JPH07215606A - Intelligent decentralized control device for elevator - Google Patents

Abstract

PURPOSE: To reduce a development cycle, facilitate maintenance, and increase reliability by characterizing the modularity, configurability, and easiness of a distributed control device. CONSTITUTION: A distributed processing unit(DPU) is applied to an elevator dispatch backup architecture. The distributed processing unit(DPU) is designed as a floor processing device for each floor a box processing device in each box 2, a group processing device and a signal processing device in a machine room 22. All these distributed processing devices have an intelligence to perform a communication and local control algorithm. The distributed control device is configured as a neural network, and a node is executed by the distributed processing unit(DPU).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is provided with a hall call registration device arranged on each floor for transmitting a hall call for elevator service in each elevator, exchanging signals with each elevator, A box including a group control device for controlling the operation of the elevator in response to a hall call and a signal received from the elevator, and further including a box call registration device for a service required by a passenger, and A box drive for performing and restraining the operation of the box, moving the box upward or downward selected in response to a signal indicative of the state of the box and a signal received from the group controller, and A box controller for providing the signal indicating the operation of the box to control the operation of the box to stop the box, and for each elevator, A plurality of elevators conventionally configured to perform services for multiple floors of an article, said group controller responding to said signals indicative of respective states of said boxes, hall calls being generated A new elevator box, which is provided with signal processor means for performing an evaluation calculation for each box when selected, and an optimal elevator box is selected based on the evaluation calculation result and dispatched to answer the hall call. To an improved intelligent distributed controller.

[0002]

Conventionally, elevator controllers are centralized systems, the operation of which is controlled by one intelligent station in the system. This station may be located in the machine room, inside the operator's cab or on the top.

Therefore, such a control device has 19
It became known from US Pat. No. 5,305,198 issued April 19, 1994 (IP389) and assigned to the applicant. In this system, the target call is
Limited and immediate allocation to individual elevators to perform calls in response to higher and lower functional requirements. This allocation is immediately indicated on the call input floor. The weighted sum corresponding to the higher functional requirements is formed from the partial operating costs, which is modified by the variable bonus penalty point factor to the operating costs in the sense of the lower functional requirements, and the target call is the elevator with the lowest operating cost. Is allocated to. The method is implemented on an industrial computer with a target call allocation algorithm that includes dependent algorithms for bonus penalty point tracking and cost calculation. This calculation of operating costs is done by the cost calculation algorithm according to a special cost formula and the rebalanced bonus
The penalty point factor multiplicatively acts on the 6-term partial cost sum. The prior art uses a computer as a centralized control unit for an elevator group of three elevators A, B, C, and is implemented by target calls only, without box calls. The different elements of the elevator group are connected to the controlling computer by a bus system. That is, the measuring and adjusting elements are connected to the computer by the elevator bus, the elements in each box by the box bus, and the floor-by-floor decimal keyboard for entering the target call by the floor bus. Elevator buses, box buses, and floor buses form a triple bus system connected to a computer by a special interface. The centralized computer may be any suitable computer, such as a separate computer for each elevator box, one of which also controls the group functions, a single computer for the entire elevator group, a single computer for two or more elevator groups. It can be configured. Common to all these computer configurations is a more or less centralized controller.

The main drawback of this type of system configuration is that despite the use of a bus system, substantial wiring effects are required. Wiring must be done from a distant location in the system to a central location. Not only is this process expensive, error prone and cumbersome. Another drawback of such a system is that it requires a lot of work for software development, testing, and maintenance. Expanding an elevator system, that is, increasing the number of floors and boxes, can overload central computers and other elevator controls. In this case, the load is not averaged and the computer processing efficiency of the entire system is low.
When using a single central station for elevator control, the control functions and data processing loads cannot be distributed and averaged.

[0005]

SUMMARY OF THE INVENTION It is therefore an object of the present invention to overcome the above mentioned drawbacks by providing a fully modular system with greatly reduced overall wiring work. Also, the control device according to the present invention shall be configured to exhibit increased system efficiency and reliability for all functions relating to banking. This problem is solved by the invention by the means characterized in the version of the independent claim. Advantageous developments are indicated in the dependent claims.

[0006]

The problems and drawbacks of prior art centralized control units are solved by the present invention by using a distributed processing unit (DPU) which also provides the following advantages. The first advantage can be seen in the modularity and configurability of the distributed controller. Elevator controllers based on DPU localize processing according to function. This results in a controller whose functions are performed locally in the respective parts of the control system. Using commonly defined interfaces, different functionality can be added simply by adding the required hardware and DPU. This allows the desired functionality without affecting other DPUs,
That is, a system with modularity and configurability will be easier to create. by this,
In the future, minimal engineering work will also add functionality for a given job. Another advantage can be seen in the simplification of the control structure as a result of distributed processing. The control system is divided into different modules, each performing a defined set of tasks. Therefore, the development and maintenance of the control system is simplified due to the modularity of this system. By specifying a pre-defined interface, each module can be developed independently and in parallel, reducing overall system development time. This reduces time to market, reduces development costs, and simplifies software maintenance. It has also been found that the system according to the invention exhibits improved overall reliability and safety, since there is no single point of failure and a reduced mode of operation is possible.

These and other advantages of the invention will be readily apparent to those of ordinary skill in the art by reviewing the following detailed description of the preferred embodiments, in view of the accompanying drawings.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENT In FIG. 1, elevators of an elevator group are designated A, B and C, a box 2 is guided by an elevator shaft 1 for each elevator and services for 16 floors E1 to E16. It is driven in a known manner via a winding cable 4 by a winding motor 3 to carry out Each drive device 3 is controlled by a drive control device, whereby the target value generation, the adjusting function and the start / stop operation are all carried out by the industrial computer 5. Measuring and adjusting element 6
Is connected to the industrial computer 5 by a first interface IF1 and an elevator bus 7.
Each box includes a load measuring device 8 for determining when a passenger enters and leaves the elevator box, a box display device 9 for notifying each operating state Z of the box, a stop indicator 10, and a box. The operation panel 11 is included. Devices 8, 9,
The units 10 and 11 are connected to the computer 5 via a box bus 12. Box calls are recorded in the elevator boxes A, B, C by a suitable push button arrangement incorporated in the box operating panel 11. The call is then serialized,
It is transmitted to the industrial computer 5 via the box bus 12 and the interface CIF together with the related information of other boxes.

Floors E1 to E16 include the lowest floor E1
"Up" hall call push button 14 located at, "down" hall call push button 15 located at the top floor E16, "up" located at each middle floor E2 to E15
Also provided is a call registration device 8 in the form of a suitable push button 13, such as a "down" hall call push button 16. Like box calls, hall calls are serialized, floor bus 17 and input interface ICF.
Is transmitted to the industrial computer 5 via the, and uses a special hall call allocation algorithm,
In the sense of the required function profile, it is assigned to the services to the individual boxes 2.

FIG. 2 shows an elevator control system based on the distributed control concept. This control system is Echel
Uses Neuron Chip based local operating networks (Lonworks) by On Corporation. This is a new technique that promises a better opportunity to create low cost distributed control systems that include unique means of implementing distributed control algorithms. The basic principle of this technique is that distributed processing equipment (D
PU) including a system. This DP
U has the intelligence to detect and control local devices and send updates to other DPUs in the system. All the distributed processing units DPU operate automatically, independently and autonomously under the control of software stored in each DPU. The DPUs are interconnected via one or more available communication media 18, thereby sharing a common message-based communication protocol. This limits the wiring to the local level,
The status of the device can be transmitted sequentially to different points in the system. Therefore, the overall wiring work is significantly reduced. Each DPU in such a system has the intelligence to execute control algorithms as well as the intelligence to send messages between different points in the system. By distributing the control functions to different DPUs, it is not necessary to have a central processor.

This is shown in FIG. In this figure, the control system consists of several intelligent DPUs interconnected by a communication medium 18. The number of DPUs will vary depending on the requirements of the control system. The control system should be configured with respect to the required processing power, not with respect to the number of elevators and floors. However, the function performed by any given DPU does not change. 3 floors E1, E2, E
3, the floor treatment device F is installed in the corridor fixing box 20 of each floor E1, E2, E3 and in the fire control center 21.
PU is located. Located in the elevator car 2 is a carton processor CPU for door operations, position indicators, landing system, and car call functions. Another group processing unit GPU and a signal processing unit SPU are located in the machine room 22. The floor processing unit FPU and the box processing unit CPU execute functions related to the sensor device and the actuator device, and execute necessary control functions. These devices also broadcast the latest status of important devices to notify other DPUs in the system.
The information thus received by other DPUs along with the state of the local device is used in making the control decisions for which the DPU is responsible. Group processing unit GPU
Mainly includes a hall call assignment function, and the signal processing unit SPU relates to a transmission function for inter-process communication.
Therefore, it is possible to create a completely modular system that includes distributed control capabilities.

An example of how to perform a typical elevator operation with such a system is shown in FIG. When the passenger presses the hallway call button 25 on the third floor, the floor processing device FPUa located in the hallway fixed box 20 on the floor recognizes that the elevators A, B, and C are operable, if any. To latch. This information is then used by the first group processor GPU in the machine room 22.
sent to a. Based on the evaluation calculation result, the group processing unit GPUa executes the evaluation calculation for determining the hall call assignment to the box 2 and controls the drive unit 26 that controls the movement of the elevators A, B and C. Responsible for performing group management of Box 2. When the information about the request on the third floor is received, the group processing device GP
Ua initiates a command to move the assigned elevator, which in this case is elevator A, towards the third floor. As elevator A moves through the building,
Box processor CPUc that monitors the landing system
Has elevator A on each floor E1. ． ． The other distributed processing unit DPU is updated when passing through. Upon receiving this information, the floor processing unit FPUa on the third floor changes the position indicator 27 to the appropriate floor position and orientation. At the same time, the first group processor GPUa also receives the landing system update from the box processor CPUc and, on the basis of this information, decides whether to continue running or to stop on the next floor. When the next floor is the target floor, the second group processing device GPUb determines that the drive device 2
Change the command to 6 to stop at that floor. When the elevator starts decelerating, the second group processor GPUb broadcasts this information to the entire network. Upon receiving this information, the floor processor FPUa on the third floor cancels the corridor call in response and at the same time turns on the hall lantern 28 for it. When the elevator A arrives at the correct horizontal position, the group processing unit GPUb moves to the landing system box processing unit CPUc.
From which the drive 26 is instructed to stop the elevator there, and the second group processor GPUb informs the door processors CPUa and CPUb that the elevator has stopped at the target floor. CPUa and CPU responsible for door operation at this point
b commands the door to open and then close after a predetermined time. All this action completes the request on the third floor. The typical elevator operation described above involves the exchange of data between various distributed processing units DPU. For this,
Signal processing for causing the group processors GPUa and GPUb to communicate with each other via the first data field 29 and the box processors CPUa, CPUb, ... To communicate with each other via the second data field 30. Device S
PUa, SPUb, ... Are provided. As shown in the example above, a fully modular system can be created without the need for a central processing unit designated 5 in FIG. By distributing the processing responsibility of the system,
The task is much simpler and the load balance can be averaged out. This simplicity is directly reflected in the system development cycle and ease of maintenance. In order to balance the load, a distributed processor with a heavy load can actually be temporarily exempted from executing the process. Even if one of the group processing units GPUa, GPUb, ... Is operative, it is possible to carry out the group control and thereby guarantee reliability in this respect. Therefore, high reliability and high system efficiency are achieved via cooperative distributed control implemented by the individual distributed processing units DPU.

FIG. 3 illustrates the neural network structure used to implement distributed control for the elevator dispatch backup architecture. The multi-box dispatch model was selected by applying the neuron MC143150 chip for low cost and high performance. The boxes to be assigned to hall calls are selected based on the results obtained using a neural net corresponding to neurons in the human brain. The neural net includes an input layer, an output layer, and an intermediate layer provided between the input layer and the output layer. In the middle layer, weighting factors are applied when combining the signals from the nodes in the input layer and when distributing the signals to the output nodes. The weighting factors are variable and are appropriately modified and corrected through learning to make more reasonable box allocations. Different sets of learning factors can be applied at different times or under different conditions of detected passenger load. Learning (correction of the network) can be performed using the backpropagation method.
Backpropagation is a method of correcting the weighting factor using the error between the output data of the network and the desired output data created from statistical data or control target values. Each primary node includes its own estimated arrival time software for executing the token ring algorithm for interbox communication and a designed backup architecture. Both of these are bank services.
Guarantee reliability against e).

In this dispatch model, node A
0, B0, C0, D0, ... are the main nodes (MC143150) for the boxes of all banks, and the nodes A1, A
2, A3, ... are auxiliary nodes for each box (MC14312
0) and the nodes Aa, Ba, Ca, Da, ... Are backups for the main nodes of all boxes A, B, C, D.
It is a node (MC143150).

In order to guarantee reliability for banking operations, rather than having one node calculate the ETA, each main node executes its own token ring algorithm for interbox communication. Includes ETA calculation software. The neuron network management also ensures parallelism of signal transmission between the auxiliary nodes A1, A2, A3, ... Or the main nodes A0, B0, C0 ,. 3150 chip has 6 for user program
It provides 4 KB and communicates with other nodes via the network common port 34.

For important nodes such as the main nodes A0, B0, C0, ..., additional backup nodes Aa, Ba, Ca, ... Can be used as shown. Each node pair 35 consists of a main node and a backup node. Each node pair starts for incoming data and also follows outgoing data by multiplexer 36. On the backup node side, the consistency checking software always sends signals to check the network common port 34 and application I / O port 37 of the main node. If an error is found, the backup node Aa, Ba, C for data communication from the node pair 35.
A signal for selecting a, ... Is transmitted. When this system is introduced, every node has its own ID.
Backup nodes Aa, B
a, Ca, ... Also have IDs. When a common error occurs, all interested nodes can look for a backup node.

The invention is not limited to the particular embodiments described above.
Various changes and modifications may be made without departing from the spirit and scope of the invention.

[Brief description of drawings]

FIG. 1 is a conventional elevator, which uses an industrial computer to perform centralized control for three elevators.
It is a schematic diagram of a group.

FIG. 2 is a schematic diagram of an elevator system using neural network based distributed control according to the present invention.

FIG. 3 is a diagram of a neural network structure used to implement distributed control according to the present invention.

[Explanation of symbols]

 A, B, C Elevator, Elevator group 1 Elevator shaft 2 Box 3 Drive device 4 Hoisting cables E1, ..., E16 16 floor 5 Industrial computer 6 Measurement setting member IF1 1. Interface 7 Elevator bus 8 Load measuring device Z Operating state 9 Operating state notification device 10 Stop indicator 11 Box operation panel 12 Box bus CIF interface 13 Push button 14 Up hall call push button 15 Down hall call push button 16 up Down call call push Button 17 Floor bus DPU Distributed processing device 18 Communication medium FPU Floor processing device 20 Corridor fixed box 21 Fire control center CPU box processing device GPU group processing device SPU signal processing device 22 Machine room 23 Sensor device 24 Actuator device 25 Corridor call button ( 3rd floor) FPUa Floor treatment equipment: 3rd floor FPUb Floor treatment equipment: 2nd floor FPUc Floor treatment equipment: 1st floor FPUd Fire control treatment equipment GPUa 1st group treatment Processing device GPUb Second group processing device 26 Drive device CPUa Front door box processing device CPUb Back door box processing device CPUc Landing system Box processing device CPUd Position display box processing device CPUe Box calling function box processing device 27 Position indicator GPUd No. 3 group processor 28 Hall lantern 29 First data field 30 Second data field A0, B0, C0. ． ． Main nodes A1, A2, A3. ． ． Auxiliary nodes Aa, Ba, Ca. ． ． Backup node 34 Network common port 35 Node pair (main node / backup node) 36 Multiplexer 37 Application I / O port

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Lichyard Day Garhardson United States, New Jersey 07869, Randolph, Center Grove Road A, 19, 44 (72) Inventor Edward El Chiyo United States , New Jersey 07430, Mahwah, Makurund Tussyu Drive 148

Claims (6)

[Claims] 1. A hall call registration device arranged on each floor for transmitting a hall call for a hoisting service in each elevator, exchanging signals with each elevator, and hall call and from the elevator. A group controller for controlling the operation of the elevator in response to received signals, and further showing a box, a box drive for performing and restraining the operation of the box, and a state of the box. Indicating the state of the box to control the operation of the box to move the box upwards or downwards selected and stop the box in response to signals and signals received from the group controller. A plurality of elevators conventionally configured to perform services for multiple floors of a building, including for each elevator a box controller that provides a signal. The group control device, in response to the signal indicating the respective states of the box, comprises a signal processor means for performing an evaluation calculation for each box when a hall call is generated, in the evaluation calculation result An intelligent decentralized controller for elevators in which an optimal elevator box is selected based on and dispatched to answer the hall call, wherein the intelligent decentralized controller comprises a distributed processing unit (DPU), preferably an input It is configured as a neural network including a layer, an intermediate layer and an output layer, and the nodes of the neural network are distributed processing devices (DP
U) and for controlling the corridor fixed box (20) on each floor to autonomously input / output information related to said floor (E1, E2, ...).
Floor processing units (FPUa, FPU)
b ,. ． ． ) For controlling each box (2) and autonomously inputting / outputting information related to the box (2).
B ,. ． ． ) Each box (2) of the box processing device (C
PUa, CPUb ,. ． ． ), And is located in the machine room (22) to perform the evaluation calculation to determine the hall call allocation to the box (2) and to perform the group management of the box (2) based on the evaluation calculation result. Group processing device (GPUa, GPUb, ...)
And the group processing device (GPUa, GPUb, ...)
And the floor processing device (FPUa, FPU
b ,. ． ． ), One or more signal processing units (SPUa, SPU provided for communicating with each other)
b ,. ． ． ) And an intelligent distributed controller. 2. A first door processor (CPUa) for operating a front door, a second door processor (CPUb) for operating a back door, and a landing system processor (CPUc) for a landing system. ), A position processing device (CPUd) for position display, and a box call processing device (CPUe) for box calling function are provided in each box. 3. The intelligent distributed control device according to claim 1, wherein the device status is transmitted to various distributed processing units (DPUs) in the system sequentially. 4. The intermediate layer comprises a first weighting factor between an individual node of the input layer and an individual node of the intermediate layer,
The intelligent distributed control device according to claim 1, further comprising a second weighting factor between the individual node of the middle layer and the individual node of the output layer. 5. The neural network comprises a main node (A0, B0, C0, ...) And an auxiliary node (A1, A).
2, A3 ,. ． ． ), Whereby each primary node (A0, B0, C0, ...) Is connected in parallel to a backup node (Aa, Ba, Ca, ...). Intelligent distributed control device described in. 6. Each main node (A0, B0, C
0 ,. ． ． 2. The intelligent distributed controller of claim 1, wherein) includes the node's own ETA software for executing a token ring algorithm for interbox communication.