WO2012089277A1

WO2012089277A1 - Manuevre analysis, direction of traffic flow and detection of grade separated crossings for network generation in a digital map

Info

Publication number: WO2012089277A1
Application number: PCT/EP2010/070965
Authority: WO
Inventors: Heiko Mund
Original assignee: Tomtom Germany Gmbh & Co. Kg
Priority date: 2010-12-31
Filing date: 2010-12-31
Publication date: 2012-07-05

Abstract

A method for maneuver calculation using traces derived from probe data is provided. The method provides a digital vector map configured Io store a plurality of line segments, shape points, and maneuvers, each of the maneuvers comprising at least three consecutive shape points or junctions spatially associated within a coordinate system. Each of the plurality of maneuvers includes a maneuver weight value. Each line segment in the digital vector map is associated with, a line weight value. A probe trace is provided and map matching criteria are established. The matched probe data is utilized to adjust ihs Sine weight values. The probe trace is associated with one of the plurality of maneuvers and the maneuver weight value is adjusted based on the probe trace. The adjusted maneuver weight value is utilized to update maneuvers available at an individual shape point or junction.

Description

MANUEVRE ANALYSIS, DIRECTION OF TRAFFIC FLOW AND DETECTION OF GRADE SEPARATED CROSSINGS FOR NETWORK GENERATION IN A DIGITAL MAP

BACKGROUND OF THE INVENTION

Field of the Invention

[0001] This invention relates generally to a method for determining manuevre availability, direction of traffic flow, and grade separated crossings using probe data, and more generally toward a method for improving traffic pattern data using incremental and time generated models.

Related Art

[0002] Navigation systems, electronic maps (also known as digital maps), and geographical positioning devices are increasingly used by travelers to assist with various navigation functions, such as to determine the overall position and orientation of the traveler and/or vehicle_* find destinations and addresses, calculate optimal routes, and provide real-time driving guidance. Typically, the navigation system includes a small display screen or graphic user interface that portrays a network of streets as a series of line segments. The traveler can be generally located on the digital map close to or with regard to a road or street.

[0003] Digital maps are expensive to produce and update, since collecting and processing road information is very costly. Surveying methods or digitizing satellite images are commonly employed techniques for creating a digital map. Furthermore, digital maps are likely to contain inaccuracies or systematic errors due to faulty or inaccurate input sources or flawed inference procedures. Once a digital map has been created, it is costly to keep map information up to date, since road geometry changes over time. In some regions of the world, digital maps are not available at all.

[0004] It may be the case that an existing roadway map or network is incomplete in its depiction of all roadways or paths within a given region. Furthermore, due to the evolving nature of networks which may include but are not limited to roadways and paths, changes may occur over time such that an existing digital map may no longer accurately portray current conditions.

[0005] Nowadays vehicle probe data are used to hold road networks up-to-date. With different approaches it is possible to generate a new network from probe data. One class of algorithms used for this purpose are known as incremental algorithms. Incremental algorithms have several advantages, although other non- incremental network generation algorithms are also useful and sometimes preferred. One primary advantage of incremental algorithms is that an incremental approach allows extending and improving a network without need to process the whole network again. Such an algorithm is for example described in the Applicant's co-pending patent PCT application titled "INCREMENTAL MAP GENERATION, REFINEMENT AND EXTENSION WITH GPS TRACES" by inventor H. Mund (PCT/EP2009/063938 filed 22 October 2009) which is herein incorporated by reference. The techniques described therein are referred to hereafter as the Viae Novae algorithm.

[0006] It is known from the Viae Novae algorithm for example, to take probe data inputs from low-cost positioning systems and handheld devices and mobile phones with integrated GPS functionality for the purpose of incrementally learning a map using certain clustering technologies. The input to be processed consists of recorded GPS traces in the form of a standard ASCII stream, which is supported by almost all existing GPS devices. The output is a road map in the form of a directed graph with nodes and edges annotated with travel time information. Travelers appropriately fitted with navigation devices and traversing a main trunk and/or branch junction may thus create a trace map. with nodes created at regular distances. The nodes and edges are stored in a digital vector map table or database. Through this technique which represents an incremental approach, road geometry can be inferred, and the collected data points refined by filtering and partitioning algorithms.

[0007] Figure 1 describes this exemplary incremental approach in terms of a simplified flow process. A plurality of probes 14 are depicted as GPS-enabled personal navigation devices such as those manufactured by TomTom NV (www.tomtom.com). However, any suitable device with GPS functionality may be used to generate probe data points, including handheld devices, mobile phones, PDAs, and the like. The probe data points may be collected and stored in a probe data table 16 or other suitable database or repository. The existing digital vector map, in this example a previously created digital map, is contained in a table 18. Of course, the digital vector map 18 can exist as a database or in other suitable form. Trace lines are generated from the rough probe data in table 16 as an initial step. A new line is selected from the probe data table at step 20. The selected line is matched to the digital vector map at step 22. During this step, each point of the trace line will be associated with a network element using a suitable map matching method. If the matching method is not able to associate any network element to the trace line, the probe data point is marked as "unmatched." AH other probe data points are attempted to be matched to the existing network in this manner. For trace line segments whose data points are not matched to any element of the existing network, these must be split from the network elements and inserted, i.e., connected, via a new or existing junction. This occurs in step 24. In some cases, it is reasonable to use a known or preexisting junction. Junctions may be referred to as intersections in a roadway network application. Once the matching and junction steps 22, 24 have been completed, the trace line segments are merged with the associated network elements at step 26. This may be accomplished with merge suitable algorithms and methodology. Finally, to reduce the number of shape points, it is possible to simplify the network element before updating the network table at step 18. This optional simplification step is indicated at function block 28 and can be accomplished through various techniques, including by application of the well-known Douglas-Peucker algorithm.

[0008] Although such an incremental approach has been contemplated to update road networks in a digital map system, there has not been an approach for improving route calculations to provide turn by turn directions to end users. In order to use road networks for route calculations, a digital map system requires additional attributes such as the direction of traffic flow, maneuvers on crossings, mean speed, presence of grade separated crossing and others. These attributes are commonly assigned to a road system and remain fixedly associated with map elements. In reality, however, route calculation attributes may change over time and may be altered without changes to the existing road network infrastructure. It would therefore be highly desirable to have a method for calculation of these attributes during a network generation updating process such as the Viae Novae process. By integrating a maneuver detection into an incremental network generation algorithm the route calculation could be improved simultaneously with the infrastructure network data.

[0009] Incremental improvement utilizes new trace data to easily refine and extend a road network in a digital map system. When using an incremental approach as described above, it is not possible to remove old data from the generated road network. This would be equally true for old maneuver data stored within the digital map system. A problem, however, lies in that the remnant of old, possibly outdated, trace data will be factored into the incremental algorithms when analyzing the extent to which a route calculations are to be extended or improved, and thus continue to exert influence over the analysis. The usage and importance of an allowed maneuver would be described by its weight value. If a maneuver is not permitted anymore, it could still have a high weight and therefore continue to negatively influence the map refinement and extension exerc ises.

[0010] Accordingly there is a not only a need for integrating maneuver detection into existing incremental road network generation techniques, but there is additionally a need for an improved method for updating and extending route calculations within digital vector maps using probe or trace data, and that is not susceptible to the negative influence of old, possibly outdated, trace data. The method should be useful in conjunction with both incremental and non- incremental network generation algorithms.

SUMMARY OF THE INVENTION

[0011] This invention overcomes the shortcomings and deficiencies of the various prior art techniques by providing a method for maneuver calculation using traces derived from probe data. The method provides a digital vector map configured to store a plurality of line segments, shape points, and maneuvers, each of the maneuvers comprising at least three consecutive shape points or junctions spatially associated within a coordinate system. Each of the plurality of maneuvers includes a maneuver weight value. Each line segment in the digital vector map is associated with a line weight value. A probe trace is provided and map matching criteria are established. The matched probe data is utilized to adjust the line weight values. The probe trace is associated with one of the plurality of maneuvers and the maneuver weight value is adjusted based on the probe trace. The adjusted maneuver weight value is utilized to update maneuvers available at an individual shape point or junction.

[0012] The novel idea is to use trace data to produce weighted maneuver information to provide a basis for calculating and updating turn restrictions, determining the direction of traffic flow, and determining the presence of grade separated crossings. The present invention contemplates the use of both simple weighting techniques as well as advanced time dependent algorithms. Exemplary, alternative methods are proposed for adjusting the weight value of the maneuvers and/or the traces. The weights can additionally depend from the accuracy and other quality measures. Combinations of all methods are possible.

[0013] By this method, an already existing digital vector map, such as that used for road maps, bicycle maps and footpath maps and the like, can be improved using collected probe data. The turn by turn instructions can be improved by increasing the accuracy of turn restrictions, traffic flow, and other maneuver generated utility. Furthermore, the general concepts of this invention can be used to improve any digital vector map, not only roadway and pathway maps.

BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a schematic diagram depicting a process for collecting probe (trace) data from a plurality of GPS-enabled navigation devices, storing the probe data in a table, and then matching individual traces to line segments in a digital vector map for the purpose of improving and/or extending a digital vector map;

[0015] Figure 2 is an illustration of a vector data map in accordance with the present invention;

[0016] Figure 3 is generalized flow diagram maneuver calculation using traces derived from probe data according to the invention;

[0017] Figure 4 is a comparative graph showing the effects of adjusting trace weights as a function of time-span from a recalculation date (Week. 0,00) according to a first alternative embodiment of the invention;

[0018] Figure 5 is a simplified diagram corresponding to Table 3, illustrating an implementation scheme according to the first alternative embodiment of the invention;

[0019] Figure 6 is a simplified diagram illustrating a hypothetical implementation scheme according to a second alternative embodiment of the invention;

[0020] Figure 7 is a simplified graph showing the exponential decay of the manoeuver and/or trace weight according to the second alternative embodiment;

[0021] Figure 8 a simplified diagram illustrating an implementation scheme according to a third alternative embodiment of the invention; and

[0022] Figure 9 shows an exemplary alternative decay function according to the third alternative embodiment of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

[0023] Nowadays vehicle probe data are used to hold road networks up-to-date. With different approaches it is possible to generate a new network from probe data. One class of algorithms are incremental algorithms. They have several advantages. The incremental approach allows extending and improving a network without to process the whole network again. Such an algorithm is described in the Viae Novae algorithm. A complete description of the incremental approach is described in the co-pending application "INCREMENTAL MAP GENERATION, REFINEMENT AND EXTENSION WITH GPS TRACES" by inventor H. Mund which is herein incorporated by reference (PCT/EP2009/063938). Although these techniques have been utilized to generate new network data, they have not been utilized to evaluate and update traffic pattern availability such as maneuver availability, traffic direction, etc. The present invention contemplates a novel approach to applying the incremental approach not only to road infrastructure, but to determination of route calculation options.

[0024] The present invention further contemplates the use of time dependent weights in order to remove the influence of road segments no longer in use from remaining an inappropriate influence on network generation. Strict incremental approach has not made it possible to remove old data from the generated road network. The usage and importance of a road element is described with reference to its weight value. If a road segment is not in use anymore it may have still a high weight. In order to avoid a re-computing of the whole network one can reduce the weight of old data. The process for use of time dependent weights is described in detail in the inventor's co-pending application "TIME DEPENDENT WEIGHTS FOR NETWORK GENERATION IN A DIGITAL MAP" by inventor H. Mund which is herein incorporated by reference (PCT/US2010/61973). The present invention contemplates the use of these time dependent weighting techniques to evaluate and update traffic pattern availability such as maneuver availability, traffic direction, etc. The use of time weighting techniques for evaluation of maneuver availability allows for the route calculation options to be continuously kept up to date. This is especially valuable as traffic patterns may be altered even more frequently than infrastructure changes.

[0025] The present invention is directed towards the introduction of weighted maneuver calculations into a system of generating, refining or extending a digital vector map using traces derived from probe data. An example of a digital vector map 30 is illustrated in Figure 2. The digital vector map 30 is comprised of a plurality of shape points 32 (or junctions) and a plurality of line segments 34 positioned there between. A maneuver 36 is defined as a triple of three consecutive shape points 32 or junctions 33. Each shape point 32 has a series of associated maneuvers 36 representing potential travel routes. In the illustrated example, shape point q would typically have only two possible maneuvers 36 (p,q,r) and (r,q,p) assuming that there are no u-turns. Although for simplicity, the present discussion will not include u-turns they would be easily implemented in light of the present disclosure by simply including maneuvers such as (p,q,p) While shape point q only has limited possibilities, junction r would have twelve possible maneuvers (q,r,s), (q,r,t), (q,r,u), (s,r,t), (s,r,q), (s,r,u), (t,r,q), (t,r,s), (t,r.u), (u,r,t), (u,r,s), and (u,r,q). It is equally contemplated that each possible maneuver may be defined in terms of an incoming line segment 38, a shape point 32 or junction 33 (such as r), and an outgoing line segment 40.

[0026] Referring to the figures, wherein like numerals indicate like or corresponding parts throughout the several views, the subject invention is described schematically in Figure 3 as a generalized flow diagram for generating, refining and/or extending maneuver information within a digital vector map using traces derived from probe (i.e., trace) data. Although the maneuver detection is illustrated as integrated into the network generation process, it should be understood that the maneuver detection may be implemented independently from the larger network generation/refining process. The method generally includes providing a digital vector map configured to store a plurality of line segments 34, junctions 33 and shape points 32 spatially associated within a coordinate system. This is represented in Figure 3 as a generated network 42. The generated network 42 stores at least one line segment 34, and more typically a substantial plurality of line segments 34. Each line segment has an associated line weight value that has been confirmed in relation to a specified date. For example, the line weight value may be comprised of a calculated sum of the individual trace weights historically collected and matched to the line segment according to a suitable method like that shown perhaps in the Viae Novae algorithm.

[0027] At least one probe trace is provided from any suitable source. Often, a plurality of traces will be provided in connection with ongoing data collection efforts or perhaps acquired in bulk from a particular source of data. The traces are shown in Figure 3 as a database or table 44. Each probe trace may include a defined creation date. The creation date is typically the time- stamped date on which the trace was generated, although another creation date may be assigned. A trace weight value is assigned to the trace in relation to its creation date. Most often, the assigned trace weight value will be 1 for every trace irrespective of its creation date, although another weight value could be used.

[0028] A matching step is depicted at 46. In this step, each trace is matched to a particular line segment in the generated network. Any suitable algorithm may be used for the step, including both incremental and non-incremental approaches like those described above. In some cases, the algorithm may be unable to map a trace to an existing line segment, in which case a new line segment is created (or an existing line segment is extended).

[0029] Next, the matched line segments are repositioned (or attributes thereof altered) at step 48. According to this invention, a recalculation date is established. This recalculation date may corresponds to (or closely precedes) the real time date at which the repositioning step 48 is carried out, although substantially varied recalculation dates may be established for particular purposes such as merging two line sengments.

[0030] The invention includes the step of adjusting the value of at least one of the line weight and trace weight. This adjustment may be a function of frequency of trace data or as a function of the time span between the recalculation date and one of the line segment specified date and trace creation date. This adjusting step is represented by either one or both function blocks 50, 52 shown in Figure 3. The adjusting step carried out in function blocks 50, 52 may be the same or different, as described further below. In one embodiment, the adjusting step may comprise reducing the line weight value in function block 50 as a function of the time span between the recalculation date and the line segment, or as a function of the time span between the recalculation date and the trace creation date. Alternatively, the adjusting step may comprise reducing the trace weight value in function block 52 as a function of the time span between the recalculation date and the trace creation date.

[0031] Those of skill in the art will understand that, in general, a network generation method does not in all cases need the map matching step 46. Non-incremental network generation methods usually do not use a map matching step 46. The map matching step 46 is primarily useful for the incremental network generation. Further still, non-incremental network generation methods usually do not use a line weight. Often, they only use the weights of the traces. On the other hand such methods may compute line weights e.g. to equip the generated network with an additional line attribute.

[0032] The present invention contemplates the use of the trace data 44 not only to generate, extend or refine the digital vector map 30 but to evaluate possible maneuvers 36 at a junction 33 or shape point 32. Therefore the present invention further includes a step of weight calculation for maneuvers 54. The present invention contemplates a variety of weighting methodologies as will be discussed in detail below. In the most basic scenario, each time a trace path follows a particular maneuver route, such as (q,r,u) the weight for that particular maneuver is increased by one. The result will be a set of weighted maneuvers at a particular junction such as junction r. This produces a plurality of weighted maneuvers 56 which may be utilized in a variety of fashions. In one intended use, the plurality of weighted maneuvers is utilized to calculate turn restrictions 58. In the example is shown in Figure 2. If at junction point r there are the following weighted maneuvers:

[0033] From this set of weighted maneuvers it can easily be determined that a vehicle approaching junction r from shape point t cannot make a right hand turn but must either proceed straight to shape point u or turn left to shape point q. This is shown by 48 traces indicating a (t,r,q) maneuver and 56 traces indicating a (t,r,u) maneuver while only a single trace indicating a (t,r,s) maneuver. Additionally, this may also be utilized to determine traffic flow direction 60 as it can easily be understood from the aforementioned weight table that traffic moves in a single direction from shape point t to shape point u. This is indicated by the 56 traces indicating a (t,r,u) maneuver and 0 traces indicating any of the (q,r,t), (u,r,t) or (s,r,t) maneuvers. It should be noted that there will likely be some weights in contradiction to actual turn restrictions such as the listed (t,r,s) with a weight of 1. These may be due to aberrant drivers, emergency vehicles, temporary road work etc. The present invention contemplates comparing the weight of a particular maneuver such as (t,r,s) =1 against the total weight of the junction r = 209 to decide if a maneuver is prohibited or allowed. This removes the low weighted maneuvers due to aberrant, illegal, or emergency activity. It is further contemplated that the weight of a particular maneuver used in conjunction with the weight of the junction could be used in combination with a defined range wherein a maneuver decision is not possible. In such a situation, turn by turn directions would select utilize maneuvers for which a decision is available.

[0034] It is further possible to utilized the weighted maneuvers 56 to detect the presence of grade separated crossing 62. In this example for the junction r, the weighted maneuver table may look like the following:

In this case there are clearly no turning maneuvers available in the table at junction r which is indicative of a grade separated crossing (indicated by 0 traces in all the turning maneuvers). This may be additionally confirmed by way of comparing time stamps of intersecting probe traces such as (q,r,s) and (t,r,u) which if correspond are a further indication of a grade separated crossing. The calculation of turn restrictions 58, the calculation of traffic flow 60, and the determination of grade separated crossings 62 are utilized to generate a snapshot of the generated network 64. The snapshot 66 represents the generated network 42 with up to date maneuver information for a particular moment in time.

[0035] In addition to using the sum of the weights of the traces to weight the maneuvers, the contribution of a particular trace to the weight of a maneuver may be varied using other criteria. For example, it is possible that the contribution of a particular trace may depend on the accuracy of the individual trace points. In this case the weight values may be floating-point numbers. In one embodiment, the trace points may be classified into accuracy classes dependent on the reception conditions of the GPS satellites, or from the type of probe utilized, or from the trace source, or from other factors. This allows the weight contribution of an individual trace to a particular maneuver to be varied as opposed to simply adding them evenly.

[0036] It should be understood that the values listed for maneuver weight above are for illustrative example only. The present invention contemplates a variety of methodologies for calculating weighted maneuvers 56 at any junction 33 or shape point 32. In the simplest case, as described above, the weight may simply be the number of traces which passed the three consecutive shape points or junctions in according order. It is contemplated, the weights may also be time dependent to provide additional accuracy. If a particular maneuver at a junction becomes disallowed, such as no left hand turns, it would retain a significant weight for some time after the new turn restrictions were implemented. It would therefore be desirable to implement time dependent weights on the weighted maneuvers 56 in order reduce the weight of old data.

[0037] A first proposed method for the updating maneuver weight uses a defined maximal time period measured rearwardly from the recalculation date. This maximal time period is divided into a plurality of bins. Each bin represents a respective portion of time between the recalculation date and the maximal time period. Figure 4 graphically shows the effects of adjusting trace weights as a function of time-span from any given recalculation date (Week 0,00).

[0038] For example, one might define a maximal time period of 1 year preceding the recalculation date, and divide this period into 12 bins corresponding each to one month. If, in this example, the recalculation date is January 1 of a given year, the bins would correspond, respectively, to the twelve immediately preceding months of the previous year. Still further according to this method, each bin is assigned a factor. The traces (determined from database 44) are grouped relating to each possible maneuver at a given intersection. The traces relating to a particular maneuver are then each associated with a specified one of the bins corresponding to their trace creation date. For each maneuver categorized trace, its weight value is then calculated as a function of the factor assigned to the bin with which that particular trace is associated. All maneuver categorized trace older than the specified maximal time period get the weight 0. Effectively, this means that maneuver categorized traces that are not assigned to a bin are not used.

[0039] For each maneuver, a table or list is stored with entries for each bin. Each entry contains the number of traces in a particular time window. As described, each of the bins has a certain factor which depends from the time window. In order to compute a cumulative weight for each bin, the number of traces of each bin is multiplied with the assigned factor. The total weight for the particular manuever is then determined as the sum of the weights of the bins. [0040] In the simple method, the weight value for a particular manever is determined by simply adding the number of traces previously mapped to it (assuming that each trace weight value is I ). Thus, if a particular manuever has 295 historic traces mapped to it, then the manuever weight value for that manuever is 295. Taking a comparative example, assume that a maximal time period is defined as 10 weeks, and that a bin length of one week is deemed adequate, so that ten separate bins of time are established. Assume still further that all 295 historic traces mapped to the manuever have creation dates within the maximal time period of the 10 weeks preceding a recalculation date. Using the principles of this invention, the 295 traces are sorted according to the bins into which their respective creation dates correspond. The following Table 1 shows an example of the weights for each bin for a time period of 10 weeks and a bin length of one week:

[0041] TABLE 1

[0042] In this example, 12 traces of the week immediately preceding the recalculation date are placed in the first bin: 25 traces having creation dates falling within the second week preceding the recalculation date; and so on. Hence all 295 traces used to compute this line segment are accounted for within this hypothetical example. If one adds a new trace to the line segment which is 4 weeks old, the table is updated at the corresponding bin as shown below in Table 2:

[0043] TABLE 2

[0044] According to this one possible implementation method of the invention, rather than simply adding the weights of all the traces to achieve a maneuver weight value of 295 (or 296 in the example of Table 2), factors are assigned to each bin. If a linear progression is chosen, the factor for week 0 may be set at "1" and each successive week decreased by 0.1. Thus, the older the trace data the lower its factor. Any non-negative and non- increasing function can be used. In the case of a linear decreasing function, the following Table 3 offers an example of the factor assigned to eac h bin, and the resulting adjusted weight value of each bin:

[0045] TABLE 3

[0046] According to this implementation method, therefore, the weight or contribution of each bin to the maneuver weight value of the maneuver is a function of the number of traces in the bin and the factor applied to that bin. Summing the adjusted weights of the bins, a total adjusted weight value of 138.8 is computed for this particular manuever. This can be compared to the maneuver weight value of 295 achieved with the simple approach. In the same maimer, if a trace that is 4 weeks old is added, as in Table 2, then that newly added trace contributes only a weight value of 0.6. After the merging of the new trace to the manuever, the total weight value of the manuever will be 139.4 (instead of 296). Figure 5 is a simplified diagram illustrating an implementation scheme corresponding to Table 3.

[0047] When a new week begins (i.e., relative to the recalculation date), all of the tables are updated. As a result, all weight values in the table have to shift one week as in the example shown in Table 4. If no new traces are mapped to the maneuver in the new week, then the weight contribution of the week immediately preceding the recalculation date (i.e., week "0") is set to 0.

[0048] TABLE 4

o [0049] If the weight contribution is 0 for every bin, it means that the maneuver does not have any traces mapped to it for the entire maximal time period. This could mean that the particular maneuver not used during the whole time period, and that it should be added to a list of turn restrictions within the generated network 42.

[0050] The second alternative approach uses a half-life concept to determine a maneuver weight value 56, and is described here in cooperation with Figures 7 and 8. In comparison to the preceding technique, the half-life concept may be preferred in situations where many bins are used, due to the need to shift large numbers of bins after each time step. In order to avoid this disadvantage one can use the half- life concept. According to this method, a half- life time T is defined, if the maneuver weight value at any given time (i.e., a "specified date") is known, then it is possible to compute the weight at any other time (i.e., the "recalculation date^"). With this concept one has only two values to store: a maneuver weight ω ₀ at a certain timestamp (the specified date).

[005] J A half-life period T is determined which, for purposes of example, could be a month. This means that after a month a trace (and/or associated maneuver) will have only half of its initial weight. If the weight for a trace (and/or a associated maneuver) is ω ₀ at a certain time stamp (i.e., specified date), then the weight w(t) after a time t is computed with

[0052] For example, there is a maneuver with a weight value of 263.8 and a time stamp of the 2^nd of August 2009. If the recalculation date is the 17^th hf August 2009, the time difference is t=\ 5 days. Assuming a half-life 30 days, as of the recalculation date the maneuver has a weight value of:

[0053] Next, suppose there is a trace from the 18 of June, 2009 (its creation date) which matches to this maneuver. Furthermore, assume that the weight of the trace on the 18^th of June is I. The adjusted weight value of the trace as of a 17^th August 2009 recalculation date is

[00S4J Now, as the trace and the manuever share a common recalculation date, their weight values can be summed to compute a new, resulting manuever weight value. Thus, after merging this trace to the manuever, the updated manuever has a weight value of 186.78 and a new time stamp (specified date) of I7^th August 2009.

[0955] Figure 9 is a simplified diagram illustrating an implementation scheme according to this third alternative embodiment of the invention. Of course it is possible to compute with this formula the weight for each time stamp in the past as well as in the future. It is also possible that the weight depends from the accuracy and other values.

[0056] According to a still further alternative approach, as referenced in Figures 9 and 10, the trace weight values with their respective corresponding creation dates can be used to compute the maneuver weight value for a maneuver to which the traces have been matched. This approach allows use of any decay function to reduce the weight depending on the timestamp. For each trace the weight value is computed with the used decay function. The total weight of a manuever is then the sum of the weights of all used traces. This approach provides a flexible way to compute the maneuver weights.

[0057] This approach is distinguished from (he first two approaches by using the trace references to compute the weight for a maneuver. This approach allows the use of any function to reduce the weight depending on the time stamp. Figure 10 illustrates an example of this alternative decay function for the weight. Until one month a trace is accepted as a new trace; in this time it keeps the weight of 1. After one month until five months, the weight decreases linearly. After five months the weight of the trace is 0. Many alternative functions are possible in this approach. In general, these decay functions are non-increasing. However for special applications other decay functions are also possible. For instance if one wants to compute how a road network was at a certain time in the past, a bell shape curve can be used.

[0058] In order to compute the total weight for a maneuver, one uses the references to the traces from which the maneuver was generated. For each trace the weight is computed with the decay function described above. The total weight of the maneuver is then the sum of the weights of all used traces. Of course it is possible to consider again the accuracy of traces and other values.

[0059] This approach can be applied in any number of ways to compute maneuver weight values under given conditions. It is also possible to change the decay function during the network generation process. However this approach may be somewhat disadvantaged relative to the other proposed approaches due to its rather higher computation effort because access to the database to obtain the dates of all referenced traces may require a substantial amount of time.

[0060] Collectively therefore, the invention describes a method for maneuver detection from probe (trace) data in which time-dependent influences are used to adjust the weight values for the manuever weight and/or trace weight data. In addition to a simplistic incremental approach, three additional approaches are proposed in order to consider the time dependency of probe (trace) data in connection with network generation algorithms, both incremental and non- incremental, but these are provided by way of example and not to be considered exclusive approaches to implementing the principles of this invention. These methods may be used, for example, in conjunction with the Viae Novae algorithm. However these methods are also useful for many other data mining activities where time-depending data are used. Furthermore, in appropriate circumstances, these methods could be used in various combinations with one another. And still further, addition factors such as accuracy, resolution or device or software depending values can be considered in a weight adjustment model.

[0061] Those of skill in this field may envision various improvements and extensions of this technology. For example, with these techniques, it is foreseeable to incorporate the computation of several road attributes like average speed, road classification, altitude and slope values. Furthermore, confidence models can be developed from these concepts for the generated geometry as well as its attributes. By these techniques, network elements generated from old data can be kept up to date without overloading data storage or data processing resources,

[0062] It will be understood that the general concepts of this invention can be used to improve any digital (vector) map, not only roadway and pathway maps. For example, circuit diagrams, schematics, and other graphical representations that can be spatially associated within a coordinate system may benefit from the techniques of this invention. [0063] Those of skill in this field may envision various improvements and extensions of this technology. For example, with these techniques, it is foreseeable to incorporate the computation of several road attributes like average speed, road classification, altitude and slope values. Furthermore, confidence models and/or codes can be developed from these concepts for the generated geometry as well as its attributes. By these techniques, network elements generated from old data can be kept up to date without overloading data storage or data processing resources.

[0064] Elements and/or features of different example embodiments can be combined with each other and/or substituted for each other within the scope of the disclosure and appended claims.

[0065] Still further, any one of the above-described and other example features can be embodied in the form of an apparatus, method, system, computer program and computer program product. For example, any of the aforementioned methods may be embodied in the form of a system or device, including, but not limited to, any of the structure for performing the methodology illustrated in the drawings.

[0066] Whilst embodiments described in the foregoing detailed description refer to GPS, it should be noted that the navigation apparatus may utilise any kind of position sensing technology as an alternative to (or indeed in addition to) GPS. For example, the navigation apparatus may use other global navigation satellite based such as the European Galileo system. Equally, it is not limited to satellite based, but could readily function using ground based beacons or any other kind of system that enables the device to determine its geographic location.

[0067] The foregoing description of the invention is exemplary rather than limiting in nature. Variations and modifications to the disclosed embodiment may become apparent to those skilled in the art and fall within the scope of the invention. Accordingly, the scope of protection afforded to this invention is defined only by the following claims.

Claims

CLAIMS:

1. A method for maneuver calculation using traces derived from probe data, comprising the steps of:

providing a digital map configured to store a plurality of maneuvers, each of said maneuvers comprising at least three consecutive shape points or junctions spatially associated within a coordinate system, each of said plurality of maneuvers having a maneuver weight value;

providing at least one probe trace;

associating said at least one probe trace with one of said plurality of maneuvers; adjusting the value of the maneuver weight value of the maneuver associated with said at least one probe trace based on said at least one probe trace; and

using the adjusted maneuver weight value to update maneuvers available at an associated shape point or junction.

2. The method of claim 1, wherein said adjusted maneuver weight value is utilized to calculate turn restrictions at said individual shape point or junction.

3. The method of any of the preceding claims, wherein said adjusted maneuver weight value is utilized to calculate traffic direction at said individual shape point or junction.

4. The method of any of the preceding claims, wherein said adjusted maneuver weight value is utilized to detect grade separated crossings at said individual shape point or junction.

5. The method of any of the preceding claims, wherein the maneuver weight value comprises the number of probe traces passing said three consecutive shape points or junctions associated with that maneuver.

6. The method of any of the preceding claims, wherein the contribution of said at least one probe trace to a maneuver weight value is varied based on the accuracy of said at least one probe trace.

7. The method of any of the preceding claims, further comprising:

assigning a trace weight value to the probe trace in relation to a probe trace creation date; establishing a recalculation date;

adjusting the weight value of the trace as a function of the time span between the recalculation date and the trace creation date; and

using the adjusted trace weight value to adjust said maneuver weight value.

8. The method of claim 7, wherein said adjusting step includes defining a maximal time period, and assigning a null trace weight value if the time span between the recalculation date and the trace creation date exceeds the maximal time period.

9. The method of claim 8, further including the step of dividing the maximal time period into a plurality of bins, each bin representing a respective portion of time between the recalculation date and the maximal time period.

10. The method of claim 9, fiirther including the steps of assigning each bin a factor, associating the trace with a specified one of the bins corresponding to the trace creation date, and calculating the trace weight value as a function of the factor assigned to the bin associated with the trace, wherein the maneuver weight value is calculated as a sum of the weight of the bins.

1 1. A method for maneuver calculation using traces derived from probe data, comprising the steps of:

providing a digital map configured to store a plurality of line segments, shape points, and maneuvers, each of said maneuvers comprising at least three consecutive shape points or junctions spatially associated within a coordinate system, each of said plurality of maneuvers having a maneuver weight value; associating each line segment in the digital vector map with a line weight value;

providing at least one probe trace;

establishing map matching criteria;

adjusting said line weight value using matched probe data;

associating said at least one probe trace with one of said plurality of maneuvers;

adjusting the value of the maneuver weight value of the maneuver associated with said at least one probe trace based on said at least one probe trace: and

12. The method of claim 1 1, wherein said adjusted maneuver weight value is utilized to calculate turn restrictions at said individual shape point or junction.

13. The method of claim 1 1, wherein said adjusted maneuver weight value is utilized to calculate traffic direction at said individual shape point or junction.

14. The method of claim 1 1, wherein said adjusted maneuver weight value is utilized to detect grade separated crossings at said individual shape point or junction.

15. The method of any preceding claims, further comprising:

defining a half-life time (T) and an exponential decay function;

adjusting the weight value of the trace as a function of the half-life time (T) and the exponential decay function; and

using the adjusted trace weight value to adjust said maneuver weight value.