation controller having protective function | |
US11531899B2 (en) | Method for estimating a global uncertainty of a neural network | |
CN107153413A (en) | A kind of automobile power back door integrated diagnostic system | |
CN104536435A (en) | Online diagnosis method for linear control system network | |
CN110658905B (en) | Early warning method, early warning system and early warning device for equipment operation state | |
CN112308336A (en) | High-speed railway high wind speed limit dynamic disposal method based on multi-step time sequence prediction | |
CN113746696A (en) | Network flow prediction method, equipment, storage medium and device | |
JP2021056983A (en) | Method and device for detecting driver abnormalities based on machine learning using vehicle can bus signal | |
CN112763967A (en) | BiGRU-based intelligent electric meter metering module fault prediction and diagnosis method | |
CN116757253B (en) | Intelligent automatic learning door opening and closing algorithm for rail transit | |
CN114559959A (en) | Automatic driving behavior decision system and method based on driving preference | |
CN112579759A (en) | Model training method and task type visual dialogue problem generation method and device | |
CN112001482A (en) | Vibration prediction and model training method and device, computer equipment and storage medium | |
CN116258242A (en) | Reactive track prediction method and system for automatic driving vehicle | |
CN115390455A (en) | Safety reinforcement learning method and system based on sequential logic | |
Kwon et al. | Hidden Markov model‐based real‐time transient identifications in nuclear power plants | |
CN112232557B (en) | Short-term prediction method for health degree of switch machine based on long-short-term memory network | |
CN114930287B (en) | Controller for a vehicle, system, method and motor vehicle having a controller | |
Liu et al. | Safe model-based control from signal temporal logic specifications using recurrent neural networks | |
CN114810292A (en) | Computer-implemented method and apparatus for maneuver identification of exhaust aftertreatment systems | |
CN108804771B (en) | Discrete event system method for analyzing and designing human-computer interaction logic | |
Das | An efficient way to enable prognostics in an onboard system |
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 |