WO2004027729A1

WO2004027729A1 - A method and system for detecting and estimating road traffic from location data of mobile terminals in a radiocommunication system, so as a program therefor

Info

Publication number: WO2004027729A1
Application number: PCT/IT2003/000552
Authority: WO
Inventors: Maurizio Morisio; Daniele Romano; Antonio Fanigliulo; Marco Vallesio
Original assignee: Politecnico Di Torino
Priority date: 2002-09-19
Filing date: 2003-09-18
Publication date: 2004-04-01
Also published as: ITTO20020814A1; AU2003274690A8; AU2003274690A1; WO2004027729A8

Abstract

The invention relates to a method and a system for detection and estimation of road traffic on the basis of position-locating information on mobile terminals of a radio communication system. The method determines if a mobile terminal under consideration is in motion or is stationary, then if it is located inside or outside of the road section under surveillance. From the knowledge of the position and speed of displacement of each mobile it is evaluated if the road traffic is flowing or slowing, and -if at least one terminal belonging to the road section is stationary, it checks if a traffic jam situation exists and if the stationary mobile is involved in it.

Description

A method and system for detecting and estimating road traffic from location data of mobile terminals in a radiocommunication system, so as a program therefor

The present invention relates to a method and a system for detection and estimation of road traffic on the basis of position information of mobile terminals of a radio communication system.

The integration of position, navigation and personal communication (mobile telephone) services represents one of the fundamental necessities of a wide range of important civil and commercial sectors . Aerial and maritime navigation, control of railway traffic, management of fleets of vehicles, the provision of comprehensive and timely- information relating to the traffic and the road conditions, emergency services and search and rescue are only some examples of the numerous fields of application which would be able to draw on the advantage of the integrated provision of these services.

The availability of a communications infrastructure makes it possible to offer the users of a generic navigation system a value in addition to just the position-locating service; as well as the transmission of data relating to position, the communication instruments allow the user to receive both information of a location-aware type (road maps, alternative itineraries, sites or services of interest such as, for example, fuel stations) and specific assistance data which makes it possible to improve the estimation of position.

An information system on road conditions in real time, and in particular motorway traffic conditions, is desirable in order to avoid tailbacks at the entrances and exits to the motorway network, or along the motorway itself, due to accidents or roadworks .

Currently, known services are based on dedicated monitoring structures such as video cameras or directly on signals from the users of the road network.

Currently cellular telephone networks are able to offer, by virtue of their nature, rudimentary position services based on the availability of some time quantities, which must be known for operation of the system itself. Among these are the Timing Advance (TA) utilised in the synchronisation of communication with a base station (BTS) and a mobile terminal, and the observed time difference (OTD) used in the handover procedures.

In consideration of the widespread diffusion of the mobile telephone, in particular of the cellular telephone, it is by now reasonable to hypothesise that the majority of drivers (and passengers) on a journey carry a telephone with them and that this is maintained turned on. It is, moreover, possible that in the course of a journey the driver of a vehicle or a passenger may engage in a telephone conversation.

Patent Application WO 99/44183 suggests deriving information on road traffic conditions by utilising the position and movement information of cellular telephones or similar such mobile communications apparatus, but does not explain in what way to treat the position information collected, which by its nature is highly imprecise . The present invention has the purpose therefore of detecting and estimating the traffic present in a predetermined section of road by exploiting the information obtained from a study of the flow of cellular telephony in the region, as a function of the position information of the communicating apparatus .

The invention relates to the aspect of detection and analysis of the traffic present in a section of motorway by making use of the information on the position of mobile radio communication terminals, with particular reference to the processing techniques of the collected position data.

By means of an instant by instant knowledge of the position of the mobiles, the track and speed of displacement thereof is identified, and, starting from this data, the traffic conditions in two directions along the roadway is determined.

A further object of the invention is to identify the possible presence of an obstacle which generates a traffic jam and to define its position in a relatively short time.

Advantageously this allows signalling to the motorists who are adding to possible tailbacks or slowdowns to offer them the possibility of varying their route.

According to the present invention this object is achieved by a method having the characteristics set out in Claim 1.

Particular embodiments of the invention are defined in the dependent claims . A further subject of the invention is an electronic processing system for carrying out the method, according to Claim 8, and a program which can be loaded into the memory of an electronic processor, comprising a code adapted to perform such method, according to Claim 9.

Further characteristics and advantages of the invention will be explained in more detail in the following detailed description of an embodiment, given purely by way of non- limitative example, with reference to the attached drawings, in which:

Figure 1 is a diagram which represents the overall structure of the traffic detection and estimation algorithm according to the invention;

Figure 2 is a schematic representation of a section of arterial road being monitored, including an area in which there is a stationary mobile phone terminal the direction of travel of which is not known, in a phase of testing for the presence of a traffic jam condition;

Figures 3, 4 and 5 show three different possible relationships between the area in which a stationary mobile phone terminal is located and that containing a mobile phone terminal in motion, the direction of travel of which is not known, in the phase of testing for the presence of a traffic jam condition.

Figures 6-10 show five different possible relationships between areas in which are located a stationary mobile terminal and a moving mobile terminal the direction of travel of which is defined, in the phase of testing for the presence of a traffic jam situation;

Figures 11a-13a. and lib-13b show six different possible relationships between areas in which are located a stationary mobile terminal the direction of movement of which is defined and a moving mobile terminal the direction of movement of which is not known, in the phase of testing for the presence of a traffic jam condition;

Figures 14-18 show five different possible relationships between areas in which are located a stationary mobile terminal the direction of travel of which is defined and a moving mobile terminal the direction of travel of which is defined, the directions of travel being in concordance, in the phase of testing for the presence of a traffic jam condition;

Figures 19-21 show three different possible relationships between areas in which are located a stationary mobile terminal the direction of travel of which is defined and a moving mobile terminal the direction of travel of which is defined, the directions of travel of which are not in concordance with one another, in the phase of testing for the presence of a traffic jam condition;

Figures 22-24 show three different possible relations between areas in which are located a stationary mobile terminal the direction of travel of which is defined and a moving mobile terminal the direction of travel of which is defined, in which the directions of travel are in concordance with one another, in a phase of testing for the end of a traffic jam condition; and

Figures 25-27 show three different possible relations between areas in which are located a stationary mobile terminal the direction of travel of which is not known and a moving mobile terminal the direction of travel of which is defined, in the phase of testing for the end of a traffic jam condition.

The diffusion of the mobile telephone and the possibility of estimating the position of a mobile telephone terminal (hereinafter, more simply, mobile) , are the two conditions which form the basis of the method forming the subject of the invention.

The diffusion of the cellular telephone in Italy and in Europe is greater than 70%. Currently, different companies provide position-locating services on request. The techniques used are often proprietary, but standard strategies exist for achieving the object, although with less precise results.

Diverse architectures exist for position-locating systems, for which the most significant parameter is the margin of error in deriving the position of a mobile .

A cellular network is one of the mobile radio communication systems used in everyday life. The term '"mobile" has been used historically to classify any type of radio terminal which is capable of moving during operation, remaining connected both during sustained high speed displacement, and for hand-held displacements. The term is also used to indicate a wireless system generally.

The Global System for Mobile Communications, commonly known as GSM is a standard for a pan-European, second generation (2G) cellular communications network, published by ETSI, which developed it to resolve the problem of the fragmentation between different analogue standards appearing during the eighties . Being a digitally based technology, GSM offers many advantages, such as the greater immunity to noise, a robust channel interface, the multiplexing of different sources of information (voice, data, video) and an increase in security (encryption procedures) . Moreover, digital transmission permits the use of codes for control of errors, signal processing techniques (source code, encryption) and equalisation procedures such as significantly to improve performance.

The architecture of the GSM system, and the aspects relating to radio access, logic channels and to the signal frames will be known to one skilled in the art for which reason these will not be described in detail hereinafter, also because they are not relevant for the purposes of performance and understanding of the invention.

Since 1997, ETSI (European Telecommunications Standards Institute) has been working in cooperation with a specific sub-group of the TIA (Telecommunications Industry Association) to develop specific and necessary techniques for the position-location service. The object of this standardisation is also to provide a measure of added value to the service to be able to meet the requirements of the entire GSM community.

The position locating service (LCS) can be considered as a technology provided by the cellular network consisting in a standardised service. This possibility allows identification of the position of a specific user in a standard format (for example geographical coordinate) and reports the information to the user, to the network operators or to those who offer services: this knowledge can be used for emergency calls, services based on the position, etc...

The characteristics of each LCS are described by a group of parameters known as location information, consisting of a geographical position, speed and quality of service (QoS) ; in particular, the latter two, indicate the accuracy (which is the typical difference between the location and the estimate) and the response time, and their values depend on the urgency of the position request. The capacity to locate the position of a cellular telephone depends on many factors, and the geographical environment constitutes a fundamental parameter which influences the precision of estimate.

LCS is classified into four categories depending on the use of the position-locating information: commercial: associated with an application which offers a service through the knowledge of the position of the user subscribing to the service; internal: used by the network operators for internal operations; emergency: forming part of a service offered to the assisted user who has called an emergency number; judicial: position information is used as support to legal or penal actions .

There are various ways of deriving the position from a measurement of a transmitted signal and these can be applied to the cellular systems included GSM. The most important measurements are based on the propagation time, on the difference in arrival time (TDOA) on the angle of arrival (AOA) and on the phase of the carrier. Each measurement defines the place in which the mobile should be located.

If the density of the base stations is such that more measurements can be made, it is possible to use a minimum squares approach to combine them and obtain a more accurate estimation of position. These measurements are known to one skilled in the art and will not be explained in more detail .

During the location process the measurements can be ambiguous. There is a large number of sources of ambiguity, including the shape of the wave used to effect the measurement (signal ambiguity) , and the nature of the transformation between the coordinates defined by the location and the cartesian coordinates (physical ambiguity) .

Moreover, delay can be an obstacle to the use of a position measurement. It is possible to determine different types of delay, including the time which elapses from when a user requests a position measurement to when he receives the result (user latency) and the time between when the measurement is effected and when the result is effectively available (movement latency) .

In position location based on GSM it is possible to determine the position of a user, even with a lower signal-to-noise ratio than that required for voice and data. For this reason two figures are defined regarding the coverage of the position-location service: network coverage and geographical coverage. The first is defined as that region over which it is possible successfully both to determine the position and to provide a service for voice transmission. For example, in this area it is possible to implement an emergency service. The second is the ratio between the total area over which it is possible to measure the position and the area over which a telephone service is provided. The second can be larger than the first. Summarising, the accuracy envisaged for the various position location techniques is of the order of one hundred to five hundred metres. Today the technology for position location of a cellular mobile terminal which has made or received a call does not allow a precision of the position obtained to be as good as that provided by GPS technology. Therefore, to exploit the current cellular communication network for the purpose of locating a transmitter associated with a motor vehicle a processing of the raw position data derivable with available technology is necessary.

Hereinafter we will analyse in detail a currently preferred embodiment of a method (which can be performed by means of a processing program) for the analysis of motorway traffic on the basis of signals coming from mobile terminals associated with drivers and/or passengers of vehicles travelling along a motorway and in the neighbouring area in which a position location service is operating as described.

The geographical region under surveillance preferably corresponds to a section lying between two entrances/exits of the motorway. The mobile terminals (useful) which are significant for estimating the changes in the traffic conditions move along the section of the motorway, whilst in the adjacent areas other mobiles (interfering) can be present which can interfere with the estimation. Each mobile is indicated by an identification datum, can commence or terminate a communication, and moves more or less rapidly within the surveyed region or remains involved in a traffic jam.

The evolution of a mobile terminal comprises the displacement information, including entry into and exit from the region covered by the position location service, and the times of commencement and termination of conversation, during which the mobile is traceable.

The specifics of the GSM system provide that in the course of a call, there is an exchange of information between the mobile and the network about every half second.

Evolutions of the cellular telephone, from current GSM systems to UMTS technology, can lead to an improved definition in the position location of the mobiles and to the possibility of making this position location even in the absence of calls, as long as the terminal is active.

In Figure 1 there is schematically shown the structure of the algorithm for analysing the traffic according to the invention.

At the input there is received, from a known position location system, a mobile terminal identification datum (ID) and the coordinates (COORD) of its position in the region under surveillance, affected by a position location error. A plurality of such data constitutes a constellation of points corresponding to the mobile terminals turned on and in conversation in the region under surveillance. For each point of the constellation information is received at regular intervals of time. Each point can correspond to one or more users of the telephone network whether associated with a vehicle or not .

The first operation envisaged by the process according to the invention is that of determining, at step 101, if the mobile terminal under consideration is in motion or is stationary. Therefore, at step 103 it is determined if the mobile is located in the road section under surveillance or is outside of it. At this time there are made available at the output the identification data (ID_road) and position (COORD_road) of the mobiles estimated as belonging to the motorway.

At step 105, after the determination at step 103, and thanks to the knowledge of the position and speed of displacement of each mobile, it is evaluated if the motorway traffic is flowing or slowing.

At step 107 on the other hand a check is made to see if there is a traffic jam and if the stationary mobile is involved in it.

The method forming the subject of the invention has been put into practice by means of a software algorithm. A man skilled in the art with experience of programming would be able easily to understand how to write a complete program based on the process described, therefore we will not describe in detail the program list, but will analyse generally the various steps of the process.

The process operates at different levels of precision of information: the raw level is the estimation of the position of each mobile, therefore, by refining the quality of the data, an evaluation of the type of traffic and the presence or absence of traffic jams on the carriageway is obtained.

At the input are received positions effected by errors (COORD) and the identification data (ID) of the mobile; the coordinates are utilised to verify if the mobile is travelling on the motorway or not. The next step, if the mobile is in motion, is the calculation of its speed and direction or identification of the carriageway in which it is located if it is stationary.

These evaluations are a basis for a subsequent estimation of the traffic. Only mobiles estimated to be on the motorway provide useful indications for this purpose. Mobiles in motion on the motorway are defined as "useful", whilst those outside the motorway section are defined as "interfering" .

A mobile the position of which is established is identifiable as a structure comprising various recognition attributes, among which are at least an identification (ID) of the terminal and a position vector in the region under surveillance acquired in succession at regular intervals in time.

A recognised and analysed terminal will also have associated therewith a speed, the duration of the conversation, the direction of movement, and the condition of being on the motorway or not . A constellation or assembly of mobiles constitutes a vector.

At each sampling instant the vector of the constellation of mobiles is updated by recording new terminals which enter the region under surveillance, cancelling those which depart from it or adding new figures to the temporal succession of position information for mobiles already recorded.

The process forming the subject of the invention estimates if a mobile terminal in communication and present in the geographical area under surveillance is on the motorway or if it is located in a neighbouring area outside the motorway through a series of tests .

Initially it is evaluated if the mobile is in motion or stationary, then the path of the mobile in motion is identified and it is determined if it is located on the motorway. In practice, this analysis is made mobile by mobile at predetermined discrete intervals of time, the duration of which can be preset.

To observe if the mobile is in motion or is stationary, the coordinates of the position of the mobile are used, which are detected at regular intervals, for example each half second, from the position location system.

The algorithm for estimating if the mobile is or is not in motion is as follows.

If the mobile is stationary, whether or not it is on the motorway, the coordinates of its position do not vary with the time. If it makes or receives a call, an exchange of information is set up between the mobile and the network, and therefore this will locate its position. By knowing the statistical distribution of errors commited by the position location system, statistical tests of hypotheses are able to recognise if the mobile is stationary, with a probability of error which reduces with an increase in the number of detections, and is in any event known.

The coordinates of the positions occupied by the mobile during the passage of time are taken to calculate its status. First of all, they must be contiguous in time and sufficiently numerous to increase the reliability of the estimation. However, points too spaced in time cannot be taken because of the risk of using old information which does not correspond to the real state of the mobile. The number of positions used by the system is also a function of the average position error; the greater is the error, the greater is the quantity of requested points .

From the cloud of position points, chosen by means of the time window, the barycentre is calculated as the average of the ordinates and abscissae. The statistical tests consist in testing if the cloud of position points is a sample compatible with the statistical distribution of the position location errors in the reference area. In particular, tests are performed on the variance of the errors on the x and y coordinates of the detected points . These errors are calculated as coordinates of the individual points with respect to the barycentre first calculated. Similarly, it will be possible to define an alternative statistical test by operating on the samples of the distances of the points from the barycentre, by suitably changing the reference distribution.

For example, it could be hypothesised that the position coordinates of a stationary mobile is located within a circle the radius of which coincides with the average position locating error. For each position its distance from the barycentre is measured, which is defined as the radius. The average of the radii is calculated and if this is less than the average position locating error then the mobile is considered stationary, otherwise it is in motion. The strategies for evaluating whether a mobile is located on a motorway or not are different depending on whether it is in motion or stationary.

If the speed of the mobile is nil, to determine if it is located on a motorway or in the adjacent zone, the average of the abscissa and ordinate in the surveyed region is used. If the value falls within an interval the lower and upper limits of which are the extreme confines of the motorway increased by a value proportional to the average positioning error, then the mobile is considered to be on the motorway.

If the mobile were in motion the estimation of the position would be made with the aid of the statistical technique of linear regression. This is a regression of particular type, that is to say an orthogonal least squares regression, since in the case under examination the detected points have both coordinates affected by casual error. By means of this powerful mathematical instrument it is possible to calculate the track of the mobile and therefore to decide if it is moving along the motorway or in the surrounding zone.

When using the regression, a greater number of points would lead to a better and more precise estimation.

The dimension of the window utilised for the regression is potentially limitless, but in practice the upper limit coincides with the duration of the call of the mobile which is limited.

Depending on the environment in which the region considered is located, the distribution of the errors committed by the position finding system is known. This distribution can be used to make a re-sampling (bootstrap) of the positions used in the linear regression calculation as necessary.

The location of a mobile is subject to various factors which, in this context, do not perform an important role. It can be observed that, in general, the unreliability of the position location has repercussions on the precision of the positions used in the analysis. The further the points are from those which have been verified as true the greater is the uncertainty of the estimation. However, in the case in question the statistical correlation of the space-time data observed with the known track of the motorway section, as with the neighbouring sections, is a very powerful means to drastically reduce the associated uncertainty in recognition. This correlation becomes gradually more discriminating with an increase in the number of points detected and suitably exploits the knowledge of the local distribution of position location errors. It is thanks to this principle that a correct recognition can be achieved with a very high and in any event controllable probability even in the case of detection of positions by a GSM system, which are characterised by very high position location errors .

To effect re-sampling it is necessary to define a technique which calculates the uncertainty in the positions due to the position location process. From a Bivariate Gaussian Probability distribution two values are extracted, one will be summed to the abscissa of the real position and the other to the ordinate. The parameters of the distribution are the average and the variance. The first is nil. The second must be in relation with the typical position location error. For each position used by the regression a number of points about it are calculated using the first cited technique; all the new positions derived are located, with a good probability, within a circle centred on the point chosen for the regression and having a radius proportional to the typical position-location error. This makes it possible to increase the quantity of positions utilised for the calculation of the regression and to obtain an estimation of the track which closely approaches the true one.

It then passes to calculations of the path of the mobile in motion and utilises the path calculated to evaluate if the mobile is on the motorway.

For simplicity of explanation, the section of motorway in the geographical area under surveillance can be considered linear and placed in the middle, parallel to the axis of the ordinate. It can be observed that this hypothesis is not at all restrictive. On the contrary, in the case of curvilinear sections, the known path of the section is more informative than in the case of a rectilinear section. In fact, in the first case, a more precise space-time positioning of the mobile is possible than in the second in that the correlation between the known path and the statistically estimated path based on the successive position-locating data is more discriminating thanks to the greater information contained in the path of the curvilinear section.

The technique for determining if the mobile is on the motorway tests if the straight line with nil angular coefficient is contained in the interval of confidence of the angular coefficient of the regression straight line. If it is effectively found in the interval, the mobile probably W

19

moves on the motorway. To eliminate the ambiguity in the decision due to the mobiles which are found outside the motorway and the path of which is parallel to the ordinate, a final test is performed. This calculates the average of the abscissae of the positions used to calculate the regression and tests if this value falls within the interval the lower and upper limits of which are the confines of the motorway increased by a value proportional to the average position location error. If the result is positive the system decides that the mobile in motion is located on the motorway.

Subsequently, for each mobile on the motorway the speed and the direction in which it moves is calculated.

The mobiles which have been estimated as being outside the motorway are not abandoned but their track is memorised and the system continues to survey them. It discontinues considering a mobile only and exclusively when it stops communicating. It is to be supposed that not all mobiles are considered correctly at the same time and that some of them will not be ever considered.

All the mobiles in conversation can be located and, via the knowledge of their assumed positions, are analysed as described. The direction and speed is only calculated for those considered to be on the motorway. For the remainder, the coordinates of their positions are memorised and at established intervals of time, different for each mobile, they are used to determine if the mobile is useful or interfering until it finishes communicating.

The value of the speed is calculated according to the relation: distance travelled/time elapsed. Assuming a monotonic evolution of a motorway section in the region under surveillance, the direction is easily identified by comparing the coordinates of the first and last detected position.

For the calculations considered, the coordinates of the relative positions of the mobiles in communication are affected by positioning errors .

The subsequent steps for estimation of the traffic and the study of a traffic jam require only and exclusively mobiles on motorways. For evaluation of the traffic it is necessary that the mobiles be on motorway and that may be in conversation. A mobile considered as useful can in reality be an interfering one and in this case the system commits an error.

The ultimate objective of the analysis is the evaluation of the traffic present on the motorway utilising only the knowledge of the positions of .the mobiles in communication.

The system effects, instant by instant, an evaluation of the type of traffic present .

There are three types of traffic which the system estimates: stationary, slow and flowing traffic.

Each is characterised by precise speed intervals. A mobile in motion is inserted in one of the types of traffic, if its speed falls in that speed interval .

For example the traffic is considered as flowing if its speed is greater than 80km/h, slow if its speed is greater than lOkm/h and stationary for speeds lying between Okm/h and lOkm/h.

When evaluating the traffic present on the motorway the system decides that the situation does not vary instant by instant, but remains constant for a predetermined finite time interval. This information serves to identify the section of the carriageway where the mobile is moving and for which it is possible to evaluate the type of traffic .

For each mobile on the motorway it is possible to identify the section along which it moves and - by means of the speed value - the type of traffic present . These two types of information identify a zone, the unit used by the system for evaluation of the traffic and for study of the presence of traffic jams.

It can be understood that in a given instant there may be no mobiles in conversation and on the motorway. Consequently it is not possible to evaluate the traffic.

If there are mobiles present on the motorway the system is able to estimate the traffic. The carriageway on which the mobile moves is identified by observing its direction.

The evaluation, instant by instant, of the traffic present is reported at the output. The first series of data (TR) reports the estimation of the traffic effected every half second, whilst a second series of data (BLOC) adds information on possible mobiles in a traffic jam.

By traffic jam is meant that situation in which the traffic is stationary and in a column. The reasons for this arrest can be various, for example an accident or road works in progress. The consequence is an interruption in the flow of motor vehicles on one of the two carriageways for a longer or shorter time interval .

To illustrate the situation, suppose that the mobiles involved in the traffic jam stop immediately and are disposed in a single row. It is hypothesised that two thirds of the mobiles involved communicate their situation and therefore are in conversation during the jam. This hypothesis makes it possible to locate the stationary mobiles and therefore have means for establishing the presence of the possible stoppage. Once the system has determined the presence of a jam the problem is to determine which of the two carriageways is involved.

The calculation of the direction of the mobiles is effected only for the mobiles considered in motion. For those which are stationary there is no directional sense given that the positions subsequently assumed are all located in the region of the point really occupied by the mobile.

The solution chosen for identifiying the carriageway of the mobile in the traffic jam uses the method of the zones defined above to define the motorway section on which it will be possible to determine the type of flow of traffic present.

For a stationary mobile it is possible to identify its zone but not its direction. For one in motion on the other hand the direction is also identified. Therefore it is possible to determine the carriageway of the stationary mobile by comparing its zone with that of a mobile in motion. If they are close, then the presence of motion on one carriageway permits the system to position the jam in the opposite carriageway where the stationary mobile is also present.

Mobiles with zero speed are subdivided into two subclasses depending on their history. In the first, the least numerous, is found mobiles already in communication before being stopped by the jam. The fact that before stopping they were in motion allows knowledge of their direction with good precision. They can be utilised to eliminate ambiguity on the choice of the blocked carriageway.

The second comprises mobiles which start to communicate when they are already stationary and the position, and therefore the carriageway on which they are located, is not known.

The steps which the system takes to manage the presence of a jam will be examined by considering the two cases separately.

The system perceives the presence of a mobile only if it is in communication. Therefore the ones which provide useful information for the study of the jam are those which are stationary and in conversation.

The most common situation is that of detected mobiles which commence to communicate when already stationary.

A mobile involved in the jam commences to talk when already stationary. The system identifies it, observes that its speed is nil and memorises its data.

In the other case it is on the other hand possible to know the direction towards which the mobile was moving before stopping. Once the system decides that the mobile is stationary its original direction is maintained valid for all the analysis. The condition needed to establish that the direction of the mobile is valid is dictated by the duration of the call . A duration greater than a predetermined threshold can lead to a greater precision in estimation.

In any case, one stationary mobile is analysed at a time, using the remainders to evaluate its state.

The module evaluates a mobile and locate it on the motorway after a series of tests. Subsequently it checks to see if it is in motion or not. If it is judged stationary this is signalled to the module which then manages the possible presence of a traffic jam. This has the function of evaluating if that mobile is truly involved in the traffic jam or if an error has been committed. However, it is not always possible to discard one of the two hypotheses and the system must then wait to collect other information in order to make a decision.

The state is an evaluation of correctness of the estimation of the stationary mobile.

There are four possible states relating to a mobile: "traffic jam", "traffic jam hypothesised", "Unknown", and "error", and each involves a different action to be performed.

The "traffic jam" state indicates that the mobile is involved in a traffic jam- but does not give any information on its position.

A mobile is considered to be in a traffic jam only if the duration of the call is greater than a predetermined threshold in order to guarantee security in the estimation. Subsequently the identification of the mobile in the traffic jam is memorised.

The state "traffic jam hypothesised" advises the system to proceed to a more accurate estimation.

The state "Unknown" cancels the memorised data of the mobile used to evaluate its state.

Finally, the state "error" advises that the mobile is not in a traffic jam and that an error has been committed. Consequently, the mobile is cancelled from the list of mobiles in the traffic jam.

Operatively, the process according to the invention provides for examination of all the mobiles estimated on the motorway. If a mobile is encountered with a zero speed, an index is memorised in a vector of traffic jammed mobiles to locate it in the memorisation vector of mobiles in conversation and it proceeds to the study of its condition.

By hypothesising the presence of an obstacle in a carriageway it is clear that the flow of automobiles is interrupted. This implies that in the zone adjacent to the stoppage there is no traffic in motion. If the condition of absence of traffic lasts for a rather long time the presence of the traffic jam can be hypothesised.

Normally, it is possible to calculate the direction of travel of the mobiles in motion by observing the positions assumed over time; for a stationary mobile all the points concentrate in a restricted zone centred on the point where it is really located. Therefore, for mobiles with a nil speed it is not possible to determine the direction.

The object of the algorithm is to verify the presence of the traffic jam and to establish in which of the two carriageways the obstacle which interrupts the flow of traffic is present .

One possible solution consists in deducing the carriageway on which the stationary mobile is located by observation of the motion of motor vehicles in both directional senses. If the absence of motion in one sense is determined and the presence of mobiles in motion in the other, it is deduced that the obstacle is located in the carriageway in which there is no traffic present.

Only mobiles in conversation for a time greater than a predetermined threshold of call duration and estimated to be on the motorway are used for the analysis .

A "valid" mobile is defined as one which respects the conditions on the call duration and which is utilised by the algorithm for the estimation of the state of the stationary mobile.

For each mobile the system calculates the zone occupied.

Each type of traffic is characterised by precise speed intervals .

The algorithm is based on the comparison between the zone of the stationary mobile and the zone of that in motion. Not all the mobiles estimated to be on the motorway contribute to the determination of the state of the stationary mobile. In order to be considered valid and useful for the purpose they must respect a condition on the duration of their conversation. The greater the duration of the call the greater is the quantity of data possessed and more precise will be the evaluation of the characteristics of the mobile.

The possible cases are the following:

1. Stationary mobile which starts a communication after stopping

When considering the case of a stationary mobile which starts a communication after stopping the value of direction is not certain; this implies the necessity of maintaining a constant observation in the neighbouring zones to the zone of the mobile with nil speed for both the carriageways. Conceptually, it is as if the zone of the stationary mobile (zone S) were to extend over both the carriageways of the motorway A as shown in Figure 2.

The relationships between the zones of the stationary mobile (zone S) and of that in motion (zone M) used for the comparison, which allow a determination of which of the two carriageways is affected by the presence of an obstacle are examined. To make the description clearer the three different values of direction of the mobile considered useful for the estimation will be dealt with separately.

A mobile the direction of travel of which is not defined, clearly cannot be utilised to identify the blocked carriageway; possibly it provides indications on the state of the stationary mobile. For this purpose a condition on the speed is required, that is to say it must also be stationary.

For the stationary mobile to be judged to be in a traffic jam it is necessary that the two zones have at least one superimposed part.

If the direction is estimated, all the mobiles are useful, independently of their speed, if their position is close to the zone of the stationary mobile.

In Figures 3, 4, 5 are shown the cases, respectively, in which the two zones are partially superimposed, the zone of the stationary mobile lies entirely within the zone of the moving mobile, or else incorporates this latter.

The value of the speed and direction, suitably combined, provides both the state of the stationary mobile and the carriageway of the possible blockage.

Schematically:

If the mobile in motion does not respect the condition on the time of duration seen above which confirms the accuracy of the estimate, the proposed state is always "Unknown" and the proposed direction is indefinite .

In the case in which the direction of travel of the useful mobile is defined the cases illustrated in Figures 6-10 are used or the respective dual cases are used if the direction of the useful mobile detected is the opposite.

The relations between zones described above are employed to verify the state of the stationary mobiles.

Of the stationary mobile the end of its zone and its barycentre are calculated. For each memorised mobile on the motorway the zone is calculated and by exploiting the algorithm seen above an attempt is made to estimate the state of the stationary mobiles .

In one structure VETT the state and direction deduced and some characteristics of the mobile which has permitted the analysis to be performed are memorised.

If after having considered all the possible mobiles estimated to be on the motorway the VETT structure is empty it signifies that in the zone neighbouring the zone of the blocked mobile the mobiles are absent or do not enter into the envisaged survey; the absence of motion is decreed.

The vector of blocked mobiles includes the stationary mobile considered, and the mobile could be involved in the blockage. A counter records the duration of the time interval in which movement in the region of the stationary mobile is not encountered. If the period is greater than a predetermined threshold then the mobile is considered to be stationary, in the traffic jam.

If the VETT structure is not empty and the data on the state of the mobile are different from "Unknown" , then the function was able to provide an evaluation of the traffic jam.

If there are present, for the same stationary mobile, different evaluations of its state, the most restrictive is "traffic jam" followed by "traffic jam hypothesised".

For the direction of the mobile considered, if opposite indications were to be present it assumes that of the majority.

The same considerations can be applied to the case of a stationary mobile which has terminated its conversation.

2. Stationary mobile which has started a communication before stopping

If the stationary mobile considered was in motion before and the conversation has remained active during change in speed, it is clearly possible to know the value of direction.

The two cases corresponding to the opposite directions of the now-stationary mobile will be treated separately. The relations between the zones are different depending on the direction of the useful mobile in motion used by the algorithm.

Tiie direction of the useful mobile is not defined

A mobile the direction of travel of which is not defined clearly cannot be utilised to identify the blocked carriageway; it possibly provides indications on the state of the stationary mobile. For this purpose it requires a condition on the speed which must be nil.

In order for the mobile to be judged to be in a traffic jam it is necessary that the two zones have at least one superimposed part .

If the two zones are partly superimposed (Figure 11a, lib) , the zone of the stationary mobile is contained in the other (Figure 12a, 12b) or contains the other (Figure 13a, 13b) then the table which relates the speed of the valid mobile and the state of the stationary mobile is as follows:

The direction of the useful mobile is defined If the directions are concordant (the same) the possible situations are graphically illustrated in Figures 14 to 18 and correspond to the cases in which the two zones are contiguous, spaced, or partly superimposed, in which the zone of the stationary mobile is contained in the other, or the zone of the stationary mobile encompasses the other.

The relations between the speed of the valid mobile and the state of the stationary mobile are expressed by the following table:

Naturally, the same considerations are applied in the reciprocal case, in which both the mobiles are on the opposite carriageway.

If the directions are not concordant, the possible situations are graphically illustrated in Figures 19-21 and correspond to the cases in which the two zones are positioned on the two carriageways, the zone of the stationary mobile is contained in the other or the zone of the stationary mobile contains the other.

Naturally, the same conditions apply in the reciprocal case, in which the mobiles are located on the mutually opposite carriageway.

For each mobile estimated to be on the road the zone is calculated and the comparison with that of the stationary mobile is calculated; if the relation is found to be within one of the listed cases then the state of the analysed mobile and the attributes of that utilised for the estimation are memorised in the vector VETT.

It can happen that different valid mobiles are present and that each provides an evaluation of the state. The number of instances for each possible state are quantified and that in the majority is chosen as the final state.

If the presence of a traffic jam is hypothesised or given a mobile with zero speed and estimated as in a traffic jam (or not) , it is finally checked if the mobile under consideration remains involved or not in the stationary traffic situation.

This function can prove useful in two situations: relating to an error or estimation and to determine the instant in which the carriageway becomes free. Below is described the algorithm listing the useful relations between zones, for the purpose of achieving the predetermined object.

The direction of the stationary mobile is known

The direction of the useful mobile is chosen in concordance. The relations between the zones are illustrated in Figures 22-24, that is the zones are partially superimposed, the zone of the stationary mobile is contained in the other, the zone of the stationary mobile contains the other.

In the first two cases the relations between the speed of the valid mobile and the state of the stationary mobile are summarised in the following table:

In the latter case the relations between the speed of the valid mobile and the state of the stationary mobile are summarised in the following table:

Naturally, the same considerations apply in the reciprocal case in which each of the mobiles are located in the opposite carriageway.

The direction of the stationary mobile is not known

The relations between the zones are illustrated in Figures 25-27, that is the zones are partially superimposed, the zone of the stationary mobile is contained in the other, and the zone of the stationary mobile contains the other.

In the latter case the relations between the speed of the valid mobile and the state of the stationary one are summarised in the following table:

Stationary Traffic Jam

Naturally, the same considerations apply in the reciprocal case in which both mobiles are located on the opposite carriageway.

Once the analysis by means of the zone method is terminated, the presence of four possible states is computed, namely End of Traffic Jam, Traffic Jam, Traffic Jam Hypothesised, Unknown.

At this point the method distinguishes the cases according to whether the direction of the stationary mobile is known or not. If its direction is known, then the number of estimations of the end of the traffic jam is tested; if it is different from zero, the mobile considered is estimated to be no longer stationary.

If its direction is not known, the condition needed to consider the mobile not estimated correctly or no longer in a traffic jam is that in both the carriageways there are present mobiles in motion. Then, as well as testing the occurrence of the estimation of End of Traffic Jam not being zero it is necessary to test the direction of the valid mobiles .

Naturally, the principle of the invention remaining the same, the embodiments and details of implementation can be varied with respect to what has been described and illustrated purely by way of non limitative example, without by this departing from the scope of protection defined in the annexed claims .

Claims

1. A method for the detection and estimation of road traffic on the basis of position-locating information of mobile terminals of a radio communication system, comprising the operations of; acquisition, at predetermined time intervals, of sets of data relating to a plurality of active mobile terminals present within a predetermined surveillance region comprising, for each terminal, terminal-identification data and a succession of respective geographical coordinate data for position-location within the said surveillance region, the said position-locating coordinate data being affected by a position-locating error; memorisation of the said acquired data in a data matrix of the said plurality of terminals; determination for each mobile, of the barycentre of the said position-locating coordinates acquired in succession over time; comparison, for each mobile, between the distance of the points identified by the succession of coordinate data of a terminal from the barycentre and a predetermined average position-locating error; memorisation of a plurality of first binary data relating to the state of a respective plurality of mobile terminals in the matrix of data of the said plurality of terminals, each datum assuming a first value indicative of a condition of stationary mobile when the average of the distances between the points identified from the succession of coordinate data of one terminal and the barycentre is less than the said average position-locating error, or a second value indicative of a condition of a moving mobile otherwise; memorisation of a plurality of second binary data relating a respective plurality of mobile terminals to a predetermined road section in the matrix of data of the said plurality of terminals, each datum assuming a first value indicative of a mobile terminal belonging to the road section, for which the mobile terminal is considered a useful mobile, i) when the barycentre of points identified from the succession of coordinate data of a terminal lies within the confines of the road section, in a stationary mobile terminal condition; or ii) when the interpolation curve of the said points locally form with the track of the road section an angle less than a predetermined threshold value, in a moving mobile terminal condition, and a second value indicative of the mobile being in the region neighbouring the road section, for which the mobile terminal is considered an interfering mobile, otherwise; calculation of the average speed of the useful mobile terminals on the basis of the temporal evolution of the points identified from the succession of coordinate data of the associated terminal, and comparison of the resultant value with a plurality of thresholds representative of a plurality of types of traffic, and generation, at predetermined time intervals, of at least one datum indicative of the estimated traffic condition on the basis of the said threshold comparison for each direction of travel.

2. A method according to Claim 1, in which the direction of travel of a useful mobile terminal is determined on the basis of evolution of the identified points from the succession of T2003/000552

39

coordinate data of the associated terminal only in a condition of the mobile terminal being in motion.

3. A method according to Claim 2, characterised in that the calculated speed of travel and the direction of travel which may possibly be determined are memorised in the said data matrix.

4. A method according to Claim 3 , characterised in that the speed and direction of travel are determined on the basis of a comparison between the first and last memorised data of the temporal succession of position-locating coordinate data.

5. A method according to any preceding claim, in which the said region of surveillance is determined as the region comprising a section of roadway under surveillance lying between two access/exit roads.

6. A method according to Claim 3, further comprising, as a function of a determined condition of stationary mobile terminal, the operation of memorisation of the data relating to the stationary terminal in a memory vector of terminals in a traffic jam.

7. A process according to Claim 6 , further comprising the operation of testing for a traffic jam condition on the road section under surveillance, as a function of the presence of data relating to one or more stationary terminals in the blocked terminals memory vector, by comparison with the speed and direction data of the useful terminals in motion memorised in the said data matrix, the said testing operation being effected on the basis of the relations expressed as: i) a first decision table:

if the direction in which the stationary mobile was previously travelling is not known; ii) a second decision table:

If the direction in which the stationary mobile was travelling is known and the direction of the useful mobile is not defined or is in concordance; and iii) a third decision table:

if the direction previously followed by the stationary mobile is known and the direction of the useful mobile is not in concordance .

8. An electronic processor system for performance of the method according to Claims 1 to 7, comprising: means for acquisition of data relation to a plurality of active mobile terminals present within a predetermined surveillance region, comprising, for each terminal, terminal- identification data and a succession of respective position- locating geographical coordinate data for locating the position within the said surveillance region, and a processor unit comprising first memory means for storing data as a matrix of data of the said plurality of terminals and arranged to determine, for each mobile, the barycentre of the said position-locating coordinates acquired in succession over time and compare, for each mobile, the distance of the points identified by the succession of coordinate data of one terminal from the barycentre with a predetermined average position-locating error; the processor unit further comprising: second memory means for memorisation of a plurality of first binary data relating to the state of a respective plurality of mobile terminals of the matrix of data of the said plurality of terminals, each datum assuming a first value indicative of a stationary mobile condition when the average of the distance between the points identified by the succession of data coordinates of one terminal and the barycentre is less than the said average position-locating error, or a second value indicative of a condition of mobile in motion otherwise; and third memory means for storing a plurality of second binary data relating to whether a respective plurality of mobile terminals belong to the predetermined road section in the matrix of data of the said plurality of terminals, each datum assuming a first value indicating that the mobile terminal belongs to the road section, for which the mobile terminal is considered a useful mobile, when the barycentre of the points identified by the succession of data of coordinates of a terminal lies within the boundary of the said road section in a stationary mobile terminal condition; or when the interpolation curve of the said points locally forms, with the hack of the road section, an angle less than a predetermined threshold value, in a mobile terminal in motion condition, and a second value indicative of whether the mobile belongs to the region neighbouring the road section, for which the mobile terminal is considered an interfering mobile otherwise; the processing unit being further arranged to: determine the direction of travel and calculate the average speed of the useful mobile terminals on the basis of the temporal evolution of the points identified by the succession of coordinate data of the associated terminal, compare the resultant value with a plurality of thresholds representative of a plurality of types of traffic, and generate at least one datum indicative of the traffic condition estimated on the basis of the said threshold comparison, for each direction of travel, at predetermined time intervals.

9. A program loadable into the memory of an electronic processor, comprising a code adapted to carry out the method according to Claims 1 to 7 when the program is run in an electronic processing system.