System and method for map matching of GNSS positions of a vehicle

Abstract

Computer-implemented method for map matching of GNSS positions (G1, G2, G3) of a vehicle (1) with location information (P1a, P1b, P1c, P2a, P2b, P3a, P3b) of a digital road map (K), comprising the steps: Acquisition (S1) of GNSS positions (G1, G2, G3) of the vehicle (1) along a route of the vehicle (1); Recording (S2) a distance (E1, E2) determined using a vehicle sensor (10), in particular a wheel speed sensor, between each pair of detected, consecutive GNSS positions (G1, G2, G3) of the vehicle (1); and Map matching (S3) of the determined distance (E1, E2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) with a route length (L1, L2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) on the digital road map (K).

Description

[0001] The invention relates to a computer-implemented method for map matching the GNSS positions of a vehicle with location information from a digital road map. The invention further relates to a system for map matching the GNSS positions of a vehicle with location information from a digital road map.

[0002] Map matching involves mapping a sequence of GNSS positions onto the road network of a digital road map. Such map matching methods aim to improve the accuracy of the mapping, for example, of a vehicle's position onto corresponding road connections. For each GNSS position, the process determines which road the vehicle traveled on.

[0003] A distinction is made between online and offline map matching. While online map matching matches each GNSS position onto the road in real time without knowledge of subsequent GNSS positions, offline map matching matches GNSS positions after recording the journey or journey segment. Both online and offline map matching can be performed in the vehicle or in the backend after GNSS positions have been transmitted. However, online map matching is more commonly used in vehicles, while offline map matching is more common in the backend.

[0004] The vehicle's GNSS position is determined within the vehicle at a frequency of 1 Hz and can be collected at this or a lower frequency for further processing. Dead reckoning is also used to determine the vehicle's position even more precisely.

[0005] For various use cases, the vehicle positions, potentially improved by dead reckoning, are mapped onto the road network either in the vehicle or after transmission to a backend using offline map matching. Map matching in the backend, rather than in the vehicle, has the advantage that the backend always has the most up-to-date digital road map. Furthermore, offline map matching generally achieves better results than map matching while driving.

[0006] A road network can be modeled as a graph, as described, for example, in Newson, Paul, and John Krumm: “Hidden Markov map Matching through noise and sparseness.”, Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems, ACM, 2009. This graph can consist of both directed and undirected edges. Unlike the publication by Newson and Krumm, a directed edge does not necessarily represent a one-way street, as roads that can be traveled in both directions can also be modeled as two directed edges. Each edge has a description of its geometry, for example, as a polyline (i.e., a line composed of multiple segments). Mapmakers offer maps in different formats with varying modeling. For example, in some models, links may only end at intersections, or there may only be directed edges.However, the aforementioned model represents the most general case.

[0007] Newson and Krumm describe a map matching method based on the Hidden Markov Model (HMM). This method calculates the most probable sequence of links the vehicle has traveled over using the Viterbi algorithm. Each GNSS position is matched to a so-called matching, the combination of link and position on the link (abbreviated as ).<Link, Position auf Link> ) shown. The position on a link can be represented, for example, as a fraction, i.e., as a number between 0 and 1.

[0008] However, the aforementioned method according to Newson and Krumm is susceptible to map errors in road topology, such as missing roads in the road map, and road geometry.

[0009] Publication US 20120209507 A1 discloses a navigation device comprising: a receiver for receiving position sensor data; a memory unit in which a starting position is stored; map data comprising a map of navigable routes in an area; and an output device comprising a display or a speaker for outputting a current position. The navigation device further comprises a processor for determining distance traveled and changes in orientation from the position sensor data and for estimating the current location on one of the navigable routes of the map data. The estimation step includes using the distance traveled and changes in orientation to extrapolate from the starting position possible paths along the navigable routes that the navigation device may have traveled.Each path is assigned a probability that the navigation device has traveled that path. When the orientation changes, the processor extends each path and recalculates the probability. The processor assigns the highest probability to the current location as the currently predicted position on the path.

[0010] US 20080262730 A1 discloses a mobile object position estimation device mounted in a mobile object. This device comprises a first estimation mechanism for estimating the position of the mobile object based on an inertial navigation sensor device (e.g., accelerometer, yaw rate sensor, etc.) that encompasses the mobile object. It further comprises a storage unit for storing map data containing at least node data, and a second estimation mechanism for estimating the position of the mobile object. The position estimation is based on a predicted trajectory of the mobile object derived from the node data, and a synthesis mechanism for determining the final estimated position of the mobile object by synthesizing the positions estimated by the first and second estimation mechanisms.

[0011] Therefore, there is a need to provide an improved map matching method and system that offers higher accuracy.

[0012] The problem is solved using a computer-implemented method for map matching of GNSS positions of a vehicle with location information from a digital road map with the features of claim 1.

[0013] Furthermore, the problem is solved with a system for map matching of GNSS positions of a vehicle with location information from a digital road map with the features of claim 10.

[0014] Furthermore, the problem is solved with a computer program having the features of claim 11 and with a computer-readable data carrier having the features of claim 12.

[0015] The present invention provides a computer-implemented method for map matching GNSS positions with location information from a digital road map. Location information includes, for example, location-specific coordinates from the digital road map.

[0016] The procedure involves capturing the vehicle's GNSS positions along a route taken by the vehicle.

[0017] The method further comprises recording the distance between each pair of consecutive GNSS positions of the vehicle, determined using a vehicle sensor, in particular a wheel speed sensor. In addition, the method includes map matching the determined distance between each pair of consecutive GNSS positions of the vehicle with the route length between each pair of consecutive GNSS positions of the vehicle on the digital road map.

[0018] The present invention further provides a system for map matching the GNSS positions of a vehicle with location information from a digital road map. The system comprises means for acquiring the vehicle's GNSS positions along a route.

[0019] The system further comprises means for recording the distance between any two consecutive GNSS positions of the vehicle, determined using a vehicle sensor, in particular a wheel speed sensor. The method also includes means for map-matching the determined distance between the two consecutive GNSS positions of the vehicle with the route length between the two consecutive GNSS positions of the vehicle on the digital road map.

[0020] The present invention furthermore provides a computer program with program code to carry out the method according to the invention when the computer program is executed on a computer.

[0021] The present invention further provides a computer-readable data carrier containing the program code of a computer program in order to carry out the method according to the invention when the computer program is executed on a computer.

[0022] One aspect of the present invention is to record the distance traveled from each GNSS position using the vehicle's odometry since the start of the journey, thereby improving the accuracy of map matching. This is achieved by comparing the distance traveled between two consecutive GNSS positions, as determined by odometry, with the route length between the two matched GNSS positions on the digital road map. This advantageously results in higher map matching accuracy.

[0023] Advantageous embodiments and further developments are described in the dependent claims and in the description with reference to the figures.

[0024] According to a preferred further development, map matching is performed in real time after recording a trip or a trip segment using an in-vehicle computing unit and / or an external server, with each captured GNSS position of the vehicle being assigned a timestamp. This allows for the advantageous use of an optimal implementation consisting of online and / or offline map matching, depending on system requirements.

[0025] According to another preferred further training, it is provided that a probability density for a transition from matching candidate c t,i in a time step t to matching candidate c t+1,j is calculated in a time step t+1 using the following equation: p(r,d)=1σ2πe−12(r−dσ)2¶ where r is the length of the fastest route between c t,i and c t+1,jis; d is the distance traveled between times t and t+1; p or p(r, d) is the probability density, where σ is the standard deviation of the acquired GNSS positions, and where i and j are numerical placeholders for Ground Truth street segments and location information respectively, where the location information in the present embodiment is designated P1a, P1b, P1c, P2a, P2b, P3a, P3b.

[0026] In the present embodiment, the matching candidates are designated M1a, M1b, M1c, M2a, M2b, M3a, and M3b. The probability density for the transition from matching candidate c t,i in a time step t to matching candidate c t+1,j In a time step t+1, for example, the probability density for the transition from matching candidate M1c to matching candidate M2b could be denoted.

[0027] In a further preferred refinement, a kernel density estimator is used instead of the equation mentioned above. This advantageously allows a probability distribution to be calculated taking into account the length of the fastest route and the distance traveled, without having to make assumptions about the structure of the probability distribution.

[0028] According to a further preferred embodiment, a standard deviation σ is determined via the square root of a sample variance of a sample of GNSS positions and vehicle sensor distance measurements from a plurality of journeys, for which the route traveled is known. Advantageously, the standard deviation is determined via the square root of the sample variance of the sample of ground truth data. The ground truth data used are the GNSS positions and odometry measurements from several journeys, for which the route traveled is known.

[0029] According to a further preferred enhancement, transitions between matching candidates are excluded if the route length of the fastest route between the matching candidates is shorter or longer than the distance determined using the vehicle sensor by a predetermined factor. This advantageously allows for the use of an additional criterion, which enables improved determination of the best possible matching candidate.

[0030] According to a further preferred embodiment, GNSS positions, current speed, and / or current average speed of a large number of vehicles are periodically transmitted to the external server. Using this information, the external server calculates traffic information, in particular an expected time of arrival and / or average speeds for road segments, and provides it to the large number of vehicles. Map matching can thus also be used to calculate a more accurate prediction of an expected time of arrival and / or average speeds for road segments and provide this information to a large number of vehicles in a fleet.

[0031] According to a further preferred embodiment, hazard data recorded by vehicle sensors, in particular concerning a slippery road surface, a traffic accident, and / or airbag activation, are transmitted to the vehicle's external server along with the vehicle's GNSS position. The external server then performs map matching of the hazard data and makes this data available to vehicles whose planned route passes through a detected hazard. Map matching can thus also be advantageously used to provide hazard information to other vehicles in the fleet.

[0032] According to a further preferred enhancement, a driver's personal route is recorded as a sequence of GNSS positions and sent to the external server. The external server then performs map matching of the GNSS positions, transmits the driver's learned personal route to the vehicle, and suggests navigation along the learned route to the driver at predetermined times. Thus, learning a driver's personal route can be advantageously used to suggest this route to the driver at suitable times when the system, based on the collected data, considers it likely that the route will be travelled.

[0033] According to a further preferred embodiment, the distance between any two consecutive GNSS positions of the vehicle, determined using the vehicle sensor, is used to validate map matching results. This involves calculating the length of the fastest route between two adjacent matched GNSS positions. If this length is shorter or longer than the distance determined using the vehicle sensor by a predetermined factor, at least one of the two determined GNSS positions of the vehicle is considered implausible. Such a validation of the map matching results also advantageously contributes to higher map matching accuracy.

[0034] The described configurations and training programs can be combined in any way desired.

[0035] Further possible embodiments, developments and implementations of the invention also include combinations of features of the invention described previously or subsequently with regard to the exemplary embodiments that are not explicitly mentioned. Brief description of the drawings

[0036] The accompanying drawings are intended to provide a further understanding of the embodiments of the invention. They illustrate embodiments and, in conjunction with the description, serve to explain the principles and concepts of the invention.

[0037] Other embodiments and many of the aforementioned advantages become apparent with reference to the drawings. The elements depicted in the drawings are not necessarily shown to scale.

[0038] They show: Fig. 1 a schematic representation of a system for map matching of GNSS positions of a vehicle with location information of a digital road map according to a preferred embodiment of the invention; Fig. 2 a schematic representation of a digital road map to illustrate the method for map matching GNSS positions of the vehicle with location information of the digital road map according to the preferred embodiment of the invention; and Fig. 3 A flowchart of a computer-implemented procedure for map matching of GNSS positions of a vehicle with location information from the digital road map.

[0039] The in Fig. The system shown in Figure 1 for map matching GNSS positions of the vehicle 1 with location information from a digital road map comprises the vehicle 1 having means 20 for acquiring GNSS positions of the vehicle 1 along a route of the vehicle 1. GPS is the preferred embodiment of GNSS in the present embodiment.

[0040] Furthermore, the system comprises means 22 for recording a distance determined using a vehicle sensor 10, in particular a wheel speed sensor, between any two recorded, consecutive GNSS positions of the vehicle 1. In addition, the system comprises means 24 for map matching the determined distance between any two recorded, consecutive GNSS positions of the vehicle 1 with a route length between any two recorded, consecutive GNSS positions of the vehicle on the digital road map.

[0041] Map matching is preferably performed in real time using an in-vehicle computing unit 12 and an external server 14, whereby each captured GNSS position of the vehicle 1 is assigned a timestamp.

[0042] Alternatively, map matching can be performed, for example, after recording a journey or after recording a journey segment using the vehicle's internal computing unit 12 and / or the vehicle's external server 14.

[0043] The 24 resources for map matching can be used as in Fig. 1 shown alternatively, for example, on the vehicle-external server 14.

[0044] Vehicle 1 also has further features in Fig. 1. Not shown, including a speed sensor, a sensor for detecting airbag activation, sensors for detecting road slipperiness and / or camera sensors.

[0045] The vehicle-external server 14 has a Fig. One receiving unit (not shown) is used. Furthermore, the external server 14 uses the data received from vehicle 1 to calculate traffic information, hazard warnings, and / or to learn the driver's personal route. The calculated data is then made available to other vehicles in the fleet.

[0046] Fig. Figure 2 shows a schematic representation of a digital road map to illustrate the method for map matching of GNSS positions of the vehicle with location information of the digital road map according to the preferred embodiment of the invention.

[0047] In the present illustration, three GNSS positions G1, G2, and G3 of vehicle 1 are shown as examples on the digital road map K. Since, in this embodiment, the GNSS positions G1, G2, and G3 of vehicle 1 are offset from their respective roads on the digital road map K, a map matching of the GNSS positions G1, G2, and G3 of vehicle 1 with location information from the digital road map K must be performed to calculate the movement path of vehicle 1.

[0048] For example, the first GNSS position G1 of vehicle 1 is located near the three road sections or location information P1a, P1b, P1c.

[0049] Similarly, a second GNSS position G2 of vehicle 1 is located near two road segments or location information points P2a and P2b. Furthermore, a third GNSS position G3 of vehicle 1 is located near two roads or road segments or location information points P3a and P3b.

[0050] Therefore, the GNSS positions G1, G2, G3 of vehicle 1 are first recorded along a route of vehicle 1. Subsequently, a distance E1, E2 is recorded between each pair of recorded, consecutive GNSS positions of vehicle 1, determined using the vehicle sensor, in particular the wheel speed sensor, for example, a distance E1 between the first GNSS position G1 and the second GNSS position G2.

[0051] Subsequently, a comparison is made between the determined distance E1 between the first GNSS position G1 and the second GNSS position G2 of vehicle 1 with a first route length L1 on the digital road map K.

[0052] A second distance E2 is also recorded between the second GNSS position G2 and the third GNSS position G3, and a comparison of a specific distance E2 between the second GNSS position G2 and the third GNSS position G3 with a route length L2 on the digital road map K is carried out.

[0053] Furthermore, a probability density for a transition from matching candidate c is calculated. t,i in a time step t to matching candidate c t+1,j calculated in one time step t+1 using the following equation: p(r,d)=1σ2πe−12(r−dσ)2¶ where r is the length of the fastest route between c t,i and c t+1,jis; d is the distance traveled between times t and t+1; p or p(r, d) is the probability density, where σ is the standard deviation of the acquired GNSS positions, and where i and j are numerical placeholders for Ground Truth street segments and location information respectively, where the location information in the present embodiment is designated P1a, P1b, P1c, P2a, P2b, P3a, P3b.

[0054] In the present embodiment, the matching candidates are designated M1a, M1b, M1c, M2a, M2b, M3a, and M3b. The probability density for the transition from matching candidate c t,i in a time step t to matching candidate c t+1,j In a time step t+1, for example, the probability density for the transition from matching candidate M1c to matching candidate M2b could be denoted.

[0055] A standard deviation σ is determined via the square root of a sample variance of a sample of GNSS positions G1, G2, G3 and vehicle sensor distance measurements of a plurality of journeys, of which the route traveled is known.

[0056] Transitions between matching candidates M1a, M1b, M1c, M2a, M2b, M3a, M3b are excluded if the route length L1, L2 of the fastest route between the matching candidates M1a, M1b, M1c, M2a, M2b, M3a, M3b is shorter or longer by a predetermined factor than the distance E1, E2 determined using the vehicle sensor 10.

[0057] GNSS positions G1, G2, G3, a current speed and / or a current average speed of a large number of vehicles are periodically transmitted to the vehicle-external server 14, using which the vehicle-external server 14 calculates traffic information, in particular an expected time of arrival and / or average speeds of road segments, and provides it to the large number of vehicles.

[0058] Hazard data recorded by vehicle sensors, in particular concerning a slippery road surface, a traffic accident and / or an airbag activation, are transmitted together with the GNSS position G1, G2, G3 of vehicle 1 to the vehicle-external server 14, whereby the vehicle-external server 14 performs a map-matching of the hazard data and makes this data available to vehicles 1 whose planned route leads through a detected hazard.

[0059] A driver's personal route is recorded as a sequence of GNSS positions G1, G2, G3 and sent to the vehicle-external server 14.

[0060] The vehicle-external server 14 performs a map matching of the GNSS positions G1, G2, G3, transmits the learned personal route of the driver to the vehicle 1 and suggests a route guidance to the driver at predetermined times along the learned personal route.

[0061] The distance E1, E2 determined using the vehicle sensor 10 between each pair of consecutive GNSS positions G1, G2, G3 of the vehicle 1 is used to verify the plausibility of map matching results.

[0062] The length of the fastest route between two adjacent matched GNSS positions G1, G2, G3 is calculated. If this length is shorter or longer than the distance E1, E2 determined using vehicle sensor 10 by a predefined factor, at least one of the two determined GNSS positions G1, G2, G3 of vehicle 1 is considered implausible.

[0063] Fig. Figure 3 shows a flowchart of a computer-implemented procedure for map matching of GNSS positions of a vehicle with location information from the digital road map.

[0064] The procedure includes acquiring S1 of GNSS positions G1, G2, G3 of vehicle 1 along a route of vehicle 1.

[0065] The method further includes recording S2 a distance E1, E2 determined using a vehicle sensor 10, in particular a wheel speed sensor, between each pair of detected, consecutive GNSS positions G1, G2, G3 of the vehicle 1.

[0066] The procedure also includes a map matching S3 of the determined distance E1, E2 between the two recorded, consecutive GNSS positions G1, G2, G3 of vehicle 1 with a route length L1, L2 between the two recorded, consecutive GNSS positions G1, G2, G3 of vehicle 1 on the digital road map K.

[0067] The term vehicle includes cars, trucks, buses, motorhomes, motorcycles, etc., which are used for the transport of people, goods, etc.

[0068] In particular, the term includes motor vehicles for passenger transport. Additionally or alternatively, a hybrid or electric vehicle, depending on its design, can be a pure electric vehicle (BEV) or a plug-in hybrid vehicle (PHEV). However, other drive systems can also be used, for example, in the form of a diesel or gasoline-powered vehicle. The vehicle can also be in the form of a rail vehicle.

[0069] Alternatively, the map matching S3 can be performed, for example, between any two GNSS positions G1, G2, G3 of vehicle 1 with a route length L1, L2 between any two corresponding GNSS positions G1, G2, G3 of vehicle 1 on the digital road map K.

Claims

[1] Computer-implemented method for map matching of GNSS positions (G1, G2, G3) of a vehicle (1) with location information (P1a, P1b, P1c, P2a, P2b, P3a, P3b) of a digital road map (K), comprising the steps: Acquisition (S1) of GNSS positions (G1, G2, G3) of the vehicle (1) along a route of the vehicle (1); Recording (S2) a distance (E1, E2) determined using a vehicle sensor (10), in particular a wheel speed sensor, between each pair of detected, consecutive GNSS positions (G1, G2, G3) of the vehicle (1); and Map matching (S3) of the determined distance (E1, E2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) with a route length (L1, L2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) on the digital road map (K). [2] Computer-implemented method according to claim 1, characterized by , that the map matching (S3) is performed in real time, after recording a journey or after recording a journey segment using an in-vehicle computing device (12) and / or an external server (14), and wherein each captured GNSS position (G1, G2, G3) of the vehicle (1) is assigned a timestamp. [3] Computer-implemented method according to claim 1 or 2, characterized by , that a probability density for a transition from matching candidate c t,i in a time step t to matching candidate c t+1,j is calculated in a time step t+1 using the following equation: p(r,d)=1σ2πe−12(r−dσ)2¶ where r is the length of the fastest route between c t,i and c t+1,jis; d is the distance traveled between times t and t+1; p or p(r, d) is the probability density, where σ is the standard deviation of the detected GNSS positions. [4] Computer-implemented method according to claim 3, characterized by , that a standard deviation σ is determined via the square root of a sample variance of a sample of GNSS positions (G1, G2, G3) and vehicle sensor distance measurements of a plurality of journeys, of which the route traveled is known. [5] Computer-implemented method according to claim 3 or 4, characterized by , that transitions between matching candidates (M1a, M1b, M1c, M2a, M2b, M3a, M3b) are excluded if the route length (L1, L2) of the fastest route between the matching candidates (M1a, M1b, M1c, M2a, M2b, M3a, M3b) is shorter or longer by a predetermined factor than the distance (E1, E2) determined using the vehicle sensor (10). [6] Computer-implemented method according to any one of claims 2 to 5, characterized by , that GNSS positions (G1, G2, G3), a current speed and / or a current average speed of a multitude of vehicles are periodically transmitted to the vehicle-external server (14), using which the vehicle-external server (14) calculates traffic information, in particular an expected time of arrival and / or average speeds of road segments, and provides it to the multitude of vehicles. [7] Computer-implemented method according to any one of claims 2 to 6, characterized by, that hazard data detected by vehicle sensors, in particular concerning a slippery road surface, a traffic accident and / or an airbag activation, together with the GNSS position (G1, G2, G3) of the vehicle (1), are transmitted to the vehicle-external server (14), wherein the vehicle-external server (14) performs a map-matching of the hazard data and makes this data available to vehicles (1) whose planned route passes through a detected hazard. [8] Computer-implemented method according to any one of claims 2 to 7, characterized by , that a driver's personal route is recorded as a sequence of GNSS positions (G1, G2, G3) and sent to the vehicle-external server (14), the vehicle-external server (14) performs a map-matching of the GNSS positions (G1, G2, G3), transmits the learned personal route of the driver to the vehicle (1) and suggests a route guidance of the learned personal route to the driver at predetermined times. [9] Computer-implemented method according to any one of claims 3 to 8, characterized by , that the distance (E1, E2) determined using the vehicle sensor (10) between each pair of consecutive GNSS positions (G1, G2, G3) of the vehicle (1) is used to validate map matching results, whereby the length of the fastest route between two adjacent matched GNSS positions (G1, G2, G3) is calculated, and if the length is shorter or longer than the distance (E1, E2) determined using the vehicle sensor (10) by a predetermined factor, at least one of the two determined GNSS positions (G1, G2, G3) of the vehicle (1) is considered implausible. [10] System for map matching of GNSS positions (G1, G2, G3) of a vehicle (1) with location information (P1a, P1b, P1c, P2a, P2b, P3a, P3b) of a digital road map (K), comprising: Means (20) for acquiring GNSS positions (G1, G2, G3) of the vehicle (1) along a route of the vehicle (1); Means (22) for recording a distance (E1, E2) determined using a vehicle sensor (10), in particular a wheel speed sensor, between each pair of detected, consecutive GNSS positions (G1, G2, G3) of the vehicle (1); and Means (24) for map matching the determined distance (E1, E2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) with a route length (L1, L2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) on the digital road map (K). [11] Computer program with program code to perform the method according to any one of claims 1 to 9 when the computer program is executed on a computer. [12] Computer-readable data carrier containing program code of a computer program for carrying out the method according to any one of claims 1 to 9 when the computer program is executed on a computer.

Images