GB2466950A

GB2466950A - Road traffic congestion detection system

Info

Publication number: GB2466950A
Application number: GB0900552A
Authority: GB
Inventors: Michael Alexander Graham Clark; Henry John Gibbs
Original assignee: CLARK SYSTEMS Ltd
Current assignee: CLARK SYSTEMS Ltd
Priority date: 2009-01-14
Filing date: 2009-01-14
Publication date: 2010-07-21
Also published as: GB0900552D0; GB2466950A8

Abstract

A road traffic congestion detection system (30) comprises a detector for detecting the presence or lack of presence of at least one vehicle in a detection zone (36) to produce a presence signal; a detector for detecting, when at least one vehicle is in the detection zone, an instantaneous speed of the vehicle or of one of the vehicles in the detection zone to produce a speed signal; and a processor (10) for producing a congestion signal (20) dependent on the presence signal and the speed signal. The detecting means produces the presence signal as a series of pulses dependent on the time the vehicle was present in the detection zone. The congestion signal is also produced as a series of pulses, the pulse width dependent on whether the speed signal satisfies predetermined criteria. The congestion signal may be compatible with the HIOCC algorithm and trigger alerts in a MIDAS system. The detector may use electromagnetic radiation, such as RADAR, and cover two lanes of the carriageway (14, 22) with means to detect only vehicle travelling in one direction.

Description

TITLE

Road traffic congestion detection systems

DESCRIPTION

This invention relates to systems for detecting road traffic congestion.

Congestion on the roads can be caused, for example, by an unduly high volume of traffic, bad weather and/or accidents. Automatic systems for detecting road traffic congestion and signalling alerts to a central station are known. Once an alert is received, any necessary action can be taken, such as deploying personnel to investigate the matter further, deploying emergency services, activating warning signs and invoking diversions and temporary speed limits.

The Highways Agency in the United Kingdom operates a Highways Agency traffic management system (or HATMS) which produces motorway incident detection and signalling alerts (or MIDAS alerts). The traffic flow at a particular location is detected by an inductive loop, at least one per lane, buried beneath the road surface, which supplies signals to a roadside signal processing unit, which can produce MIDAS alerts which are conveyed to a central station by a network of the HATMS. The roadside unit processes the signals using a high-occupancy (or HIOCC) algorithm, or a more-advanced HIOCC2 algorithm.

Referring to Figure 1 of the drawings, in a typical implementation of the HIOCC algorithm by a roadside unit 10, a single inductive loop 12 is buried in a particular lane 14 of a carriageway 16 of a road to define a detection zone, and the loop 12 is connected via a relay unit 18 to the roadside unit 10. The HIOCC algorithm divides time into main intervals of one second each, and each main interval is divided into ten sub-intervals. An instantaneous occupancy value is determined for each main interval by counting the number of sub-intervals in that main interval in which a vehicle is detected by the inductive loop 12. The instantaneous occupancy value for a particular main interval can therefore range between zero and ten. If the instantaneous occupancy values for two consecutive main intervals are both ten, then a MIDAS alert is generated and transmitted on the HATMS network 20. The roadside unit 10 also calculates smoothed values of the instantaneous occupancy values using single-stage exponential smoothing, and maintains the MIDAS alert until the current smoothed occupancy value falls below an average of particular smoothed occupancy values that were calculated before the alert was generated. In the case of a carriageway 16 having a further lane 22, a further inductive loop 24 is buried in the further lane 22 of the carriageway 16 and connected via a relay unit 18 to the roadside unit 10. The roadside unit 10 processes the signals from the two inductive loops 12,24 independently in order to determine whether to generate a MIDAS alert.

There are a number of problems or disadvantages associated with the arrangement described above: A. First, it is necessary to install an inductive loop 12 in the lane 14 of the carriageway 16 that is to be monitored, and in order to do that it is necessary to close the lane 14, the carriageway 16 or the whole road to traffic.

B. Second, it is necessary to install a separate inductive loop 12,24 in each lane 14,22 that is to be monitored and to process the data from the inductive loops 12,24 separately.

C. Third, generation of an alert is dependent on when, in relation to the main interval time frame, a vehicle is first detected by the inductive loop 12. For example, referring to row Ri in Figure 2, if a vehicle is initially detected in the first sub-interval of a main interval, it needs to remain detected for the next nineteen sub-intervals in order to trigger an alert, requiring a continuous occupancy of about two seconds. However, referring to row R2, if a vehicle is initially detected in the second sub-interval of a main interval and remains detected for the next nineteen sub-intervals, it does not trigger an alert. Instead, it would need to remain detected for a further nine sub-intervals before it would trigger an alert, requiring a continuous occupancy of about 2.9 seconds.

Therefore, at speeds above an upper threshold, a particular vehicle will not trigger an alert; at speeds below a lower threshold, the vehicle will trigger an alert; and in an intermediate range of speeds between the two thresholds, the vehicle may or may not trigger an alert depending on when, in relation to the main interval time frame, it is first detected.

D. Fourth, the generation of an alert is dependent not only on the speed of the vehicle but also on its length, in that a long vehicle travelling at a particular speed will be detected by the inductive loop 12 for a longer period of time than a short vehicle travelling at the same speed. Tests on a typical arrangement have shown that a small car, three metres in length, will trigger a MIDAS alert at speeds up to 6 km/h, and may possibly trigger an alert at speeds between 6 km/h and 9 km/h, whereas a long truck or coach, seventeen metres in length will trigger an alert at speeds up to 24 km/h, and may possibly trigger an alert at speeds between 24 km/h and 34 km/h. Accordingly some vehicles may trigger a MIDAS alert when travelling as fast as 34 km/h, whereas others may not trigger an alert when travelling as slowly as 6 km/h.

The HIOCC2 algorithm attempts to deal with problems "C" and "D" mentioned above.

Referring to row R4 of Figure 3, in the case where a vehicle is initially detected in a sub-interval other than the first sub-interval of a main interval and remains detected for the remainder of that main interval, those detections are delayed until the vehicle ceases to be detected, as shown in row R4' of Figure 3. This aspect of the HIOCC2 algorithm therefore attempts to align the arrival of a vehicle with the start of a main interval. Also, the HIOCC2 algorithm requires, as shown in Figure 4, a second inductive loop 26,28 to be buried in each lane 14,22 of the carriageway 16 with a predetermined pitch P of, for example, 4.5 metres from the first inductive loop 12,24 to define a second detection zone in the same lane. The second loop 26,28 supplies signals to the roadside unit 10 via a further relay unit 18. The roadside unit attempts to correlate the signals from the pair of inductive loops 12,26; 24,28 in each lane 14,22 in order to estimate the speed of the vehicle in that lane 14,22. Then, if the estimated speed is greater than a threshold, a MIDAS alert is not triggered despite the occupancy data.

Tests have shown that, with an example of the HIOCC2 algorithm, the range of speeds at which a vehicle may or may not trigger an alert, depending on the vehicle's length, is reduced to between about 19 km/h and 24 km/h. Although the HIOCC2 algorithm ameliorates problems "C" and "D" of the HIOCC algorithm, it does however exacerbate disadvantage "A" above, in that double the number of inductive loops are required.

An aim of the present invention, or at least of specific embodiments of it, is to provide a system for detecting road traffic congestion that does not require two detection zones in the same lane in order to deal with the problem of long vehicles, that does not necessarily require inductive loops to be buried in the carriageway, that does not require duplication of detection systems for carriageways having more than one lane, and that is compatible with existing HIOCC and HIOCC2 or similar processing equipment.

In accordance with a first aspect of the present invention, there is provided a road traffic congestion detection system, comprising: presence detecting means for detecting the presence or lack of presence of at least one vehicle in a detection zone to produce a presence signal; speed detecting means for detecting, when at least one vehicle is in the detection zone, an instantaneous speed of the vehicle or of one of the vehicles in the detection zone to produce a speed signal; and processing means for producing a congestion signal dependent on the presence signal and the speed signal. By comparison with the prior art system employing the HIOCC2 algorithm, the system of the present invention does not require two detection zones in the same lane in order to calculate an estimate of the average speed of the vehicle from the time taken to travel between two fixed points. Instead, instantaneous vehicle speed is detected. (Although "instantaneous speed" is referred to in this specification, this is not intended to exclude the case where the instantaneous speed is smoothed over a brief interval.) Although other types of equipment may be employed, the presence detecting means and speed detecting means preferably comprise means for projecting electromagnetic radiation above ground into the detection zone, means for detecting electromagnetic radiation reflected by vehicles in the detection zone, and means for producing the presence signal and the speed signal from the detected radiation. Indeed, a known microwave radar vehicle detector may be used to produce the presence signal and speed signal. This has the advantage, unlike inductive loops, that is unnecessary to close at least part of the carriageway in order to install the system.

In the case where the road carriageway has more than one lane, the detection zone is preferably arranged to cover at least two lanes of the carriageway in order to avoid duplication of equipment. Although the system does not then produce separate congestion signals for each lane, it will nevertheless detect congestion in all of the lanes that are covered.

The presence detecting means and the speed detecting means are preferably arranged to produce the presence signal and the speed signal only for vehicles travelling in one particular direction along the carriageway. Unwanted signals will therefore not be produced in the event that the system is set up so that the detection zone covers or partly covers a lane in which traffic travels in the opposite direction.

In one embodiment of the invention, the detecting means is operable to produce the presence signal as a series of pulses each having a pulse width dependent on (for example substantially equal to) the period of time for which at least one vehicle is present in the detection zone; the processing means is operable to produce the congestion signal as a series of pulses; and the processing means is operable to determine whether the speed signal satisfies a first criterion and if so to produce a pulse in the congestion signal in a first predetermined manner.

The first criterion may be that the speed signal represents a speed less than a first predetermined speed, such as 20 km/h. More specifically, the first criterion is preferably that the speed signal, substantially at the time when the detection signal is indicative that a vehicle is entering the detection zone, represents a speed less than the first predetermined speed.

If the speed signal satisfies the first criterion, the processor may be operable to produce such a pulse in the congestion signal having a first predetermined pulse width, for example not less than the longer threshold pulse width (typically 2.9 seconds) of the HIOCC algorithm.

Accordingly, vehicles travelling at less than, for example, 20 km/h will be guaranteed to trigger a MIDAS alert. Alternatively, if the speed signal satisfies the first criterion, the processor may be operable to produce such a pulse in the congestion signal having a pulse width dependent on (for example substantially equal to) the pulse width of a respective pulse of the presence signal.

If the speed signal does not satisfy the first criterion, the processing means may be operable to produce a pulse in the congestion signal in a second predetermined manner, for example with a second predetermined pulse width, which may, for example be not greater than the shorter threshold pulse width (typically 2.0 seconds) of the HIOCC algorithm. Accordingly, vehicles travelling at more than, for example, 20 km/h will be guaranteed not to trigger a MIDAS alert. Furthermore, if for example two long vehicles, 17 metres in length each, are travelling at similar speeds in different lanes of a carriageway covered by the same detection zone, and if the vehicles are just overlapping in the field of view of the detector system so that they are detected together as a single vehicle, they will not trigger a MIDAS alert provided that they are travelling in excess of the first predetermined speed (e.g. 20 km/h) despite the long duration of the presence signal (about 6 seconds at 20 km/h).

With the above features, it will be appreciated that the congestion signal can be input into the basic HIOCC algorithm and yet provide the advantage of the HIOCC2 algorithm in preventing false alerts with fast moving, but long, vehicles.

It is possible that a vehicle may enter the detection zone at a sufficiently high speed to produce a pulse in the congestion signal with the second predetermined pulse width, but that the vehicle may be braking sharply, so that it would be preferable for the HIOCC algorithm to determine whether or not to produce a MIDAS alert in the normal way. In order to deal with this, if the speed signal does not satisfy the first criterion, the processing means is preferably operable to produce a pulse in the congestion signal having the second predetermined pulse width, unless the speed signal at a time during the production of the pulse in the congestion signal satisfies a second criterion. The second criterion may be that the speed signal, at a time during the production of the pulse in the congestion signal, represents a speed less than a second predetermined speed, the second predetermined speed being less than the first predetermined speed. If the speed signal does not satisfy the first criterion but does satisfy the second criterion, the processing means is then preferably operable to produce a pulse in the congestion signal having a pulse width dependent on (for example substantially equal to) the pulse width of the respective pulse of the presence signal.

If the system is to be used with the HIOCC2 algorithm, for example to take advantage of the alignment of the arrival of a vehicle with the start of a main interval of the HIOCC algorithm, the processing means is preferably operable to produce a second congestion signal as a series of pulses, each pulse in the second congestion signal being generated from a respective pulse in the first-mentioned congestion signal delayed by a time substantially inversely proportional to the speed represented by the speed signal at a time during production of the respective pulse in the first congestion signal. The first and second congestion signals can then be processed in accordance with the HIOCC2 algorithm.

The detection system of the first aspect of the invention may be used with algorithms other than the HIOCC or HIOCC2 algorithm. In this case, the processing means may be operable to produce the congestion signal in accordance with a predetermined algorithm dependent on how many vehicles having occupied the detection zone travelling at a speed satisfying a third criterion and dependent on how many vehicles having occupied the detection zone travelling at a speed satisfying a fourth criterion. For example, the algorithm may be such that the congestion signal is changed to indicate congestion once a vehicle has, or a predetermined number of successive vehicles have, occupied the detection zone travelling at a speed less than a third predetermined speed, and the congestion signal is subsequently changed to indicate a lack of congestion once a vehicle has, or a predetermined number of successive vehicles have, occupied the detection zone travelling at a speed greater than a fourth predetermined speed, the fourth predetermined speed being greater than the third predetermined speed.

In accordance with a second aspect of the invention, there is provided a road traffic congestion detection method, comprising the steps of: detecting the presence or lack of presence of at least one vehicle in a detection zone; detecting, when at least one vehicle is in the detection zone, an instantaneous speed of the vehicle or of one of the vehicles in the detection zone; and producing a congestion signal dependent on the detected presence and detected speed.

Specific embodiments of the present invention will now be described, purely by way of example, with reference to the accompanying drawings, in which: Figure 1 is a bird's eye view of a section of carriageway having a known congestion monitoring system; Figure 2 is a timing diagram illustrating operation of a known HIOCC algorithm used in the system of Figure 1; Figure 3 is a timing diagram illustrating how the known HIOCC2 algorithm modifies the timing diagram of Figure 2; Figure 4 is a bird's eye view of a section of carriageway having a known congestion monitoring system utilising the HIOCC2 algorithm; Figure 5 is a schematic block diagram of a known radar vehicle detector; Figure 6 is a bird's eye view of a section of carriageway having a congestion monitoring system employing the detector of Figure 5; Figure 7 is an aerial view of the section of carriageway of Figure 6 illustrating a problem with overlapping vehicles; Figure 8 is a schematic block diagram showing a modification to the detector of Figure 5 in accordance with a first embodiment of the invention; Figure 9 is a state diagram illustrating operation of the detector of Figure 8; Figure 10 is a schematic block diagram showing another modification to the detector of Figure 5 in accordance with a second embodiment of the invention; Figure 11 is a flow diagram illustrating operation of the detector of Figure 10; Figure 12 is a schematic block diagram showing a further modification to the detector of Figure 5 in accordance with a third embodiment of the invention; and Figure 13 is a state diagram illustrating operation of the detector of Figure 12.

Figures 5 and 6 show a known radar vehicle detector 30 that is mounted, above ground, on a pole 32, lamppost or signpost at the roadside and detects vehicles as they pass the detector 30. The detector 30 incorporates a K-band microwave patch antenna 34 which has a well-defined beam pattern, a low-noise microwave receiver 38, and a digital signal processor 40 which processes the received signals and produces two output signals, PRESENCE and SPEED.

The detector 30 is mounted so that the axis of the beam pattern 36, when viewed vertically from above, is inclined at an angle A of about thirty degrees towards the direction B of the oncoming traffic, and so that the beam pattern 36 covers both of the lanes 14,22 of the carriageway 16.

The SPEED signal indicates the instantaneous speed of any vehicle detected by the detector 30 in the detection zone travelling in the direction B, and, in the case where more than one vehicle is in the detection zone, the speed of the vehicle which produces the strongest signal received by receiver 38. The PRESENCE signal is TRUE or FALSE depending on whether or not any vehicles are within the detection zone 36 travelling in the direction B. In the event that a vehicle is detected by the detector 30, but its speed falls to zero while it is being detected, the PRESENCE signal automatically resets to FALSE after a preset time of, for example, thirty seconds so that the detector 30 does not hang up in the event that the vehicle has slowly moved out of the detection zone 36 without being noticed. The detector 30 does not provide any output in response to any vehicles which may be travelling in the opposite direction to the direction B. Detectors of the type described in this paragraph are known and available on the market, an example being the RVD 1501 radar vehicle detector marketed by Clark Systems Limited, of Southampton, S030 2GX, GB. The PRESENCE signal produced by the detector 30 described above could be directly input to a roadside unit 10 of the HATMS utilising the basic HIOCC algorithm. However, it could then produce false MIDAS alerts due to the detection zone 36 covering more than one lane 14,22 of the carriageway 16. For example, Figure 7 shows the case where two long vehicles 42A,42B each having a length of, for example, 17 metres are travelling along the lanes 14,22, respectively, at generally similar speeds of, for example, 42 km/h, which is well in excess of the upper and lower speed thresholds of 34 and 24 km/h for long vehicles using HIOCC and an inductive loop. If the two vehicles 42A,42B were just overlapping as they passed through the detection zone 36, they would generate a continuous PRESENCE signal for at least 2.9 seconds, which would guarantee that the roadside unit 10 would generate a MIDAS alert. In order to avoid this problem of false detection, in accordance with the embodiment of the invention, the SPEED and PRESENCE signals are processed further, as will now be described with reference to Figures 8 and 9.

Referring to Figure 8, a further digital signal processor 44 is included, or the existing digital signal processor 40 is modified, so as to produce an OCCUPANCY-A signal. It is the OCCUPANCY-A signal, rather than the PRESENCE signal, which is output to the roadside unit 10 utilising the HIOCC algorithm. The processor 44 is provided with three preset values: PERIOD is a period of an OCCUPANCY-A signal insufficient to trigger a MIDAS alert.

Typically PERIOD would be set to 1 second.

SPEED-A is a vehicle speed below which the period of the OCCUPANCY-A signal will be directly determined from the period of the PRESENCE signal. Typically SPEED-A SPEED-B is a speed indicative that a vehicle that was travelling at a speed in excess of SPEED-A is about to stop. Typically SPEED-B is set to 2 km/h.

The operation of the further processor 44 is illustrated by the state diagram of Figure 9.

The OCCUPANCY-A signal is initially set in step 46 to FALSE, and the processor 44 then waits in state 48 for the PRESENCE signal to become TRUE. In response to that, in step/state 50, the OCCUPANCY-A signal is set to TRUE, and the processor 44 then immediately determines whether the SPEED signal is less than the preset SPEED-A value. If so, then in state 52, the processor 44 waits for the PRESENCE signal to become FALSE, whereupon in step 46 the OCCUPANCY-A signal is set to FALSE, and the processor 44 reverts to waiting state 48.

Accordingly, if a vehicle enters the detection zone 36 at a speed less than 20 km/h, an OCCUPANCY-A pulse is output to the roadside unit 10 for as long as a vehicle continues to be detected by the detector 30. In response, the roadside unit 10 may or may not generate a MIDAS alert depending on the period of the OCCUPANCY-A pulse.

However, if, in step/state 50, the SPEED signal is determined to be not less than the preset SPEED-A value, then in step 54 a TIMER signal is started, and then in state 56 the processor 44 waits for the TIMER signal to reach the preset PERIOD value. When it does, the OCCUPANCY-A signal is set to FALSE in step 58, and the processor 44 waits in state 60 for the PRESENCE signal to become FALSE, whereupon processor 44 reverts to waiting state 48.

Accordingly, if a vehicle enters the detection zone 36 at a speed not less than 20 km/h, an OCCUPANCY-A pulse is output to the roadside unit 10 for 1 second, which is insufficient for the roadside unit 10 to trigger a MIDAS alert, and a further OCCUPANCY-A pulse is prevented, by waiting state 60, until the vehicle has left the detection zone 36, or, in the case of more than one overlapping vehicle, all of the vehicles have left the detection zone 36. The feature therefore prevent a false MIDAS alert in the circumstances described with reference to Figure 7.

A problem that may arise if a vehicle enters the detection zone 36 at a speed not less than the preset SPEED-A value, but is braking sharply, so that it ought possibly to generate a MIDAS alert. In order to deal with this, while in state 56 waiting for the TIMER signal to reach the preset PERIOD value, the processor 44 also monitors the SPEED signal, and if it falls to the preset SPEED-B value, the state changes to waiting state 52, rather than proceeding to step 58, -10 -whereupon the processor 44 waits for the PRESENCE signal to become FALSE before setting the OCCUPANCY-A signal to FALSE in step 46. The processor therefore waits either for the vehicle to move out of the detection zone 36, or for the PRESENCE signal to reset automatically after 30 seconds (as described above in connection with Figure 5).

It will be appreciated that, with the algorithm of Figure 9, if a vehicle enters the detection zone 30 travelling at a speed slower than the threshold speed SPEED-A (e.g. 20 km/h), the width of the OCCUPANCY-A pulse is substantially equal to the width of the PRESENCE pulse. In a modification of the algorithm of Figure 9, if a vehicle enters the detection zone 30 travelling at a speed slower than the threshold speed SPEED-A, an OCCUPANCY-A pulse produced with a predetermined pulse width PERIOD-B which is guaranteed to cause the roadside unit 10 to trigger a MIDAS alert. For example, PERIOD-B may be set to 3.4 seconds.

It will be appreciated from the above that the single detector 30 described with reference to Figures 8 and 9 (i) is compatible with a roadside unit 10 utilising the HIOCC algorithm, but (ii) does not require any inductive loops 12,24 to be buried in the lanes 14,22 of the carriageway and (iii) does not require separate detection zones for different lanes 14,22 and yet (iv) deals with the problems of overlapping vehicles and vehicles braking sharply in the detection zone 36.

In the case where the detector 30 is required to be compatible with a roadside unit 10 utilising the HIOCC2 algorithm, it may be modified, using the SPEED signal, to produce a second output emulating the outputs from the further inductive loops 26,28 described above with reference to Figure 4. It will be recalled that the HIOCC2 algorithm attempts to align the arrival of a vehicle with the start of its main interval. It can perform this function on the OCCUPANCY-A signal without any modification. The HIOCC2 algorithm also attempts to suppress MIDAS alerts triggered by fast-moving long vehicles using a signal on its second input which will now be referred to as the OCCUPANCY-B signal. The modification that will now be described with reference to Figures 10 and 11 emulates the OCCUPANCY-B signal produced by the further inductive loops 26,28 for fast moving vehicles by producing a delayed form of the OCCUPANCY-A signal with a variable delay inversely proportional to the SPEED signal.

The modified detector 30 of Figure 10 produces its OCCUPANCY-A signal in an identical manner to that described with reference to Figures 8 and 9. The further digital signal processor 44 (or modified existing digital signal processor 40) also parallel processes in accordance with the flow diagram of Figure 11. The processor 44 maintains a rotating circular -11 -buffer which can be considered to be addressed in a clockwise direction and to be rotatable as a whole one address at a time in an anti-clockwise direction. The buffer may have, for example, 64 bits of memory, one for each 1/10 second interval in a 6.4 second period. The processor 44 also provides a clock producing clock pulses at 1/10 second intervals. Upon initialisation in steps 62 and 64, the clock is started and the buffer is completely cleared. In step 66, a DELAY is calculated from the current SPEED signal (in km/h) and the pitch P (in metres) of the inductive loops 12,26;24;28 that the detector is emulating by the formula DELAY = 36 x P / SPEED (the factor of 36 being derived from the number of tenths of a second taken to travel a distance of 1 m at a speed of 1 km/h). However, if DELAY calculates to more than 63, it is set to 63. In step 68, the buffer at the address DELAY is written with the value of the current OCCUPANCY-A signal. In step 70, the value currently stored in the buffer at current address zero is read and set as the OCCUPANCY-B signal which is output to the roadside unit 10. In step 72, the value at address zero is cleared. In step 74, the buffer is rotated one address anticlockwise so that address 1 becomes address 0, address 2 becomes address 1, and so on, and address 0 becomes address 63. In step 76, the processor 44 waits for the next clock pulse, whereon the process returns to step 66. It will therefore be appreciated that the OCCUPANCY-B signal is produced by delaying the OCCUPANCY-A signal with a variable delay inversely proportional to the SPEED signal and that the value of the OCCUPANCY-2 signal is updated every tenth of a second.

In a modification of the process of Figure 11, the processor 44 does not employ a rotating buffer to produce the OCCUPANCY-B signal. Instead, it uses two timers, START-TIMER and END-TIMER. When the OCCUPANCY-A signal is changed to TRUE, for example in step 50 of Figure 9, the processor 44 calculates a DELAY value as described above.

Also, the processor 44 immediately starts the START-TIMER and subsequently sets the OCCUPANCY-B signal to TRUE when the START-TIMER reaches the DELAY value.

Furthermore, when the OCCUPANCY-A signal is changed to FALSE, for example in step 46 or 58 of Figure 9, the processor 44 immediately starts END-TIMER and subsequently sets the OCCUPANCY-B signal to FALSE when the END-TIMER reaches the DELAY value. The OCCUPANCY-B signal is therefore produced by delaying the OCCUPANCY-A signal with a variable delay inversely proportional to the SPEED signal.

The detector 30 may also be modified for use in congestion monitoring systems that do not employ the HIOCC or HIOCC2 algorithms. For example, Figure 12 and 13 illustrate a system in which an alert is raised (i.e. the signal ALERT becomes TRUE) when one vehicle -12 -has, or a plurality of successive vehicles have, entered the detection zone 36 below a lower threshold speed, and the alert is cleared when one vehicle has, or a plurality of successive vehicles have, entered the detection zone 36 above a higher threshold speed. Specifically, the further digital signal processor 44 (or modified existing digital signal processor 40) maintains two counters having values ALERT-COUNT and CLEAR-COUNT and is provided with four preset values: ALERT-SPEED is the speed below which a vehicle counts as contributing to an alert.

Typically, ALERT-SPEED may be set to 15 km/h.

CLEAR-SPEED is the speed above which a vehicle counts as contributing to the clearing of an alert. Typically, CLEAR-SPEED may be set to 30 km/h.

ALERT-THRESH is the number of vehicles required to generate an alert. Typically, ALERT-THRESH may be set to one.

CLEAR-THRESH is the number of vehicles required to clear an alert. Typically, CLEAR-THRESH may be set to five.

Referring to Figure 13, in initialisation step 80, ALERT-COUNT and CLEAR-COUNT are cleared and the ALERT signal is set to FALSE, and then the processor 44 waits, if necessary, in state 82 for the PRESENCE signal to become FALSE, whereupon it waits in state 84 for the PRESENCE signal to become TRUE. When that happens, the processor 44 determines in step 86 whether or not the SPEED signal is less than the ALERT-SPEED value.

If it is not, then the processor 44 returns to waiting state 82, but if it is then in step 88 the processor 44 increments the ALERT-COUNT counter and determines whether or not ALERT-COUNT has reached the ALERT-THRESH threshold. If it has not, then the processor 44 waits in state 90 for the vehicle to cease being detected, and then waits in state 92 for the next vehicle to be detected. When that happens, the processor 44 determines in step 94 whether or not the SPEED signal is less than the ALERT-SPEED value. If it is not, indicative that there have not been two successive vehicles travelling below the ALERT-SPEED speed, then in step 96 the ALERT-COUNT counter is reset, and the processor 44 returns to waiting state 82. However, if in step 94 the processor 44 determines that the SPEED signal is less than the ALERT-SPEED value, the processor 44 returns to step 88.

If in step 88 the processor determines that ALERT-COUNT has reached the ALERT-THRESH threshold, then in step 98 the processor 44 sets the ALARM signal to TRUE and then waits in state 100 for the PRESENCE signal to become FALSE, whereupon it waits in state 102 for the PRESENCE signal to become TRUE. When that happens, the processor 44 determines in step 104 whether or not the SPEED signal is greater than the CLEAR-SPEED value. If it is -13 -not, then the processor 44 returns to waiting state 100, but if it is then in step 106 the processor 44 increments the CLEAR-COUNT counter and determines whether or not CLEAR-COUNT has reached the CLEAR-THRESH threshold. If it has not, then the processor 44 waits in state 108 for the vehicle to cease being detected, and then waits in state 110 for the next vehicle to be detected. When that happens, the processor 44 determines in step 112 whether or not the SPEED signal is greater than the CLEAR-SPEED value. If it is not, indicative that there have not been two successive vehicles travelling above the CLEAR-SPEED speed, then in step 114 the CLEAR-COUNT counter is reset, and the processor 44 returns to waiting state 100.

If in step 104 the processor 44 determines that the CLEAR-COUNT has reached the CLEAR-THRESH threshold, then the processor 44 returns to the initialisation step 80 in which both counters ALERT-COUNT and CLEAR-COUNT are reset and the ALERT signal is set to FALSE.

It will therefore be appreciated that the ALERT signal is set to TRUE when ALERT- THRESH successive vehicles have entered the detection zone 36 at speeds below ALERT-SPEED, and that the ALERT signal is subsequently reset when CLEAR-THRESH successive vehicles have entered the detection zone 36 at speeds above CLEAR-SPEED.

In a modification to the algorithm described with reference to Figures 12 and 13, CLEAR-THRESH is not a fixed value, but is varied according to historical vehicle flow over a period of typically five minutes prior to the alert being raised. For example, for a low flow rate as might occur during the middle of the night, CLEAR-THRESH might become a number as low as one, whereas for high vehicle flow rates as might occur during peak hours, CLEAR-THRESH might be increased to a higher number such as twenty.

It should be noted that the embodiments of the invention has been described above purely by way of example and that many modifications and developments may be made thereto within the scope of the present invention.

Claims

-14 -CLAIMS1. A road traffic congestion detection system, comprising: presence detecting means for detecting the presence or lack of presence of at least one vehicle in a detection zone to produce a presence signal; speed detecting means for detecting, when at least one vehicle is in the detection zone, an instantaneous speed of the vehicle or of one of the vehicles in the detection zone to produce a speed signal; and processing means for producing a congestion signal dependent on the presence signal and the speed signal.
2. A system as claimed in claim 1, wherein: the detecting means is operable to produce the presence signal as a series of pulses each having a pulse width dependent on the period of time for which at least one vehicle is present in the detection zone; the processing means is operable to produce the congestion signal as a series of pulses; and the processing means is operable to determine whether the speed signal satisfies a first criterion and if so to produce a pulse in the congestion signal in a first predetermined manner.
3. A system as claimed in claim 2, further including means for processing the congestion signal in accordance with a HIOCC algorithm having a shorter threshold pulse width for which such a pulse of the congestion signal may or may not trigger an alert and a longer threshold pulse width for which such a pulse of the congestion signal will trigger an alert.
4. A system as claimed in claim 2 or 3, wherein the first criterion is that the speed signal represents a speed less than a first predetermined speed.
5. A system as claimed in claim 2 or 3, wherein the first criterion is that the speed signal, substantially at the time when the detection signal is indicative that a vehicle is entering the detection zone, represents a speed less than a first predetermined speed.
6. A system as claimed in any of claims 2 to 5, wherein, if the speed signal satisfies the first criterion, the processor is operable to produce such a pulse in the congestion signal having a first predetermined pulse width.

-15 -
7. A system as claimed in claim 6 when directly or indirectly dependent on claim 3, wherein the first predetermined pulse width is not less than the longer threshold pulse width of the HIOCC algorithm.
8. A system as claimed in any of claims 2 to 5, wherein, if the speed signal satisfies the first criterion, the processor is operable to produce such a pulse in the congestion signal having a pulse width dependent on the pulse width of a respective pulse of the presence signal.
9. A system as claimed in claim 8, wherein, if the speed signal satisfies the first criterion, the pulse width of the congestion signal is substantially equal to the pulse width of the respective pulse of the presence signal.
10. A system as claimed in any of claims 2 to 9, wherein if the speed signal does not satisfy the first criterion, the processing means is operable to produce a pulse in the congestion signal in a second predetermined maimer.
11. A system as claimed in claim 10, wherein, if the speed signal does not satisfy the first criterion, the processor is operable to produce such a pulse in the congestion signal having a second predetermined pulse width.
12. A system as claimed in claim 10, wherein, if the speed signal does not satisfy the first criterion, the processing means is operable to produce such a pulse in the congestion signal having a second predetermined pulse width, unless the speed signal at a time during the production of the pulse in the congestion signal satisfies a second criterion.
13. A system as claimed in claim 12, wherein the second criterion is that the speed signal, at a time during the production of the pulse in the congestion signal, represents a speed less than a second predetermined speed, the second predetermined speed being less than the first predetermined speed.
14. A system as claimed in claim 12 or 13, wherein if the speed signal does not satisfy the first criterion but does satisfy the second criterion, the processing means is operable to produce a pulse in the congestion signal having a pulse width dependent on the pulse width of a respective pulse of the presence signal.

-16 -
15. A system as claimed in any of claims 11 to 14 when directly or indirectly dependent on claim 3, wherein the second predetermined pulse width is not greater than the shorter threshold pulse width of the HIOCC algorithm.
16. A system as claimed in any of claims 2 to 15, wherein the processing means is operable to produce a second congestion signal as a series of pulses, each pulse in the second congestion signal being generated from a respective pulse in the first-mentioned congestion signal delayed by a time substantially inversely proportional to the speed represented by the speed signal at a time during production of the respective pulse in the first congestion signal.
17. A system as claimed in claim 16, further including means for processing the first and second congestion signals in accordance with the HIOCC2 algorithm.
18. A system as claimed in claim 1, wherein the processing means is operable to produce the congestion signal in accordance with a predetermined algorithm dependent on how many vehicles having occupied the detection zone travelling at a speed satisfying a third criterion and dependent on how many vehicles having occupied the detection zone travelling at a speed satisfying a fourth criterion.
19. A system as claimed in claim 18, wherein the algorithm is such that the congestion signal is changed to indicate congestion once a vehicle has, or a predetermined number of successive vehicles have, occupied the detection zone travelling at a speed less than a third predetermined speed, and the congestion signal is subsequently changed to indicate a lack of congestion once a vehicle has, or a predetermined number of successive vehicles have, occupied the detection zone travelling at a speed greater than a fourth predetermined speed, the fourth predetermined speed being greater than the third predetermined speed.
20. A system as claimed in any preceding claim, wherein the presence detecting means and speed detecting means comprise means for projecting electromagnetic radiation into the detection zone, means for detecting electromagnetic radiation reflected by vehicles in the detection zone, and means for producing the presence signal and the speed signal from the detected radiation.

-17 -
21. A system as claimed in any preceding claim, wherein the detection zone is arranged to cover at least two lanes of a road carriageway.
22. A system as claimed in claim 21, wherein the presence detecting means and the speed detecting means are arranged to produce the presence signal and the speed signal only for vehicles travelling in one particular direction along the carriageway.
23. A road traffic congestion detection system, substantially as described with reference to Figures 8 to 13 of the drawings.
24. A road traffic congestion detection method, comprising the steps of: detecting the presence or lack of presence of at least one vehicle in a detection zone; detecting, when at least one vehicle is in the detection zone, an instantaneous speed of the vehicle or of one of the vehicles in the detection zone; and producing a congestion signal dependent on the detected presence and detected speed.
25. A road traffic congestion detection method, substantially as described with reference to Figures 8 to 13 of the drawings.