Map Matching and Lanes Number Estimation with Openstreetmap
Summary (3 min read)
Introduction
- In recent years, significant progress have been made in Driver Assistance Systems (DAS) towards automated driving.
- In [3], a vision-based application is proposed to estimate lanes number in road.
- In addition to the spatial information, the geodata contains detailed information such as the name, the limitation of speed and the number of lanes [10].
- Other mathematical tools has been used to perfom the map matching, in [15] a probabilistic criteria is calculated to select the right link on which the vehicle travels.
- Section IV presents the real-world experimental results.
II. THE OSM GEODATA
- For local sections of the planet, an XML file containing the latest revision of the OSM map can be downloaded from the official website of OpenStreetMap.
- Furthermore, regularly updates of the geodata are available at that website.
- According to the OSM specifications, the OSM data model consists of three basic geometric elements called Nodes, Ways and Relations.
A. Nodes
- Nodes ni are point-shaped geometric elements which are used to represent GPS points in term of latitude (lat) and longitude (lon).
- Moreover, Nodes are the basic points to represent the geometry of the Way.
B. Ways
- Throughout this paper, a single Way is defined as follows: W = (idw, Nw, Tw) (1) Where idw denotes a unique identification number of the Way.
- According to the OSM specifications, each Way can have up to 255 tags.
- Each tag consists of two elements, a key k and the corresponding value v : ti = (k, v) (4) As illustrated on Figure 1 an example of an OSM Way with its corresponding tags.
D. The proceeding of the OSM data
- For this paper, only two OSM tags from the 255 possible are used.
- The tags used are listed in the Table I. The OSM map consists of a set of Ways.
- To increase the robustness of their map matching a preprocessing stage is performed.
- Only Ways representing roads on which the vehicle may travel are selected as shown on Figure 2.
III. MAP MATCHING ALGORITHM
- As discussed before, an OSM map consists of a set of Ways, each Way is constructed from multiple Nodes.
- In the remaining sections, words segment and Way are switchable.
- To identify the Way matched with a GPS point, it is necessary to perform for each GPS node a ’discrimination stage’ to remove all the Ways which are not suitable for the map matching.
- Afterwards, an OSM Way is chosen from the remaining OSM Way candidates by calculating a probabilistic criterion.
- In their paper, three methods for calculating this criterion are presented.
A. Discrimination stage
- The aim of this stage is to remove all the Ways which are not suitable for map matching depending on several factors:.
- In their work, the authors take the closest distance as being the shortest distance to a Way (d0 in Figure 4a and d2 in Figure 4b).
- Ways with a distance smaller than a threshold are picked as Way candidates.
- As shown in Figure 5, the authors compute the angle difference between the GPS trace and the Way.
- If the speed of the vehicle is greater than a threshold then the Way is eliminated.
B. The probabilistic criteria
- The distance and orientation criteria are not always sufficient to select the right Way.
- There might still be an ambiguity on choosing the right Way, as pictured on Figure 6.
- Three methods of calculating this probabilistic criterion are presented: Probabilistic criterion based on Euclidean distance, Probabilistic criterion based on Mahalanobis distance, Probabilistic criterion based on the probability of be- longing to a segment.
- The distance to a road d is modeled as zero mean, normally distributed random variable with standard deviation σd described as follows: σd = .DOP 1 (9) Due to the uncertainty of the GPS data and the OSM map (beetween 6-9 m [16]).
- 2) Probabilistic criterion based on Mahalanobis distance: Using the covariance matrix associated with the pose of the vehicle ΣX , the authors compute the Mahalanobis distance for each Way candidate.
IV. REAL-WORLD EXPERIMENTAL RESULTS
- The GPS data were collected in the region of Paris for a total of 6596 GPS points.
- The results obtained on Table III support their argument.
- In addition the wrong estimations have also decreased, meaning that the wrong estimations are not only due to a wrong map matching.
V. CONCLUSION
- In this paper the authors presented a method for estimating the number of lanes in a road using a GPS receiver and the OSM geodata.
- The authors also estimated the lanes numbers by extracting the information about the number of lanes from the OSM database.
- To improve the results, a heuristic about the change in lanes number has been used.
- The authors plan to use this map matching algorithm on other maps in order to compare the results.
- The authors also plan to implement their algorithm in real time.
Did you find this useful? Give us your feedback
Citations
90 citations
23 citations
15 citations
Cites methods from "Map Matching and Lanes Number Estim..."
...• Endangered Lane based Critical Time (E-Lane): Depending on the values of tcritical for each lane and for a road configuration of two lanes (lane information are estimated from OpenSteetMap (OSM) [27] for example), this node has 3 states: Lane 1, Lane 2, Both Lanes on the lanes are endangered and emergency braking is not possible....
[...]
...Depending on the values of tcritical for each lane and for a road configuration of two lanes (lane information are estimated from OpenSteetMap (OSM) [27] for example), this node has 3 states: Lane 1, Lane 2, Both Lanes on the lanes are endangered and emergency braking is not possible....
[...]
14 citations
Additional excerpts
...As mentioned in [12], the OSM map does not provide information about the accuracy of its data....
[...]
...In our previous work [12], a multi-criteria map-matching algorithm based on multiple proba-...
[...]
...To do so, we introduced three discrimination criteria [12]: The first one is based on the distance between the ego-...
[...]
...2 the proposed module is an upgrade of our work presented in [12]....
[...]
...Based on equation (1) three probabilistic criteria [12] were developed: Ce: Criterion based on Euclidean distance....
[...]
12 citations
References
20,541 citations
"Map Matching and Lanes Number Estim..." refers methods in this paper
...By assuming the uniformly distributed prior probability for the Ways P (Wi) and using Bayes formula [17], we get:...
[...]
2,757 citations
"Map Matching and Lanes Number Estim..." refers methods in this paper
...Using the canonical representation for a Gaussian [18], we get:...
[...]
...Thus the probability we are looking for is defined as follows: Cp = PAB = ∫ N (X̄,ΣX)N (SAB ,ΣAB)dX (19) The product of two Gaussian distributions is a denormalized Gaussian distribution, we can rewrite (19) as follows: Cp = ∫ N (X̄,ΣX)N (SAB ,ΣAB)dX = ∫ kN (µ,Σ)dX (20) Using the canonical representation for a Gaussian [18], we get: Cp = k = 1 p exp(g1 + g2 + 1 2 µTΣ−1µ) (21) with: Σ = [Σ−1X + Σ −1 AB ] −1 µ = [Σ−1X + Σ −1 AB ] −1[Σ−1X X̄ + Σ −1 ABS̄AB ] p = log [(2π)−n/2|Σ|−1/2] g1 = log [(2π) n/2|ΣX |−1/2]− 12X̄ T Σ−1X X̄ g2 = log [(2π) n/2|ΣAB |−1/2]− 12 S̄ T ABΣ −1 ABS̄AB...
[...]
125 citations
"Map Matching and Lanes Number Estim..." refers methods in this paper
...In litterature, the use of the OpeenStreetMap geodata is limited to navigation tasks such a localisation [11]....
[...]
122 citations
"Map Matching and Lanes Number Estim..." refers methods in this paper
...Therefore, for our future work we want to use the information about the number of lanes as prior for a information driven lanes recognition process using a visionbased sensors [19]....
[...]
111 citations
"Map Matching and Lanes Number Estim..." refers background in this paper
...In [13] a map matching based on a weighting system is proposed....
[...]
Related Papers (5)
Frequently Asked Questions (11)
Q2. What have the authors stated for future works in "Map matching and lanes number estimation with openstreetmap" ?
The authors plan to use this map matching algorithm on other maps in order to compare the results. The authors also plan to implement their algorithm in real time. Therefore, for their future work the authors want to use the information about the number of lanes as prior for a information driven lanes recognition process using a visionbased sensors [ 19 ].
Q3. How many tags are used to represent the geometry of the Way?
In addition, the set Nw regroup the m Nodes ni representing the geometry of the Way as follows:Nw = {n1, n2, ..., nm} (2)To specify the semantic of each Way, a subset Tw of n tags is related to the Way.
Q4. How long did the algorithm take to perform?
their algorithm was performed on the integral data set under Matlab for a computation time of 12 minutes and 30 secon a intel core i7-6820HQ CPU 2.70 GHz.
Q5. What is the reason why the lanes number estimation is false?
The authors match the GPS traces with the existing map according to a probabilistic multicriteria map matching algorithm that takes into account several factors like distance, uncertainties on GPS data, angle between trace and road and the speed limitation.
Q6. What is the probability of a way being chosen?
2) Probabilistic criterion based on Mahalanobis distance: Using the covariance matrix associated with the pose of the vehicle ΣX , the authors compute the Mahalanobis distance for each Way candidate.
Q7. what is the probability of a road being mapped?
The distance to a road d is modeled as zero mean,normally distributed random variable with standard deviation σd described as follows:σd = .DOP 1 (9)Due to the uncertainty of the GPS data and the OSM map (beetween 6-9 m [16]).
Q8. What is the probability matrix for the way?
To address these issues, an extended Kalman-filter (EKF) is used in order to get the uncertainty matrix related to the pose of the vehicle.
Q9. What is the reason why the OSM map has been updated?
During the time interval between the two runs, the OSM map has been updated and therefore the map should provide more information about the lanes number.
Q10. What is the reason why the map matching is less accurate?
The map matching based on the Euclidean distance Cd providesthe less accurate results, which makes sense since this probabilistic criterion uses only inaccurate GPS measurements to calculate the uncertainty about the pose of the vehicle and the orientation of the vehicle.
Q11. What is the probability of a way being mapped?
The map matching task can be formulated as the calculation of the highest posterior probability of a GPS measurement Zk belonging to a Way Wi:arg max iP (Wi|Zk) (5)By assuming the uniformly distributed prior probability for the Ways P (Wi) and using Bayes formula [17], the authors get:arg max i p(Wi|Zk) = arg max i (p(Zk|Wi)P (Wi)p(Zk) ) (6)Since p(Zk) is constant for each Way candidate, then the authors have:arg max i p(Wi|Zk) ≡ arg max i (p(Zk|Wi)P (Wi)) (7)Since p(Wi) is uniformly distributed, the authors get:arg max i P (Wi|Zk) ≡ arg max i p(Zk|Wi) (8) The probability p(Zk|Wi) is modeled by two random independent variables : •