GB2348528A

GB2348528A - Counting apparatus

Info

Publication number: GB2348528A
Application number: GB0007461A
Authority: GB
Inventors: Knud Erik Jermer; Peter Sorensen
Original assignee: Honeywell Control Systems Ltd
Current assignee: Honeywell Control Systems Ltd
Priority date: 1999-03-31
Filing date: 2000-03-29
Publication date: 2000-10-04
Anticipated expiration: 2020-03-29
Also published as: GB0007461D0; GB2348528B

Abstract

A railway traffic control system 1 has a databank of valid sensing signals for a sensing location having at least three sensors. Patterns of sensor outputs for significant positions are produced and used to produce state diagrams.

Description

COUNTING METHOD AND APPARATUS The present invention relates to a method of, and apparatus for, counting objects.

The present invention is particularly suited for the monitoring of railway carriages on railway lines whereby the number of carriage wheels entering a sensing zone is counted and compared with the count of wheels leaving the zone, but the present invention is also applicable to analogous situations with similar monitoring of different objects ; it is equally suited to situations in which there is required merely the cumulative counting of objects moving past a point on a predetermined path.

A conventional system for counting railway carriages has two sensors such that, depending on the order of signals output from these sensors, the system is able to determine the direction of the movement between the two sensors. As the direction cannot be determined by only one sensor, the counting function is inoperable if there is a fault on either sensor.

The present invention provides a method of counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the method comprising : a) defining a set of sensing characterisation values for objects to be sensed ; b) for a first characterisation value, producing a pattern of outputs from a single sensing location for a first significant position of an object sensed relative to the sensing location ; c) producing another pattern of sensor outputs for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) repeating steps (b) and (c) for all the remaining significant relative positions of the object and the sensing location ; e) for at least a second sensing characterisation value, producing a sequence of output patterns in accordance with steps (b) to (d).

The method may include any one or more of the features of dependent Claims 2 to 14.

The present invention also provides a databank produced by a method of the present invention.

The present invention also provides a computer program product stored on a computer usable medium, comprising : a) computer readable program means for causing a computer to define a set of sensing characterisation values for objects to be sensed ; b) computer readable program means for causing the computer to, for a first characterisation value, produce a pattern of outputs from a single sensing location for a first significant position of an object sensed relative to the sensing location ; c) computer readable program means for causing the computer to produce another pattern of sensor outputs for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) computer readable program means for causing the computer to repeat the steps indicated as (b) and (c) hereinabove for all the remaining significant relative positions of the object and the sensing location ; and e) computer readable program means for causing the computer to, for at least a second sensing characterisation value, produce a sequence of output patterns in accordance with steps (b) to (d).

In a variant, there may be provided a corresponding program for mechanical, chemical, optical, analogue or other computing.

By"software code", there can be included program code and lookup tables.

The present invention further provides apparatus for counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the apparatus comprising : a) means to define a set of sensing characterisation values for objects to be sensed ; b) means, for a first characterisation value, to produce a pattern of outputs from a single sensing location for a first significant position of an object sensed relative to the sensing location ; c) means to produce another pattern of sensor output for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) means to repeat steps (b) and (c) for all the remaining significant relative positions of the object and the sensing location ; e) means, for at least a second sensing characterisation value, to produce a sequence of output patterns in accordance with steps (b) to (d).

The apparatus may include any one or more of the features of dependent Claims 21 to 31.

The present invention is particularly suited to the monitoring of railway carriages on a railway line by the counting of wheels of the carriages as they pass individual sensors within a sensing zone defined by a group of sensors (typically 4 sensors, but also groups of 5 or 6 sensors are particularly beneficial and preferred, and in any event a group is not limited to those specific numbers). However, the present invention is also suited to a wide variety of other applications, for example : vehicles on a road ; manufactured goods on a conveyor belt ; food cans/bottles on a bottle process line ; components on a manufacturing/assembly line ; luggage/items on a cargo-handling conveyor belt.

A wide variety of types of sensor and techniques of sensing can be used readily in the present invention, including (but not limited only to) : proximity sensors, light sensors, radar sensors, magnetic sensors, infra-red sensors, mechanical sensors, ultrasonic proximity sensors.

The present invention is not limited to binary sensing whereby a sensor merely detects the presence or absence of an object, but also encompasses tertiary or more states of sensing whereby a sensor detects three or more possible states for example : i) absence of any object ; ii) presence of object of type one (eg. coloured blue) ; iii) presence of object of type two (eg. coloured red).

The method may include defining sensing characterisation values for : i) sensing resolution of a sensor ; or ii) separation of sensors ; or iii) sensing the nature of an object.

The nature of an object can be any one or more important aspects of an object, for example the size, colour, cross-section, profile, density, opacity, reflectivity, speed, separation.

A wide variety of electrical and electronic components and computer hardware, software and firmware can be used to implement parts of, and functional elements of, the present invention in various forms according to individual applications as appropriate, including mainframe computers, pcs, hard disks, RAMs, EPROMs.

According to another aspect of the present invention, there is provided a method of counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the method comprising : a) defining a set of sensing characterisation values for objects to be sensed ; b) for at least one characterisation value, producing a pattern of outputs from a single sensing location for a first significant position of an object relative to the sensing location ; c) producing another pattern of sensor outputs for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) repeating step (c) for all the remaining significant relative positions of the object and the sensing location ; e) when an outputs pattern is produced, determining if the pattern is a permitted transition from the previous pattern ; f) if the pattern is a permitted transition, continuing the pattern production operation ; and g) for at least a second sensing characterisation value, producing a sequence of outputs in accordance with steps (b) to (f).

This method may comprise any one or more of the following features : if the pattern is not a permitted transition, initiating an error-routine. the error-routine comprises suspending further repetition of steps (c) and (d). the error-routine further comprises monitoring for a permitted pattern to occur. the error-routine further comprises monitoring for a neutral pattern. once the monitoring step determines a specified pattern, the pattern production operation is resumed. comprising determining if the pattern is a permitted transition by consulting a store of permitted patterns. determining if the pattern is a permitted transition, by consulting the store of output patterns produced by the method before consulting the store of permitted patterns. producing a sequence of output patterns relating to travel in one direction only. producing a sequence of output patterns including travel in two opposite directions. at least one change in direction of travel in a single sensing sequence which corresponds to the completed motion of the object at the sensing location. producing a sequence of output patterns with a noise signal occurring, in turn, at each sensor location. producing a sequence of output patterns for a single fault for each sensor in turn, the fault being one or more of the following : i) any one of the sensors is permanently sensed ; ii) any one of the sensors is permanently unsensed ; iii) transition from"no fault"to"fault"for any sensor ; iv) transition from"fault"to"no fault"for any sensor. producing the patterns manually. producing the patterns by computer. producing the patterns by computer and checking with manually generated signal patterns, or vice versa. checking for duplications of patterns produced in different circumstances. the patterns are converted to representations in a state diagram.

* defining sensing characterisation values for : 1) sensing resolution of a sensor ; and/or 2) separation of sensors ; and/or 3) sensing the nature of an object. a sensing location has 4 sensors.

'objects to be sensed comprise wheels of railway carriages on a railway track.

According to this aspect of the present invention, there is also provided apparatus for counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the apparatus comprising : a) means to define a set of sensing characterisation values for objects to be sensed ; b) means, for at least one characterisation value, to produce a pattern of outputs from a single sensing location for a first significant position of an object relative to the sensing location ; c) means to produce another pattern of sensor output for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) means to repeat step (c) for all the remaining significant relative positions of the object and the sensing location ; e) means, when an outputs pattern is produced, to determine if the pattern is a permitted transition from the previous pattern ; f) means, if the pattern is a permitted transition, to continue the pattern production operation ; and g) means, for at least a second sensing characterisation value, to produce a sequence of outputs in accordance with steps (b) to (f).

The apparatus may comprise any one or more of the following features : means, if the pattern is not a permitted transition, to initiate an error routine. the error-routine means comprises means to suspend further repetition of steps (c) and (d).

'the error-routine means comprises means to monitor for a permitted pattern to occur. the error-routine means comprises means to monitor for a neutral pattern to occur. means, once the monitoring means determines a specified pattern, the pattern production operation is resumed. means to determine if the pattern is permitted by consulting a store of patterns already produced by the apparatus before consulting the store of permitted patterns.

'means to produce a sequence of output patterns relating to travel in one direction only. means to produce a sequence of output patterns including travel in two opposite directions. the patterns include those relating to at least one change in direction of travel in a single sensing sequence which corresponds to the completed motion of the object at the sensing location.

< the patterns include those relating to a noise pulse occurring, in turn, at each sensor location. the patterns include those relating to a single fault for each sensor in turn, the fault being one or more of the following : i) any one of the sensors is permanently sensed ; ii) any one of the sensors is permanently unsensed ; iii) transition from"no fault"to"fault"for any sensor ; iv) transition from"fault"to"no fault"for any sensor. means to check computer-generated results with manually-generated signal patterns, or vice versa.

< means to check for duplications of patterns produced in different circumstances.

< means to convert the patterns to representations in a state diagram. means to define sensing characterisation values for : i) sensing resolution of a sensor ; and/or ii) separation of sensors ; and/or iii sensing the nature of an object.

# a sensing location has 4 sensors. objects to be sensed comprise wheels of railway carriages on a railway track.

According to another aspect of the present invention, there is provided a method of counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the method comprising : a) defining a set of sensing characterisation values for objects to be sensed ; b) for at least a first characterisation value, producing patterns of outputs from a single sensing location for significant positions of an object sensed moving relative to the sensing location ; c) comparing the patterns to detect duplication of patterns ; and d) when a duplication is detected, making appropriate modification to the patterns and/or databank ; The method may include any one or more of the following features : the step of comparing patterns is effected upon completion of the production of a new pattern. the new pattern is compared with the valid patterns already produced. the new pattern is compared with a store of valid patterns.

'the modification step comprises linking the sets of different circumstances which result in the same pattern. the patterns of outputs are converted to representations in a state diagram. upon production of each new pattern, it is compared with the existing patterns and, if it is different to the existing patterns, it is converted to a new representation in a state diagram. upon production of each new pattern, it is compared with the existing patterns and, if it is the same as an existing pattern, it is converted to that existing representation in a state diagram. the patterns and/or diagrams are stored in a databank. producing a sequence of output patterns to travel in one direction only. producing a sequence of output patterns including travel in two opposite directions. at least one change in direction of travel in a single sensing sequence which corresponds to the completed motion of the object at the sensing location. producing a sequence of output patterns with a noise signal occurring, in turn, at each sensor location. producing a sequence of output patterns for a single fault for each sensor in turn, the fault being one or more of the following : i) any one of the sensors is permanently sensed ; ii) any one of the sensors is permanently unsensed ; iii) transition from"no fault"to"fault"for any sensor ; iv) transition from"fault"to"no fault"for any sensor. producing the patterns and/or sequences manually.

< producing the sequences by computer. producing the sequences by computer and checking with manually generated signal patterns, or vice versa. defining sensing characterisation values for : 1) sensing resolution of a sensor ; and/or 2 separation of sensors ; and/or 3. sensing the nature of an object. a sensing location has 4 sensors. objects to be sensed comprise wheels of railway carriages on a railway track.

According to this aspect of the present invention, there is provided apparatus for counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the apparatus comprising : a) means to define a set of sensing characterisation values for objects to be sensed ; b) means, for at least a first characterisation value, to produce patterns of outputs from a single sensing location for a first significant position of an object sensed moving relative to the sensing location ; c) means to compare the patterns to detect duplication of patterns ; and d) means, when a duplication is detected, to make appropriate modification to the patterns and/or databank.

The apparatus may include any one or more of the following features : means to compare patterns upon completion of the production of a new pattern. means to compare the new pattern with valid patterns already produced. means to compare the new pattern with a store of valid patterns. the modification step comprises linking the sets of different circumstances which result in the same pattern. means to convert the patterns of outputs to representations in a state diagram. means, upon production of each new pattern, to compare with the existing patterns and, if it is different to the existing patterns to convert it to a new representation in a state diagram. means, upon production of each new pattern, to compare with the exiting patterns and, if it is the same as an existing pattern, to convert it to that existing representation a state diagram.

'the patterns and/or state diagrams are stored in a databank.

'means to produce a sequence of output patterns relating to travel in one direction only. means to produce a sequence of output patterns including travel in two opposite directions. the patterns include those relating to at least one change in direction of travel in a single sensing sequence which corresponds to the completed motion of the object at the sensing location. the patterns include those relating to a noise signal occurring, in turn, at each sensor location. the patterns include those relating to a single fault for each sensor in turn, the fault being one or more of the following : i) any one of the sensors is permanently sensed ; ii) any one of the sensors is permanently unsensed ; iii) transition from"no fault"to"fault"for any sensor ; iv) transition from"fault"to"no fault"for any sensor. means to check computer-generated results with manually-generated signal patterns, or vice versa.

'means to eliminate duplications of patterns generated. means to define sensing characterisation values for : i) sensing resolution of a sensor ; and/or ii) separation of sensors and/or. iii) sensing the nature of an object. a sensing location has 4 sensors. objects to be sensed comprise wheels of railway carriages on a railway track.

Any of the aspects of the invention may include a computer program product directly loadable into the internal memory of the digital computer, comprising software code portions for performing the invention.

Any of the aspects of the invention may include electronic distribution of such a software program.

In order that the present invention may more readily be understood, a description is now given, by way of example only, reference being made to the accompanying drawings, in which : Figure 1 is a schematic diagram of a rail traffic control system ; Figure 2 is a block diagram of the train sensing system of the present invention being part of the system of Figure 1 ; Figure 3 is a more detailed diagram of part of the system of Figure 2 ; Figure 4 is a block diagram of the circuitry of a unit in Figure 2 ; Figures 5 (a) to (f) show sensing signals for movement of various wheel types ; Figure 6 shows part of a databank for signals sensed in accordance with Figure 5 ; Figure 7 shows an example of a state diagram for the signal sequences as shown in Figure 6 ; and Figure 8 (a) to (c) show the construction of a state diagram.

Figure 1 shows a railway traffic control system, generally denoted by reference numeral 1, having a mainframe computer traffic controller 2 to which is input signals from a plurality (only three being shown) of railway line monitor units 3A, 3B, 3C, and various other sources of relevant data (for example data on historic operation/incidents of system 1, on timetables/schedules for passenger and/or freight movements, on historic weather conditions, on weather forecasts).

Each line monitor unit 3A, 3B, 3C has its respective designated railway line 4A, 4B, 4C and a number of sensing locations 5A, 5B, 5C positioned along each line typically at a separation of 1000 metres in addition to one near the end of each line ; in circumstances involving very great traffic, the separation between adjacent sensing locations may be reduced significantly, for example of the order of 200 metres. A line unit 3A, 3B, 3C has a respective train detector unit 6A, 6B, 6C corresponding to each sensing location 5A, 5B, 5C.

Figure 2 shows in greater detail the major elements of one sensing zone 5A and train detector unit 6A for one line 4A, namely four identical proximity sensors 10 positioned on th e side of one metal rail of the railway line 4A at a separation of 0. 125 metres between sensors. Outputs of sensors 10 pass to sensing-processor unit 11 which has two sets 12, 13 of sensing signals analysis units 14 such that, at any one time, one set (for example 12 as shown in Figure 3) is in operation monitoring the signals from sensing zone 5A while the other set (13 as shown in Figure 3) is being tested by traffic control system 2 to ensure that it is able to operate correctly. After 10 seconds, system 2 sends unit 11 an instruction effecting a change in mode whereby set 12 is switched to the test mode and set 13 is switched to the operational mode in which it monitors the signals from sensing zone SA. After a further 10 seconds, system 2 sends unit 11 a further instruction to change set 12 to the operational mode and set 13 to the test mode. This mode-switching of sets 12, 13 occurs every 10 seconds continuously throughout operation of train detector unit 6A. If an error in a set is detected during testing, then the system enters the "fail-safe"state.

The mode-switching operation operates to ensure that, at all times, there is continuous processing of sensing signals being output from sensing zone 5A, regardless of the type or intensity of train movements on line 4A. It is essential that, in order to maintain the exceptionally high standards of safety and reliability required of a train detecting system of the present invention, effective testing of the counting equipment must be achieved, and this arrangement of two sets 12, 13 ensures that this can be done without any interruption of the counting operation.

Each set 12, 13 of sensing signal analysis units has two units 14, both working with the signals from the sensing zone 5A. One is passing output to the true part of the comparator system 15 and the other passing output to the inverse part of the comparator system 15 to check for consistency. The comparator system 15 is included in the line unit 3A.

Each unit 14 has a databank 16 (typically an EPROM) of valid sensing signals, a level controller 17 (typically a Field Programmable Gate Array) to run sensing signal verifications and test routines and modeswitching, latches 18 and counters 19.

When one of the sets 12, 13 is switched into the test mode by the appropriate instruction signal from controller 3A, and units 14 of that set begin a test routine performed by its internal processor, the results of the tests being output to controller 3A for verification of the test and confirmation that the units 14 are operating correctly ; once these issues have been validated, controller 3A can then effect mode-switching of the units 12, 13 at the end the appropriate time interval (10 seconds).

Consider a situation whereby unit 12 is in the operational mode and is processing the signals being output from sensing zone 5A, while unit 13 is under test.

There are five different test modes which together demonstrate and ensure correct functioning of the entire hardware and software of units 14. Each test operation is directed to a specific area of the sensing signal analysis unit 14, with a degree of overlap between the test operations. The tests done include the following : 1. A test of all the output stages from unit 14 and a test that the unit can execute the databank logic ; 2. A test of the content of databank 16 including that the contents are intact and that unit 14 can take decisions in relation to the counting operation based on signals received from the sensing location and on the contents of databank 16 ; 3. A test of the calculator within unit 14 ; 4. A test of the latch 18 which is the key function of the Test Ring mode. This test of latch 18 sources a sequence of"one high"bit patterns through the 16-bit wide latch to demonstrate that it is functioning correctly ; 5. A test of the state address of the set 13 of units under test, achieved by copying the state address of set 12 units in the operational mode into the set 13 in the test mode, executing it immediately and latching both the original and the copied state address into the state address latch. In this way, this procedure confirms that the state address was copied and executed correctly. As a result, it is known that the two sets 12, 13 are running exactly in parallel, so that the set 12 is now ready to be switched to the test mode and set 13 switched into the operational mode.

As part of the last test above, set 13 goes into a Preset mode which prepares it for switching into operational mode synchronised with the state of set 12 before switching set 12 into the test mode. A synchronisation signal from operational set 12 to set 13 under test triggers the Preset mode ensuring synchronisation of the two sets 12, 13 in a few nanoseconds ; the synchronisation, transfer of state address and the latching of actual and copied state address are completed within a few microseconds, and in considerably less time than it takes for a unit 14 to execute one processing operation of the signals from a sensing location 5A so a switch sequence of units 14 does not influence, and cannot be affected by, monitoring of trains.

Figure 4 shows in greater detail the level controller unit 17, which controls all functional modes of unit 14, which again is controlled and monitored by the controller 3A. The functional modes are : -Operational mode. Unit 17 monitors normal operation of unit 14.

Components used are : 17F, J, L, M and N.

-Test Simulation mode. Unit 17 tests that unit 14 can execute the databank logic unit 16, can operate latch unit 18 and can output to unit 19.

Components used are : 17F, J, L, M and N.

-Test Binary mode. Unit 17 tests that databank logic unit 16 is valid.

Components used are : 17A, C, G, H, I, F, J, M and N.

-Test Calculation mode. Unit 18 tests that the calculator unit 20 is valid.

Components used are : 17A, C, G, H, I, F, J, M and N.

-Test Ring mode. Unit 17 tests that latch unit 18 is functioning correctly.

Components used are : 17A, B, D, E, F, J, M and N.

-Test Preset mode. Unit 17 tests that ie. set 12 unit 14 can take over state address from set 13 unit 14. Components used are : 17F, J, K, L, M and N.

The features of the specific example described with reference to the Figures can be implemented in a variety of ways in practice, and may make use of technically equivalent procedures and techniques as appropriate including EPROMs, RAMs, computer hardware, software, firmware ; Table 1 provides exemplary but not limitative implementations.

In the first part of the process for producing the data on sensing signals, the four sensors 10A in a sensing zone are named A, B, C, and D. Each sensor can be in two states :"sensed" (namely it detects a wheel above the sensor) and"unsensed" (it detects no wheel present).

The method is described with reference to Figures 5 to 7. The width of a wheel can differ so, for the purpose of the production of this data, the wheels are divided into the following types, based on the number of sensors which can be sensed at the same time by a given wheel : Type 1. Only one sensor can be sensed at the same time ; Type 2. Two sensors can be sensed at the same time ; Type 3. Three sensors can be sensed at the same time ; Type 4. Four sensors can be sensed at the same time.

In addition there are types intermediate the above-mentioned types, these limiting types are named"1. 5" ;"2. 5" ; and"3. 5".

Figure 5 shows the movement pattern for the individual wheeltypes in the direction from sensors A to D, and also shows the resulting pulse sequences, caused when that wheel-type passes the sensors.

In Figure 5a, the four sensors are illustrated by four boxes, named A, B, C and D, and the pair of horizontal lines under the boxes illustrates the possible positions of the wheel during the movement, a sensor being sensed when the line is under the sensor.

The horizontal lines also represent the possible movements of a wheel. So the first line represents a wheel, which comes from the left, stops as soon sensor A is sensed, and then reverses and returns to the left ; the second line represents a wheel which stops just before sensor B is sensed ; the third line represents a wheel which stops as soon as sensor B is sensed, etc. The last line in each Figure represents a wheel which passes all the sensors, i. e. does not reverse.

Under the horizontal lines, the resulting sensor output sequences are shown. In these sequences,"0"means that no sensors are sensed ;"A" means that sensor A is sensed ;"AB"that both sensor A and B are sensed, etc.

Figures 5b to 5f show the equivalent movement patterns for the other wheel-types and the corresponding resultant signal sequences.

Table 1 below illustrates the possible sequences for all significant positions for the individual wheel-types, when the wheel is passing from sensor A to sensor D (no reversing). The pulses are represented by a block of four digits, each having the value 1"if the corresponding sensor is sensed, and"0"if the corresponding sensor is unsensed.

TABLE 1 : Signal sequences for wheel-types

TYPE 1.5 2 2.5 3 3.5 4 SEQUENCE 0000 0000 0000 0000 0000 0000 1000 1000 1000 1000 1000 1000 0100 1100 1100 1100 1100 1100 0010 0100 0110 1110 1110 1110 0001 0110 0011 0110 0111 1111 0000 0010 0001 0111 0011 0111 0011 0000 0011 0001 0011 0001 0001 0000 0001 0000 0000 0000 This process of producing the possible sequences for individual wheel-types is repeated, but this time for movement of the wheel when passing through sensing zone 5A in the direction from sensor D to sensor A.

The process is repeated again, but this time for a shunting movement whereby the wheel passes into sensing zone 5A in one direction, halts while in sensing zone SA, changes direction of movement and then exists zone 5A from the same side that it entered. The signal sequence for each variant of this motion in respect of the previous permutations is determined, including movement from each side, halting at each position, for each wheel-type.

The process is repeated again, but this time it is assumed that, for each of the above variants, there is a faulty sensor, and so the signal sequence is determined for when each of four forms of fault occurs at each of the sensors A, B, C, D in turn. The four forms of fault are : i) any one of the sensors is permanently"sensed" ; ii) any one of the sensors is permanently"unsensed" ; iii) transition from"no fault"to"fault"for any sensor ; iv) transition from"fault"to"no fault"for any sensor.

The process is repeated, but this time assuming that a noise pulse occurs, in turn, in each sensor output.

All the sensor output patterns are compiled in a data-table with identification as to the respective circumstances, Figure 6 showing a section of the data-table. In the shown section of the data-table which concerns sequences relating to faults for wheel-type 3 and involving movement from sensor A to sensor C and back to sensor A for mode 1, the possible faults on the sensors are listed in the column"Sensor Faults" (not complete in the shown section), the sensor output sequences in the columns 1-17 and the incidents attributable to a given sequence in the column"Output". By"mode 1", there is specified which of a number of possible movements from A to C to A is involved ; thus, for example, "mode 1"may be represented by line 5 in Figure 5d, while"mode 2" may be represented by line 6 in Figure 5d. In the lines 2-25 in the "Sensor Faults"column, the"permanently-sensed"sensors are represented by capital letters, e. g."C"for sensor C being"permanently- sensed", while"permanently-unsensed"sensors are represented by small letters, e. g. b for sensor B being"permanently-unsensed". In the "Output"column, the incidents are described by a capital letter, followed by a small letter. The capital letter describes the type of the incident ("C" for count ;"F"for a transition to a fault and"W" (for"working") for a transition from a fault to a normal-working condition. The small letter indicates to which sensor the incident is related (a-d).

It can be seen that certain of the lines are repeated within the table 1, e. g. line &num;1 and line &num;9, and therefore the table can be rationalised somewhat. Consolidation of the signal sequences results in a data-table having more than 12000 patterns and 200 state diagrams.

The contents of the data-table is represented in a set of state diagrams. Figure 7 shows an example of these diagrams. The implementation of the data-table in the state diagrams further reduces the table by deleting any recurring duplications of the patterns. The circles represent the individual states in the sensor output patterns, and the connecting arrows the possible connections between the states. The letter code on some of these arrows shows the resulting output (the codes are described hereinabove). Each type of sensor fault (10 in total) : i) no faulty sensors ; ii) one of the sensors is permanently"sensed" ; iii) one of the sensors is permanently"unsensed" ; and each direction (A-D and D-A) is represented by a separate diagram.

The patterns may be generated manually, and/or by computer ; generation of the patterns by one such method may be checked by comparison of the results produced by another such method.

Line by line, the contents of the data-table is converted to the representations in the state diagrams. The reduction is carried out by using as much of the existing state diagram as possible, every time a new datatable line is implemented.

Typically, when all state diagrams have been produced, the contents of these is transformed into binary form and burned into the databank EPROM 16.

In the second part of the process for producing the data on sensing signals, there is an analysis of how the sensing pattern of the four sensors 10A in a sensing zone can evolve. If, for example, sensor B is sensed by a wheel, this pattern can evolve in the following ways. The pattern is represented by a block of four digits, as described previously.

0100 = > 1100 1000 0110 0010 In other words, analysis is made to determine which transitions are possible from a given state to other states. When all possible patterns and their evolutions are analysed, these are gathered in a transitions list. This transitions list is implemented in a special software tool, which is able to generate the databank EPROM 16.

Conversion of the patterns to state diagrams is as follows, using as an example the pattern for line &num;2 in Figure 6.

1. Start with the first state in the pattern which will always be equal to the idle state (in this case"1000"). As that state already exists in the diagram, nothing is added.

2. For the second state, as in this case it is identical to the first, so nothing is added to the state diagram.

3. For the third state, one checks whether the transition from the previous state is represented in the state diagram. In this case it exists and so nothing is added.

4. For the fourth state (i. e."1110"), one checks whether the transition from the previous state is represented in the state diagram. In this case it is not already existing, and so the state diagram is extended as shown in Figure 8 (b) by the addition of state 6.

5. For the fifth and sixth states, they are identical to the fourth, so nothing is done.

6. For the seventh state (i. e."1100"), one checks whether the transition from the previous state is represented in the state diagram. In this case it is not represented, and so the state diagram is extended as shown in Figure 8 (c).

7. For the eight state (i. e."1000"), one checks whether the transition from the previous state is represented in the state diagram. In this case it is, and so nothing is to be added.

The pattern is now implemented, as a result of which one new state (state 6) and two new transitions have been added.

The basic rule is, that if a given state or transition in a given pattern already exists, nothing is added to the state diagram ; if a given state or transition does not already exist, it is added to the state diagram.

As a result, when the very first pattern is implemented, all states and transitions in the sequence are added to the then empty state diagram. Table 2 : Examples of preferred implementations of elements of traffic control system 1.

Implementation 1 Implementation 2 Traffic control 2 channel fail-safe 2 of 3 channel fail-safe system 1 computer computer Railway and monitor 2 channel fail-safe 2 of 3 channel fail-safe Unit 3A, 3B, 3C computer computer Train detector unit State-machine Computer technology 6A, 6B, 6C technology Sensing signal Programmable logic, Central Processing Analysis unit 14 Storage device and Unit discret logic Databank 16 Look up table Pattern recognition The individual elements of railway traffic control system 1 can be implemented in practice in a variety of forms according to particular circumstances of use, application and ancilliary equipment, and can incorporate various technologies, including mainframe computers, pc's, hard disks, RAM's, EPROM'S, hardware, software firmware. Table 2 indicates two particularly advantageous (but not the sole) practical implementations of principal elements of system 1.

Claims

CLAIMS 1. A method of counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the method comprising : a) defining a set of sensing characterisation values for objects to be sensed ; b) for a first characterisation value, producing a pattern of outputs from a single sensing location for a first significant position of an object sensed relative to the sensing location ;

c) producing another pattern of sensor outputs for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) repeating steps (b) and (c) for all the remaining significant relative positions of the object and the sensing location ; e) for at least a second sensing characterisation value, producing a sequence of output patterns in accordance with steps (b) to (d).
2. A method according to Claim 1, comprising producing a sequence of output patterns relating to travel in one direction only.
3. A method according to Claim 1, comprising producing a sequence of output patterns including travel in two opposite directions.
4. A method according to any of Claims 1 to 3, comprising at least one change in direction of travel in a single sensing sequence which corresponds to the completed motion of the object at the sensing location.
5. A method according to any preceding claim, comprising producing a sequence of output patterns with a noise signal occurring, in turn, at each sensor location.
6. A method according to any preceding claim, comprising producing a sequence of output patterns for a single fault for each sensor in turn, the fault being one or more of the following :

i) any one of the sensors is permanently sensed ;

ii) any one of the sensors is permanently unsensed ;

iii) transition from"no fault"to"fault"for any sensor ;

iv) transition from"fault"to"fault"for any sensor.
7. A method according to any preceding claim, comprising producing the patterns manually.
8. A method according to any of Claims 1 to 6, comprising producing the patterns by computer.
9. A method according to any of Claims 1 to 6, comprising producing the patterns by computer and checking with manually-generated signal patterns, or vice versa.
10. A method according to any preceding claim, comprising checking for duplications of patterns produced in different circumstances.
11. A method according to any preceding claim, wherein the patterns are converted to representations in a state diagram.
12. A method according to any preceding claim, comprising defining sensing characterisation values : 1. Sensing resolution of a sensor ; and/or 2. Separation of sensors ; and/or 3. Sensing the nature of an object.
13. A method according to any preceding claim, wherein a sensing location has 4 sensors.
14. A method according to any preceding claim, wherein objects to be sensed comprise wheels of railway carriages on a railway track.
15. A method of counting objects substantially as hereinbefore described with reference to, and/or as illustrated in, any one or more of the Figures in the accompanying drawings.
16. A databank produced by a method according to any one or more of Claims 1 to 15.
17. A computer program product directly loadable into the internal memory of a digital computer, comprising software code portions for performing the method of any one or more of Claims 1 to 15 when said product is run on a computer.
18. A computer program product stored on a computer usable medium, comprising : a) computer readable program means for causing a computer to define a set of sensing characterisation values for objects to be sensed ; b) computer readable program means for causing the computer to, for a first characterisation value, produce a pattern of outputs from a single sensing location for a first significant position of an object sensed relative to the sensing location ;

c) computer readable program means for causing the computer to produce another pattern of sensor outputs for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) computer readable program means for causing the computer to repeat the steps indicated as (b) and (c) hereinabove for all the remaining significant relative positions of the object and the sensing location ; and e) computer readable program means for causing the computer to, for at least a second sensing characterisation value, produce a sequence of output patterns in accordance with steps (b) to (d).
19. Electronic distribution of a database according to Claim 16, and/or of a software program according to Claim 17 or 18.
20. Apparatus for counting objects comprising producing a databank of valid signal patterns for a sensing location having at least three sensors, the apparatus comprising : a) means to define a set of sensing characterisation values for objects to be sensed ; b) means, for a first characterisation value, to produce a pattern of outputs from a single sensing location for a first significant position of an object sensed relative to the sensing location ;

c) means to produce another pattern of sensor output for the next successive significant position, in a predetermined direction, of the object relative to the sensors at the sensing location ; d) means to repeat steps (b) and (c) for all the remaining significant relative positions of the object and the sensing location ; e) means, for at least a second sensing characterisation value, to produce a sequence of output patterns in accordance with steps (b) to (d).
21. Apparatus according to Claim 20, comprising means to produce a sequence of output patterns relating to travel in one direction only.
22. Apparatus according to Claim 20 or 21, comprising means to produce a sequence of output patterns including travel in two opposite directions.
23. Apparatus according to any of Claims 20 to 22, wherein the patterns include those relating to at least one change in direction of travel in a single sensing sequence which corresponds to the completed motion of the object at the sensing location.
24. Apparatus according to any of Claims 20 to 23, wherein the patterns include those relating to a noise signal occurring, in turn, at each sensor location.
25. Apparatus according to any of Claims 20 to 24, wherein the patterns include those relating to a single fault for each sensor in turn, the fault being one or more of the following :

i) any one of the sensors is permanently sensed ;

ii) any one of the sensors if permanently unsensed ;
26. Apparatus according to any of Claims 22 to 25 wherein the means to check computer-generated results with manually-generated signal patterns, or vice versa.
27. Apparatus according to any of Claims 22 to 26, comprising means to check for duplications of patterns produced in different circumstances.
28. Apparatus according to any of Claims 22 to 27, comprising means to convert the patterns to representations in a state diagram.
29. Apparatus according to any of Claims 22 to 28, comprising means to define sensing characterisation values for :

i) sensing resolution of a sensor ; and/or

ii) separation of sensors and/or

iii) sensing the nature of an object.
30. Apparatus according to any of Claims 22 to 29, wherein a sensing location has 4 sensors.
31. Apparatus according to any one of Claims 22 to 30, wherein objects to be sensed comprise wheels of railway carriages on a railway track.
32. Apparatus for counting objects substantially as hereinbefore described with reference to, and/or as illustrated in, any one or more of the Figures in the accompanying drawings.