Determining representative vehicle routes through the road network

Abstract

This invention provides a method, processing system, and software for determining representative vehicle routes along a sequence of road sections in a road network. [Solution] For each road section of the road network, at least one section-specific route index is determined for each of the multiple vehicle traces. The section-specific route index is selected from a predetermined set of route indices for the road section, where each index corresponds to a different set of possible routes along the road section. This determines each sequence of route indices for each vehicle trace. The vehicle traces are divided into groups based on the similarity between the sequences of route indices. For each of the groups of vehicle traces, a representative vehicle route along the sequence for the road section is determined.

Description

Technical Field

[0001] The present invention relates to determining a representative vehicle route through a road network.

Background Art

[0002] Collecting and analyzing vehicle movement data helps autonomous vehicles drive more humanely, which can lead to safer interactions with other road traffic. It also helps assist human drivers when traveling through dangerous or complex road sections. Existing systems, such as Advanced Driver Assistance Systems (ADAS), can directly control vehicle movement or provide additional assistance information to the driver and are already installed in many vehicles.

[0003] The present disclosure aims to provide an improved method for determining a typical vehicle route that helps a vehicle move through a road network.

Summary of the Invention

[0004] As a first aspect, the present invention is a method for determining a representative vehicle route along a sequence of road segments of a road network, the method comprising: accessing a plurality of vehicle traces, each vehicle trace including a respective sequence of geographical positions of each vehicle as the vehicle passes through the sequence of road segments; for each of the road segments of the sequence of road segments, processing each of the vehicle traces to determine a respective segment-specific route index of at least one of the vehicle traces, the segment-specific route index being selected from a predetermined set of one or more route indexes of the road segment, and each segment-specific route index corresponding to a different set of possible routes along the road segment, thereby determining a respective sequence of route indexes for each of the plurality of vehicle traces; Determining a similarity-based partition of the multiple sequences of the root index, thereby dividing the multiple vehicle traces into multiple groups of vehicle traces, For each of the groups of vehicle traces, the sequence of geographical locations of one or more vehicle traces in the group is processed in order to determine a representative vehicle route of the group along the sequence of road sections. This provides a method that includes [something].

[0005] The applicant understands that using recorded vehicle traces to determine representative vehicle routes along a portion of a road network can provide useful information about typical vehicle behavior on the network. However, dividing a large number of vehicle traces based on the similarity between sequences of their respective geographic locations can be computationally very demanding. Instead, embodiments of the present invention divide vehicle traces into groups based on the similarity between corresponding sequences of route indices (e.g., sequences of integer values). Dividing based on route indices rather than geographic coordinates allows for more efficient determination of representative vehicle routes.

[0006] Some or all of the vehicle traces may include multiple geographical locations within a single road section of the sequence of road sections. Therefore, each vehicle trace may contain a relatively large amount of data for each road section (e.g., including several latitude and longitude coordinates). In contrast, each section-specific route index may be a single character or value (e.g., an integer) selected from a predetermined set of characters or values. In some embodiments, each set of one or more route indices may consist of a sequence of indices, e.g., a sequence of integers. For example, for each road section, each set of route indices may contain up to 30 members. Thus, a sequence of route indices can provide a very compact representation of each vehicle trace. This saves space and allows for more efficient subsequent processing.

[0007] In yet another aspect, the present invention provides software (for example, on a non-temporary computer-readable storage medium) that, when executed on a processing system, includes instructions causing the processing system to perform a method for determining a representative vehicle route along a sequence of road sections of a road network, wherein the method is Accessing multiple vehicle traces, wherein each vehicle trace includes a sequence of the geographical locations of each vehicle as it travels through the sequence of road sections. For each of the road sections in the sequence of road sections, processing each of the vehicle traces to determine at least one section-specific route index for each of the vehicle traces, wherein the section-specific route index is selected from a predetermined set of one or more route indices for the road section, and each section-specific route index corresponds to a different set of possible routes along the road section, thereby determining each sequence of route indices for each of the plurality of vehicle traces. Determining a similarity-based partition of the multiple sequences of the root index, thereby dividing the multiple vehicle traces into multiple groups of vehicle traces, For each of the groups of vehicle traces, the sequence of geographical locations of one or more vehicle traces in the group is processed in order to determine a representative vehicle route of the group along the sequence of road sections. Includes.

[0008] In a further embodiment, the present invention provides a processing system configured to determine a representative vehicle route along a sequence of road sections in a road network, wherein the processing system Access multiple vehicle traces, each vehicle trace including each sequence of the geographical locations of each vehicle as it passes through the sequence of road sections. For each of the road sections in the sequence of road sections, each of the vehicle traces is processed to determine at least one section-specific route index for each of the vehicle traces, the section-specific route index being selected from a predetermined set of one or more route indices for the road section, each section-specific route index corresponding to a different set of possible routes along the road section, thereby determining each of the sequence of route indices for each of the plurality of vehicle traces. Determine the similarity-based partitioning of the multiple sequences of the root index, thereby dividing the multiple vehicle traces into multiple groups of vehicle traces. For each of the groups of vehicle traces, the sequence of geographical locations of one or more vehicle traces of the group is processed in order to determine a representative vehicle route of the group along the sequence of road sections. It is structured in this way.

[0009] The processing system may be able to access the vehicle traces from its local memory, or it may be communicatively connected to a server system via a network (which may include the Internet), and may be configured to access the multiple vehicle traces by obtaining the vehicle traces from the server system via the network. The server system and / or the local memory may be configured to store historical vehicle traces generated using data collected from vehicles traveling through the road network. The server system may be physically separated from the processing system or located in the same location.

[0010] The processing system may be configured to store the determined representative vehicle routes in local or remote memory, such as a database. The processing system may be configured to communicate with the vehicle control system (e.g., a driver assistance system) of a vehicle traveling through the road network. The processing system may be further configured to transmit data, including the representative vehicle routes, to the vehicle control system of the vehicle. The processing system may be configured to transmit data to the vehicle control system periodically, for example, once daily, once monthly, or at other appropriate periodic intervals. Alternatively, the processing system may be configured to transmit data to the vehicle control system in response to the fulfillment of an update condition. The update condition may be fulfilled when the number of newly determined representative vehicle routes exceeds a threshold. Alternatively or additionally, the update condition may be fulfilled when a predetermined amount of time has elapsed since the last time the processing system transmitted data to the vehicle control system.

[0011] In another aspect, the present invention provides a vehicle control system for vehicles traveling through a road network, the vehicle control system is Access data containing one or more representative vehicle routes along a sequence of road sections in the aforementioned road network, Determine the geographical location of the aforementioned vehicle, Based on the determined geographical location of the vehicle, one of the representative vehicle routes is selected. The information derived from the selected representative vehicle route is used to control the movement of the vehicle or to communicate information to the driver of the vehicle. It is structured in this way.

[0012] The vehicle control system may be configured to access the data, which includes one or more representative vehicle routes, by receiving it from the processing system disclosed herein. The processing system may be located away from the vehicle, located inside the vehicle, or distributed between a remote location and a location inside the vehicle. The memory or storage device of the vehicle control system may store representative vehicle routes of the road network (e.g., those received from the processing system or those loaded manually). Accessing the data, which includes one or more representative vehicle routes through the road network, may include retrieving the data from a database that may be located inside the vehicle, the database which includes the representative vehicle routes of the road network.

[0013] The vehicle control system may determine the geographical location of the vehicle from sensor data output by one or more sensors inside or on the vehicle. The one or more sensors may include Global Navigation Satellite System (GNSS) sensors arranged to receive satellite positioning signals. In some embodiments, the one or more sensors may additionally include one or both of an accelerometer and a camera. The camera may be an outward-facing camera arranged to capture images of the vehicle's environment. Using multiple sensors can improve the accuracy of the determined geographical location of the vehicle.

[0014] The processing system and one or more such vehicle control systems can together form a vehicle management system.

[0015] In another aspect, the present invention provides a vehicle management system that combines the processing system disclosed herein with the vehicle control system disclosed herein, wherein the vehicle control system is The processing system receives the data, which includes the representative vehicle route along the sequence of road sections in the road network, Determine the geographical location of the aforementioned vehicle, Based on the determined geographical location of the vehicle, one of the representative vehicle routes is selected. The information derived from the selected representative vehicle route is used to control the movement of the vehicle or to communicate information to the driver of the vehicle. It is structured in this way.

[0016] The vehicle control system or the processing system may be configured to process representative vehicle routes or associated driving attribute data in order to identify one or more sections of the sequence of road sections as potentially hazardous.

[0017] In any aspect or embodiment disclosed herein, each sequence of geographical locations of each vehicle used to determine the representative route may be determined using one or more sensors in or on each of the respective vehicles. Some embodiments may include generating the multiple vehicle traces by one or more vehicles, while in other embodiments, these are generated before determining the representative vehicle route. The one or more sensors may include Global Navigation Satellite System (GNSS) sensors positioned to receive satellite positioning signals. In some embodiments, the one or more sensors may additionally include one or both of an accelerometer and a camera. The camera may be an outward-facing camera positioned to capture images of the vehicle's environment. Using multiple sensors can improve the accuracy of each vehicle trace.

[0018] In some embodiments, each geographical location in the vehicle trace is determined by one or more simultaneous sensor measurements (e.g., from a GNSS receiver). Furthermore, in some other embodiments, one or more of these geographical locations may be calculated, for example, by interpolation from other geographical locations in the sequence.

[0019] Conventional navigation systems based on the output from GNSS sensors can use GNSS sensor data to match the position of a vehicle to points on a route on a digital map, thereby determining the map-matched position of the vehicle. This map-matched position can replace the original position indicated by the sensor data when the navigation system uses the position. However, detailed information regarding the position of the vehicle (e.g., which lane the vehicle is traveling in) may not be fully grasped solely based on the map-matched position.

[0020] Some of the road segments of the sequence may be associated with only a single respective route index. Desirably, one or more of the road segments are associated with respective plural route indexes from which a segment-specific route index can be selected.

[0021] The segment-specific route index may represent the lateral position of each route along the road segment. The segment-specific route index may be a segment-specific lateral position index. The vehicle trace may be processed to determine a sequence of lateral displacement values of the vehicle trace. In a set of embodiments, for each vehicle trace, a respective lateral displacement value is determined for all or a subset of the geographical positions within the sequence of geographical positions. Each lateral displacement value may indicate the lateral (e.g., orthogonal) displacement of the geographical position with respect to a reference route along the sequence of road segments.

[0022] The applicant has recognized that determining the direction displacement of the vehicle with respect to a reference route along the sequence of road segments can efficiently capture information useful for dividing the plurality of vehicle traces into groups. Thus, representing the position of the vehicle as a sequence of one or more lateral displacement values can provide a more efficient input to the step of determining a representative vehicle route, compared to, for example, directly dividing the sequence of geographical positions.

[0023] The reference route along the sequence of road sections may be determined before performing the method. The reference route may be stored, for example, with digital map data associated with the road network and retrieved by the processing system. However, in some embodiments, the reference route is determined by processing a set of vehicle traces, each vehicle trace containing a sequence of the geographical locations of each vehicle as it passes through the sequence of road sections. This set of vehicle traces may be all or a subset of the plurality of vehicle traces described above. In some embodiments, determining the reference route includes selecting one of the vehicle traces from the plurality of vehicle traces. The vehicle trace may be selected arbitrarily (e.g., randomly) or may represent an average vehicle trace. In some embodiments, the reference route may be determined by processing the plurality of vehicle traces to calculate an average vehicle trace. The reference route may be the average vehicle trace or one that follows it.

[0024] Calculating the average vehicle trace is Selecting a first vehicle trace from the aforementioned multiple vehicle traces, and (for example, arbitrarily), Determining a series of orthogonal lines at points (which may be spaced regularly or irregularly) along the first vehicle trace, For each orthogonal line, The intersection points of the orthogonal lines and each of the vehicle traces of the plurality of vehicle traces are determined, thereby determining a set of intersection points of the orthogonal lines. Calculating the average intersection point from the aforementioned set of intersection points, Constructing an average vehicle trace that includes the sequence of the respective average intersection points of each orthogonal line, It may include.

[0025] Calculating the average intersection may involve determining the median or average intersection from the aforementioned set of intersections.

[0026] For each of the plurality of vehicle traces, the lateral displacement value relative to the calculated average vehicle trace can be determined for all or a subset of the geographical locations of the vehicle traces.

[0027] The applicant found that using a reference route determined by processing actual vehicle traces may be advantageous compared to using a reference route determined according to topology data associated with a digital map (although this may be done in some embodiments). The applicant found that the lateral displacement values ​​subsequently determined relative to the reference route may better represent the variation in the vehicle's position as it travels along the route through the road network.

[0028] In some embodiments, the method includes processing each of the sequence of lateral displacement values ​​to determine at least one section-specific route index for each vehicle trace for each road section in the sequence of road sections. Thus, in these embodiments, the assignment of each section-specific route index depends on the lateral displacement values ​​determined for each vehicle trace along the road sections. This makes it possible to identify groups of vehicle traces that follow similar routes along the road sections with respect to lateral displacement values.

[0029] A route index specific to each section may be a lane index specific to that section. It may correspond to traffic lanes, for example, the section of the road intended to separate traffic lanes. Such traffic lanes may be demarcated by markings on the road surface. Thus, the predetermined set of route indices for each road section may be the respective set of lane indices corresponding to the lanes or lane markings of each road section. However, the applicant understands that in practice, the flow of traffic through a road section does not have to be limited to being contained within the lanes demarcated by markings on the road surface. Nevertheless, traffic may exhibit predictable behavior by generally following one of a typical set of routes along the road section. For example, in some road sections, vehicles may typically turn corners at certain junctions rather than precisely following the marked lanes on the road. In other scenarios, there may be road sections without marked lanes, yet traffic may still follow a predictable route—for example, toll booth exits, paved roads, or asphalt roads. Therefore, in some embodiments, the route index specific to each section may correspond to a route that is regularly followed, or expected to be regularly followed, by vehicles passing through the road section. In this sense, the route index specific to each section may be considered as the lane index of the corresponding traffic lane, but it does not need to correspond to a lane marked on the road surface.

[0030] Routes regularly followed by vehicles passing through each road section can be identified using historical vehicle movement data obtained from or about vehicles passing through the road section. Thus, in some embodiments, for each road section, the set of section-specific route indices is determined by processing a plurality of vehicle traces, each vehicle trace containing a sequence of the geographical locations of each vehicle as it passes through the road section. These plurality of vehicle traces may be the same as, overlap with, or different from the plurality of vehicle traces used to determine the representative vehicle route along the route through the road network. The plurality of vehicle traces may include a subset of the vehicle traces used to determine the representative vehicle route, or may consist of a plurality of entirely different vehicle traces.

[0031] Each section-specific route index may correspond to a different cluster of possible routes that can be taken along the road section. The possible routes may be spatially clustered. Each possible route may correspond to a historical vehicle trace. In some embodiments, for each road section in the sequence of road sections, each set of section-specific route indices may be determined by grouping vehicle traces into one or more road section clusters and assigning a different section-specific route index to each road section cluster. In embodiments where each sequence of lateral displacement values ​​is determined for each vehicle trace, the vehicle traces through the road sections may be grouped according to the similarity between the sequences of lateral displacement values.

[0032] Determining the set of route indices by grouping actual vehicle traces for each road section may have the advantage of allowing the number of section-specific route indices for each road section to vary depending on the actual vehicle movement through each road section. For example, it is possible to avoid artificially constraining the number of section-specific route indices for each road section to the number of known lanes marked on the road. This may result in a better representation of vehicle movement through each road section. Adjacent road sections may turn out to have the same number of section-specific route indices, or they may turn out to have different numbers of section-specific route indices as vehicle behavior and / or lane markings change along the sequence of road sections.

[0033] A road section may be defined in any suitable way. In at least some embodiments, each road section is determined such that the majority of vehicles follow a fixed route (i.e., corresponding to each route index) along the entire road section. Each road section may have a length such that most vehicles remain in a fixed lane for the duration of the road section. However, some vehicles may still change their route (e.g., atypical lane changes) during the duration of the road section. Such vehicle traces may be represented by multiple route indices of the road section. In some embodiments, each vehicle trace may be assigned a best-matching section-specific route index. This may help stabilize the data and / or provide a more compact representation. Thus, in some embodiments, for each road section, exactly one section-specific route index is determined for each vehicle trace, and as a result, the method determines, for each of the multiple vehicle traces, each sequence of route indices corresponding to the sequence of road sections.

[0034] In some embodiments where each of a plurality of vehicle traces is processed to determine each sequence of lateral displacement values, determining the sequence of route indices is done for each of the plurality of road sections, To generate a road section dataset containing some or all of the lateral displacement values ​​from each sequence of lateral displacement values ​​along the road section, Modeling the distribution of lateral displacement values ​​within the aforementioned road section dataset, Identifying one or more peaks within the aforementioned distribution, Each lateral displacement value is assigned to a cluster associated with the peak closest to the lateral displacement value among the one or more peaks, This involves assigning a different interval-specific root index to each cluster of lateral displacement values, thereby assigning an interval-specific root index to each lateral displacement value. For each vehicle trace passing through the aforementioned road section, a single section-specific route index is determined for the vehicle trace, depending on the section-specific route index assigned to the lateral displacement value of each sequence of lateral displacement values ​​determined for the vehicle trace, Includes.

[0035] A section-specific route index for each vehicle trace can be determined from the sequence of lateral displacement values ​​as the modal section-specific route index (i.e., the route index that occurs most frequently for the vehicle trace along the road section). This ensures that each vehicle trace through the road section is associated with the single section-specific route index that best represents the path of the vehicle through the road section.

[0036] The applicant understands that performing the step of reducing the sequence of geographical locations for each vehicle trace to a simplified sequence of route indices may improve the outcome of the division of the multiple vehicle traces into multiple groups of vehicle traces. This simplification of the input data may lead to faster convergence of the algorithm used to divide the multiple vehicle traces. It may also help reduce the impact of outliers in the data that may have been caused by errors in the original sensor data used to determine the sequence of geographical locations. Similarly, it may help reduce the impact of errors in the digital map data.

[0037] Dividing multiple vehicle traces into multiple groups using similarity-based partitioning may involve inputting the sequence of root indices for each vehicle trace into a clustering algorithm. The clustering algorithm may include a k-means clustering algorithm. Typically, multiple vehicle traces are divided into multiple groups of vehicle traces, but the processing system may be configured such that for some sequences of road sections, the partitioning results in only a single group of vehicle traces.

[0038] Determining a representative vehicle route for each of the aforementioned groups of vehicle traces may include calculating the average vehicle trace by processing the sequence of geographical locations from all of the vehicle traces in the group. Each representative vehicle route may include a sequence of geographical locations (e.g., longitude-latitude coordinates). Determining a representative vehicle route for a group may include calculating a sequence of average geographical locations at points along the sequence of road sections, and may include using spline interpolation of the sequence of average geographical locations to determine a geometric line through the sequence of road sections. Each average geographical location may include the lateral (e.g., perpendicular to the direction of travel) geographical location of the average or median of the vehicle traces for each of the respective groups at each point along the route.

[0039] The representative vehicle paths may represent only the geographical shape (geometry) of the respective average vehicle traces of each group, but in some embodiments they may also represent or be associated with vehicle attribute data. Some embodiments may include determining a sequence of driving attribute values ​​for each of the groups of vehicle traces, each driving attribute value associated with each point along the sequence of road sections. Each point along the sequence of road sections may be located in a different road section. In some embodiments, a driving attribute value is determined for each road section of the sequence of road sections. The driving attribute values ​​may include, but are not limited to, values ​​associated with vehicle movement, such as vehicle speed and / or vehicle acceleration. Furthermore or alternatively, the driving attribute values ​​may include the number or percentage of vehicle traces associated with each group of vehicle traces at each point along the sequence of road sections. In some embodiments, each driving attribute value in the sequence of driving attribute values ​​may be a measure of vehicle speed. For example, each measure of vehicle speed may be calculated as the average (e.g., harmonic mean) of the vehicle speeds associated with each vehicle trace at a point along the sequence of road sections. Each measurement of vehicle speed may further, or instead, include a free-flow speed, for example, a speed measurement calculated as the average of a subset of vehicle speeds associated with each vehicle trace at a point along the sequence of road sections. The subset of vehicle speeds may be selected as the top n percentiles of vehicle speeds at a point along the sequence of road sections, for example, the top 25th percentile, 50th percentile, or 75th percentile.

[0040] The representative vehicle route or associated driving attribute data may be further processed to identify one or more sections of the sequence of road sections as potentially hazardous. If a vehicle is approaching or in a section of the road network identified as potentially hazardous, a warning may be issued to the driver of the vehicle. For example, in embodiments where the determined representative vehicle route is transmitted to the vehicle's driver assistance system, if the vehicle is approaching or in a section of the road network identified as potentially hazardous, the driver assistance system may issue a warning to the driver of the vehicle. In some embodiments, the section of the road network identified as potentially hazardous may be transmitted to the vehicle's driver assistance system without additional transmission of the determined representative vehicle route. Visual, auditory, or tactile warnings may be issued individually or in any combination.

[0041] The vehicle control system (e.g., a driver assistance system) may be configured to identify one of the plurality of representative driving paths as matching the vehicle's path (e.g., as the one that best fits the path the vehicle is currently following). The vehicle may be autonomous (i.e., unmanned) or it may have a driver. In either case, the vehicle's movement may be controlled, influenced, or proposed, at least in part, by the vehicle control system, relying on or referring to driving attribute information associated with the identified representative vehicle path. For example, to determine a typical speed for a vehicle moving along this representative driving path, the vehicle control system may control the vehicle's movement to conform to the surrounding traffic and / or environment by, for example, comparing the vehicle's position to the identified representative driving path and using the attribute information associated with the representative driving path. In embodiments where the vehicle control system is implemented in an autonomous vehicle, the vehicle control system may directly control the vehicle's speed to match a typical speed determined, for example, using the identified representative driving path.

[0042] The sequence of road sections may represent a section of the road network without junctions, or it may span one or more junctions. The sequence may represent traffic making specific choices at each junction. In addition to being configured to determine multiple representative vehicle routes along a sequence of road sections in the road network, the processing system may be configured to determine a single representative vehicle route along several sequences of road sections in the road network (for example, along a single-lane highway where all traffic follows the same route). It may also be configured to determine representative vehicle routes corresponding to two or more partially overlapping sequences of road sections in the road network, such as when the road network branches or merges at junctions.

[0043] The processing system may be located in a single location (e.g., including a single server or workstation) or it may be a distributed system. The server that stores the vehicle trace may be a single server or it may be a distributed server system.

[0044] Any of the processing steps disclosed herein may be performed by software or hardware, or a combination of software and hardware, running on one or more processors. The information may be communicated by wired and / or wireless channels.

[0045] Information such as vehicle traces, route indices, and representative routes can be encoded in an appropriate manner. Data containing such information can be compressed, decompressed, encoded, or decoded during storage, transmission, or processing.

[0046] Features of any aspect or embodiment described herein may apply to any other aspect or embodiment described herein, where appropriate. Where different aspects or sets of embodiments are referred to, it should be understood that they are not necessarily different and may overlap. [Brief explanation of the drawing]

[0047] Specific preferred embodiments of the present invention will be described only by reference to the following accompanying drawings. [Figure 1] Figure 1 is a schematic diagram of an exemplary system embodying the present invention. [Figure 2] Figure 2 is a satellite image of a road network intersection with geometric lines representing different vehicle routes superimposed on it. [Figure 3] Figure 3 is a flowchart showing the method performed by the vehicle behavior processing system to determine a representative vehicle trace according to an embodiment of the present invention. [Figure 4] Figure 4 is a schematic diagram of a vehicle trace along a reference route corresponding to the centerline of a road on a digital map. [Figure 5] Figure 5 is a plot showing lateral displacement data generated by converting vehicle traces into lateral displacement values ​​using the road centerlines of a digital map as the reference route. [Figure 6] Figure 6 is a schematic diagram showing the original vehicle trace along the route section. [Figure 7] Figure 7 is a schematic diagram showing a series of orthogonal lines calculated along one of the lengths of the vehicle trace shown in Figure 6. [Figure 8] Figure 8 is a schematic diagram showing the average vehicle trace constructed using the orthogonal lines shown in Figure 7. [Figure 9] Figure 9 is a schematic diagram showing a close-up view of an example of a set of intersections determined between perpendicular lines and vehicle traces. [Figure 10] Figure 10 is a plot showing lateral displacement data generated using the average vehicle trace as the reference route. [Figure 11] Figure 11 is a plot of the lateral displacement data shown in Figure 10, divided into smaller intervals. [Figure 12]Figure 12 is a plot of isolated intervals of lateral displacement data extracted from the lateral displacement data shown in Figures 10 and 11. [Figure 13] Figure 13 is a plot showing the modeled distribution of the lateral displacement data intervals shown in Figure 12. [Figure 14] Figure 14 is a plot showing intervals of the lateral displacement data shown in Figure 12, labeled according to interval-specific indices assigned to the lateral displacement values. [Figure 15] Figure 15 is a distance-aligned chart showing the plot from Figure 10 over a modeled route index, illustrating two exemplary sequences of route indices determined for two vehicle traces according to an embodiment of the present invention. [Figure 16] Figure 16 is a plot depicting numerous vehicle traces along a sequence of road sections, represented as a sequence of route indices. [Figure 17] Figure 17 is a plot showing several groups of vehicle traces identified by clustering the sequence of route indices shown in Figure 16. [Figure 18] Figure 18 shows three exemplary vehicle routes overlaid on a section of the digital map. [Figure 19] Figure 19 shows a typical vehicle route section from Figure 18, along with the relevant speed data. [Modes for carrying out the invention]

[0048] Gaining insights into typical vehicle movement is beneficial for safely implementing autonomous vehicle driving systems or driver assistance systems for driver-controlled vehicles. Many modern vehicles are equipped with advanced driver assistance systems (ADAS) that provide additional information to the driver or directly control the vehicle's movement. ADAS can typically control the vehicle's speed and direction based on information such as speed limits and road layout. ADAS can implement this by directly controlling the vehicle's speed and direction, or by suggesting changes to the driver, for example, via a dashboard display. Systems that warn the driver and / or control the vehicle's movement (ADAS, whether autonomous or not) may be referred to here as vehicle control systems.

[0049] The applicant understands that it would be beneficial for a vehicle control system to have access to information about the typical movement of a vehicle along a stretch of road, in addition to topology data associated with digital maps. This means that autonomous vehicles can better mimic human drivers and drive more safely alongside other road users. The same information can also be useful for drivers who are driving a particular stretch of road for the first time.

[0050] The applicant understands that, in particular, a more accurate record of the routes a vehicle typically takes through a section of the road network could be useful for providing information to, for example, vehicle control systems. There may be several scenarios where the actual route a vehicle takes deviates from the expected route based on a topology map. For example, when a vehicle passes through a junction, the driver may take a route that does not precisely follow the topology shape (e.g., centerline) of the corresponding digital road map. In another example, there may be multiple different routes that a vehicle typically takes along a particular stretch of road. For example, a vehicle approaching a junction may travel in one lane or road position across a stretch of road before changing lanes before an exit ramp. However, other vehicles continuing on the same road may remain in the same lane. Providing vehicle control systems with information on how traffic typically changes lanes before an exit ramp can improve the safety of vehicles traveling on the road, for example, by enabling the system to automatically slow down the vehicle if it follows a route that matches the route the vehicle typically takes before the driver enters the exit ramp.

[0051] Typical driver behavior can be determined by processing historical data, including GNSS (e.g., GPS) traces of routes taken by numerous vehicles along road stretches. However, mapping vehicle movement across a wide geographical area presents significant challenges due to the large volume of data involved. The described methods and systems can provide an efficient way to determine typical vehicle routes and provide them to vehicle control systems.

[0052] High-resolution location data is required to collect data with the necessary resolution to map different routes taken by vehicles on the same road stretch. Vehicles used to collect appropriate motion data may include Global Positioning System (GPS) sensors positioned to receive satellite signals. In addition to GPS, other sensors can be used to generate more accurate vehicle traces. For example, a vehicle may include outward-facing cameras positioned to capture images or artifacts such as road signs, lanes, bridges, and other recognizable features at known locations. The position of these features relative to the vehicle can be used in conjunction with data from GPS sensors to determine a more accurate vehicle position. A sequence of geographical locations (i.e., a time series) as a vehicle moves along a sequence of road stretches in a road network can be stored as a vehicle trace. These vehicle traces may be associated with attribute data that describes, for example, the vehicle's speed and / or acceleration as it moves along the route. Data from GPS sensors may include vehicle speed, or vehicle speed may be derived from GPS sensor data. Vehicle acceleration may be derived from GPS sensor data using, for example, vehicle speed or vehicle position.

[0053] In some embodiments of the present invention, historical data containing numerous vehicle traces may be processed to generate a set of representative vehicle routes that characterize typical vehicle behavior through a road network. A vehicle control system can access a database of these representative vehicle routes. When a vehicle is traveling on a road, the vehicle control system can determine the section of road it is currently traveling on and retrieve one or more corresponding representative routes from a database containing numerous determined representative routes within the road network. Such a database may here be referred to as a behavior map database.

[0054] Figure 1 is a schematic diagram of a system 100 suitable for carrying out the embodiments of the present invention described herein. The vehicle 110 comprises a vehicle positioning system 112 and a vehicle control system 114. The vehicle control system 114 is communicated with a behavior map server system 130, and the behavior map server system 130 is communicated with a vehicle trace server system 150. In Figure 1, the behavior map server system 130 is shown separately from the vehicle trace server system 150, but they can also be located in the same place.

[0055] The behavioral map server system 130 includes a central behavioral map database 132 containing a set of representative routes that characterize typical vehicle movement through the road network. The behavioral map server system 130 also includes a vehicle behavior processing system 134 (embodying the processing system disclosed herein) configured to generate data to be included in the central behavioral map database 132. The vehicle behavior processing system 134 includes one or more processors and memory for storing software to be executed by one or more processors to perform the processing operations described herein. The vehicle trace server system 150 includes a vehicle trace database 152 containing numerous historical vehicle traces through the road network. The vehicle behavior processing system 134 is communicated with the vehicle trace server system 150 to access the vehicle trace database 152. The vehicle 110 and the server systems 150, 130 may each include communication circuits for exchanging data over wired and / or wireless networks (e.g., cellular radio to and from the vehicle 110).

[0056] The vehicle positioning system 112 includes two sensors: an outward-facing camera 116 and a GPS sensor 118. These two sensors 116 and 118 output data used by the vehicle positioning system 112 to determine the precise geographical location of the vehicle 110. In other examples (not shown), one or more additional sensors, such as accelerometers, may be present on the vehicle 110. By combining data from multiple sensors, the vehicle positioning system 112 can determine the vehicle 110's location more precisely.

[0057] In the example shown in Figure 1, the vehicle positioning system 112 is also connected to the vehicle trace server system 150. The vehicle 110 is configured to transmit location information determined by the vehicle positioning system 112 to the vehicle trace server system 150. This location information is stored in the vehicle trace database 152 and can be used for future analysis of the vehicle's movement patterns. The location information can be transmitted to the vehicle trace server system 150 intermittently and / or periodically, for example, daily, weekly, or monthly, or when the total amount of historical location information locally stored in the vehicle reaches a threshold. However, in other examples, the vehicle 110 does not necessarily need to upload data to the vehicle trace database 152, but can instead generate data using data from other sources.

[0058] The vehicle control system 114 comprises a processing system and a memory that stores a digital map 120 of the road network and local behavior map data 122. The vehicle control system 114 is communicated with the vehicle positioning system 112. The vehicle control system 114 uses the position of the vehicle 110 determined by the vehicle positioning system 112 and the digital map 120 to determine the position of the vehicle 110 within the road network and obtains information about road attributes such as road shape and speed limits from the digital map 120. The local behavior map data 122 includes a copy of data from a behavior map database 132 related to the area where the vehicle is located. This typically includes one or more representative vehicle paths along the sequence of road sections the vehicle is currently traversing. The vehicle control system 114 can obtain this typical vehicle movement information from the local behavior map data 122 and use it to control one or more actions of the vehicle 110, such as the vehicle's trajectory and / or speed.

[0059] In the example shown in Figure 1, vehicle 110 is shown to be in communication with the behavior map server 130 and the vehicle trace server system 150, but this may not be the case in other examples. Alternatively, vehicle 110 may have behavior map data 122 preloaded for offline use and / or may not be configured to send historical location information to the vehicle trace server system 150 for future analysis.

[0060] The operation of the behavioral map processing system 134 for generating the central behavioral map database 132 will be described in detail below.

[0061] To create the Central Traffic Map Database 132, a digital map of the road network, including a network of connected road sections, is analyzed, and sections of the road network are identified for analysis of typical vehicle movement. These sections may be simple stretches of road or may include complex junctions. Intersections and other complex features within the road network can trigger a variety of vehicle behaviors, including different typical driving routes. For example, as a vehicle approaches an intersection, a vehicle attempting a first turn may drive differently than a vehicle attempting a second turn.

[0062] After sections of the road network are identified for further analysis, the digital map data associated with each section is processed to roughly determine all viable routes through it. An example of an intersection (roundabout) is shown in Figure 2, where geometric lines representing different vehicle routes are superimposed. The upstream and downstream road sections of the intersection are added to the road section that is part of the intersection itself, forming a sequence of straight road sections leading to, through, and away from the intersection. One example of these identified routes through the intersection is highlighted by a solid line in Figure 2.

[0063] The same process is repeated for all sections of the road network of interest. Depending on the memory constraints used to store information about the central behavioral map database 132, the vehicle behavior processing system 134 may be configured to analyze only a subset of road sections within the road network for inclusion in the central behavioral map database 132. This subset may be identified using a list of known complex intersections, or the most complex road sections may be automatically identified by the vehicle behavior processing system 134.

[0064] Once a section of the road network is identified for analysis, historical vehicle traces collected for that section of road are retrieved from the vehicle trace database 152. These may be collected from a large number of different vehicles over time, forming a large dataset that can determine typical driving behavior. Each vehicle trace includes a sequence of the vehicle's geographical location (e.g., latitude and longitude coordinates) as the vehicle traveled along the sequence of road sections (e.g., a series with timestamps).

[0065] Figure 3 shows a flowchart illustrating a method 300 performed by the vehicle behavior processing system 134 to determine a representative vehicle trace of a route through the road network from the acquired historical vehicle traces.

[0066] In step 310, the vehicle behavior processing system 134 determines a reference route along the sequence of road sections. In step 320, for each vehicle trace along the route, all or a subset of the geographical locations of the trace are converted into positive and negative lateral displacement values ​​that represent the lateral displacement of the geographical location from the reference route (i.e., along lines perpendicular to the reference route or perpendicular to the vehicle trace and intersecting the geographical location).

[0067] The first approach is to use the centerline of the road shape on the digital map as the reference route. While this information is readily available, the applicant recognizes that this can cause confusion in the calculation of the resulting lateral displacement values. The actual path of a vehicle traveling on a road network typically deviates from the topological road shape associated with the digital map. This topological shape is usually simplified, and in some cases, acute angles that do not actually exist on the road may appear on the digital map, for example, due to resolution limitations where the road curves. An example of this discrepancy is shown in Figure 4. Line 410 on the digital map includes an acute corner 412. However, the vehicle trace 420 does not follow the same acute angle.

[0068] Figure 5 shows an example of lateral displacement data calculated using line 410 on the digital map as the reference route. As shown in Figure 5, the lateral displacement data deviates significantly from the topological centerline of the road at positions 510 and 520. This artificially bulges the maximum and minimum boundaries of the lateral displacement data, creating unnatural discontinuities that can cause problems in subsequent processing.

[0069] While this approach may be used in several embodiments, a second approach is preferred, which involves processing the vehicle trace to determine a reference route. This provides a reference route that better represents the shape of the path actually followed by the vehicle.

[0070] Several different methods can be used to determine a similarly effective reference route for a stretch of road. For example, one can randomly select vehicle traces from multiple vehicle traces, select a representative vehicle trace from multiple vehicle traces, or calculate the average trace.

[0071] A method for constructing a reference route along a route used in several embodiments is described with reference to Figures 6, 7, and 8. Figure 6 shows the original vehicle trace along a section of the route. A random trace 710 is selected, and a series of orthogonal lines 712 are calculated at regular intervals along the length of the random trace 710, as shown in Figure 7. For each orthogonal line 712, the respective intersections of the orthogonal line 712 and each vehicle trace are determined. The average (average displacement) intersection of each orthogonal line is selected, and an average vehicle trace 810 is constructed by fitting a curve through these average intersections from each orthogonal line, as shown in Figure 8 (using an appropriate curve fitting algorithm).

[0072] Figure 9 shows an enlarged view of an example of a set of intersections 910 determined between the orthogonal line 912 and the vehicle trace 914. The mean intersection 916 can be determined as the mean or median intersection from the set of intersections 910.

[0073] Figure 10 shows the results of calculating lateral displacement data using a reference route corresponding to the determined average vehicle trace. As can be seen from Figure 10, the variance of the lateral displacement data has been significantly reduced compared to the variance of the lateral displacement data shown in Figure 5 for the same input vehicle trace.

[0074] Referring again to Figure 3, in step 330, the lateral displacement data is divided into segments, and each data segment is analyzed individually. Each segment of lateral displacement data corresponds to one road segment in an identified sequence of road segments across the road network. Figure 11 shows an example of dividing the lateral displacement data shown in Figure 10 into smaller segments 1100a-1100n. As can be seen from the scale on the lower axis of Figure 11, road segments can be on the order of 50m-100m, but they can also be shorter or longer. In the example shown, the size of the road segments varies along the road segments (stretches) depending on the changes in route complexity. Figure 12 shows the extracted segment of lateral displacement data 1100i alone.

[0075] Referring again to Figure 3, the following two steps 340 and 350 are performed for each section of lateral displacement data. As a result of these two steps, each vehicle trace through each road section is described using a single section-specific route index.

[0076] Section-specific route indices are used to compactly represent the route taken by a vehicle through a road section in a single integer value. For example, traffic traveling through a road section may be traveling in one of three different lanes, and the section-specific route index assigned to a vehicle trace can describe which lane the vehicle was traveling in for the duration or for most of the time through the road section. These traffic lanes may be demarcated by markings on the road surface. However, the applicant understands that the flow of traffic through a road section may not actually be limited to the lanes demarcated by markings on the road surface. Traffic may nevertheless exhibit predictable behavior by following a typical set of routes along the road section. In this sense, section-specific route indices can be considered lane indices even if they do not necessarily correspond to traffic lanes marked on the road surface.

[0077] To determine how many of these de facto lanes there are through each road section, step 340 processes the lateral displacement data to determine a set of section-specific route indices for the road section. These route indices are determined by clustering similar lateral displacement values ​​within the lateral displacement data for the road section. Figures 13 and 14 illustrate the steps of this process for the section of lateral displacement data 1100i shown in Figure 12. As shown in Figure 13, the lateral displacement data for the road section is first modeled as a distribution 1310. After testing various algorithms, the applicant found that kernel density estimation provides the most robust results for modeling the distribution. This distribution 1310 is analyzed to detect peaks in the data. In the example shown in Figure 13, three peaks, the first peak 1312, the second peak 1314, and the third peak 1316, are identified. The peak detection algorithm is tuned with various pre-set parameters to identify only the most important peaks in the data. These parameters can be implemented as configurable parameters that can be adjusted according to the level of detail required for the behavioral map data.

[0078] By dynamically identifying several significant peaks within the data, the number of traffic lanes in each section can be determined. This effectively avoids artificially constraining the process to the expected number of traffic lanes in each road section, which may not adequately represent typical driving behavior as mentioned earlier (e.g., based on observed road markings).

[0079] Next, each lateral displacement value is assigned to an interval-specific index, with each index corresponding to its respective distribution peak. As shown in Figure 14, the first group of lateral displacement values ​​is assigned to the first index 0, associated with the first peak 1312. The second group of lateral displacement values ​​is assigned to the second index 1, associated with the second peak 1314, and the third group of lateral displacement values ​​is assigned to the third index 2, associated with the third peak 1316.

[0080] In step 350, a section-specific route index is determined for each vehicle trace passing through the road section. To this end, individual lateral displacement values ​​are mapped to the vehicle trace to which they belong, and for each vehicle trace, the index that best fits is selected based on the index to which the majority of its individual lateral displacement values ​​are assigned. This process makes it possible to identify groups of vehicle traces that follow similar routes through the road section.

[0081] Returning to Figure 3, once each vehicle trace is labeled with a section-specific index for each road section, the sections are combined in step 360, determining the respective sequence of route indices for each vehicle trace. As a result of steps 310-360, we can see that the sequence of geographical locations for each vehicle trace has been reduced to a simplified sequence of route indices. Figure 15 shows an example of the sequence of route indices for two traces generated using the process described above. For example, trace 1 starts in the first lane of a two-lane road, labeled with index 0, for one section. It then moves in the second lane of a two-lane road, labeled with index 1, for one section. It then moves in the third lane of a three-lane road, labeled with index 2, for nine sections. Finally, it moves in the second lane of a two-lane road, labeled with index 1, for six sections, until the end of the route sequence. Compared to the lateral displacement data shown in Figure 10 (duplicated in the upper half of Figure 15 for easier reference), we can see that the amount of data used to describe the movement of this vehicle on the road has decreased significantly (i.e., to the sequence {0,1,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1}).

[0082] Figure 16 shows numerous vehicle traces along a section of road, represented as a sequence of route indices. As shown in Figure 16, there are initially two identified traffic lanes at the beginning of the section. Therefore, all vehicle traces are assigned to index 0 or index 1. In the middle section of the route, additional traffic lanes appear, and some traces from index 0 and index 1 move to index 2. By the end of the route, the traffic flow returns to only two lanes.

[0083] In step 370, vehicle traces are grouped based on the similarity between sequences of section indices. The applicant recognizes that simplifying the input data for this grouping process significantly reduces the processing load required, while still allowing groups of vehicles to identify common routes they typically take.

[0084] The applicant found that clustering algorithms such as k-means clustering are effective in identifying common sequences of root indices shared by vehicles. The k-means clustering algorithm iterates through several different sets of clusters of vehicle traces, and the vehicle traces are clustered using sequences of root indices as input. A quality score is determined for each clustering, and the iteration of the clustering process that yields the highest score is selected to divide the vehicle traces into groups.

[0085] Figure 17 shows examples of different groups identified by a clustering process on the sequence of route indices shown in Figure 16. The first group 1710 includes vehicles that remained in lane 0 throughout the entire route. The second group 1720 includes vehicles that started in lane 0 and then moved to lane 1. The third group 1730 includes vehicles that started in lane 1 and remained in lane 1. The fourth group 1740 includes vehicles that started in lane 1, moved to lane 2, and then returned to lane 1.

[0086] Finally, returning to Figure 3, a representative vehicle route is determined for each group identified in step 380 and then stored in the central behavioral map database 132 in step 390. Geographic data associated with each vehicle trace is reintroduced to determine the representative vehicle route in step 380. To determine the spatial geometry of the representative vehicle route, an average vehicle trace is calculated using a sequence of geographic locations from each of the group's vehicle traces. A sequence of average geographic locations along the route through the road network is calculated, and spline interpolation of the sequence of average geographic locations, e.g., a cubic spline, is used to create a geometric line that represents one of the typical routes taken by a vehicle along the identified sequence of road sections through the road network.

[0087] Figure 18 shows three exemplary representative vehicle paths overlaid on a digital map of the analyzed section. The first representative path 1810 corresponds to the movement of the first group identified as described with respect to Figure 17; that is, the representative path starts in lane 1 and remains in lane 1 for the duration of the section. The second representative path 1830 corresponds to the movement of the third group identified as described with respect to Figure 17; that is, the second representative path 1830 starts in lane 2 and remains in lane 2 for the duration of the section. The third representative path 1840 corresponds to the movement of the fourth group identified as described with respect to Figure 17; that is, the fourth representative path 1840 starts in lane 2, transitions to lane 3, and then returns to lane 2.

[0088] Determining a representative vehicle route also involves determining a sequence of driving attributes associated with a sequence of average geographical locations. For example, the average vehicle speed at each average geographical location is calculated as the harmonic mean of the driving speeds associated with each vehicle in the vehicle trace at that location. Figure 19 shows the change in speed at each average geographical location for a section of the representative vehicle trace. As indicated by the legend in Figure 19, locations associated with speeds from 40 km / h to 60 km / h are represented by circles. Locations associated with speeds from 20 km / h to 40 km / h are represented by triangles. Locations associated with speeds from 0 km / h to 20 km / h are represented by squares.

[0089] The same process can be performed to determine other typical driving attributes such as acceleration and deceleration. Considering these attributes allows for a more comprehensive understanding of typical vehicle movements (e.g., driving behavior). For example, by analyzing changes in acceleration and velocity, the location of sections where the vehicle frequently stops can be identified. When the vehicle control system 114 accesses this information, it can warn the driver in advance that the vehicle is approaching a stopping zone, allowing the driver to gradually decelerate the vehicle to avoid a collision, or the vehicle control system 114 can directly decelerate the vehicle (e.g., by overriding the driver, or if the vehicle is unmanned).

[0090] Data associated with typical vehicle routes can also be used to identify other potentially hazardous sections of road. For example, sections of road where traffic lanes typically merge, or sections with significant variability in typical vehicle behavior, indicate potentially uncertain or complex driving conditions. Similar to identified stopping zones, when the vehicle control system 114 accesses this information, it can warn the user in advance that the vehicle is approaching a potentially more hazardous area, allowing the driver to slow down the vehicle and approach the next road section cautiously, or the vehicle control system 114 can directly slow down the vehicle (for example, by overriding the driver, or if the vehicle is unoccupied).

[0091] The central movement map database 132 can be periodically updated by using new vehicle traces as input and repeating the method described above. The resulting updated central movement map 132 can be transmitted to the vehicle 110 to periodically update the local movement map 122.

[0092] Those skilled in the art will understand that the present invention is not limited to the embodiments shown by describing one or more specific embodiments, and that many variations and modifications are possible within the scope of the appended claims.

Claims

1. A method for determining a representative vehicle route along a sequence of road sections in a road network, wherein the method is: Accessing multiple vehicle traces, wherein each vehicle trace includes a sequence of the geographical locations of each vehicle as it travels through the sequence of road sections. For each of the road sections in the sequence of road sections, the process of each of the vehicle traces to determine at least one section-specific route index for each of the vehicle traces, wherein the section-specific route index is selected from a predetermined set of one or more route indices for the road section, and each section-specific route index corresponds to a different set of possible routes along the road section, thereby determining a sequence of route indices for each of the plurality of vehicle traces. Determining a similarity-based partition of the multiple sequences of the root index, thereby dividing the multiple vehicle traces into multiple groups of vehicle traces, For each of the groups of vehicle traces, the sequence of geographical locations of one or more vehicle traces in the group is processed in order to determine a representative vehicle route of the group along the sequence of road sections. Methods that include...

2. The method according to claim 1, further comprising determining, for each vehicle trace, a lateral displacement value for all or a subset of the geographic locations in the sequence of geographic locations, wherein the lateral displacement value represents the lateral displacement of the geographic locations relative to a reference route along the sequence of road sections, thereby determining a sequence of lateral displacement values ​​for each of the plurality of vehicle traces.

3. The method according to claim 2, wherein the reference route is determined by processing all or a subset of the plurality of vehicle traces to calculate the average vehicle trace, the reference route being the average vehicle trace or following the average vehicle trace.

4. The method according to claim 2 or 3, further comprising processing each of the sequences of lateral displacement values ​​for each of the road sections in the sequence of road sections to determine the route index specific to each section of each vehicle trace.

5. The method according to any one of claims 1 to 4, comprising: determining the set of section-specific route indices by grouping the vehicle traces through the road sections into one or more road section clusters for each road section in the sequence of road sections; and assigning a different section-specific route index to each road section cluster of the vehicle traces.

6. The method according to any one of claims 1 to 5, wherein the predetermined set of one or more route indices for each road section is each set of one or more lane indices corresponding to the lanes of each road section.

7. The method according to any one of claims 1 to 6, wherein determining a representative vehicle driving route for each of the groups of vehicle traces includes processing the sequence of geographical locations of each of the vehicle traces in the group in order to calculate an average vehicle trace for the group.

8. The method according to any one of claims 1 to 7, further comprising determining a sequence of driving attribute values ​​for each of the groups of vehicle traces, wherein each driving attribute value is associated with a point along the sequence of road sections.

9. The method according to claim 8, wherein each sequence of driving attribute values ​​includes a sequence of vehicle speed measurements.

10. The method according to any one of claims 1 to 9, further comprising processing the representative vehicle path to identify one or more sections of the sequence of road sections as potentially hazardous.

11. Software that, when executed on a processing system, includes instructions causing the processing system to perform a method for determining a representative vehicle route along a sequence of road sections of a road network as described in any one of claims 1 to 10.

12. A processing system configured to perform a method for determining a representative vehicle route along a sequence of road sections of a road network according to any one of claims 1 to 10.

13. The processing system according to claim 12, further configured to communicate with a vehicle control system of a vehicle traveling through the road network, wherein the processing system is configured to transmit data including the representative vehicle route to the vehicle control system of the vehicle.

14. A vehicle control system for vehicles traveling on a road network, wherein the vehicle control system is Access data containing one or more representative vehicle routes along a sequence of road sections in the aforementioned road network, Determine the geographical location of the aforementioned vehicle, Based on the determined geographical location of the vehicle, one of the representative vehicle routes is selected. Using the information derived from the selected representative vehicle route, the movement of the vehicle is controlled, or the information is transmitted to the driver of the vehicle. A vehicle control system configured in such a way.

15. The vehicle control system according to claim 14, configured to determine the geographical location of the vehicle using sensor data output by one or more sensors located inside or on the vehicle, which optionally include a global navigation satellite system sensor positioned to receive satellite positioning signals.

Images