Harman Becker Automotive Systems Gmbh

Volodymyr Ivanov

Alexey Pryakhin

Ievgenii Borshch

Eric Theisinger

Roland Preiss

Tobias Emrich

A map fusing method includes receiving a source graph and a target graph. The source graph is representative of a source map and the target graph is representative of a target map and includes nodes and edges that connect the nodes. The method further includes processing each of the source graph and the target graph in a graph convolutional layer to provide graph convolutional layer outputs related to the source graph and to the target graph, processing each of the graph convolutional layer outputs for the source graph and the target graph in a linear rectifying layer to output node feature maps related to the source graph and the target graph. The method further includes selecting pairs of node representations from the node feature maps related to the source graph and the target graph and concatenating the selected pairs to output selected and concatenated pairs of node representations.

Road map fusion

In one embodiment, an attribute application associates content with a road segment. In operation, the attribute application generates a spatial reference identifier based on coordinates associated with the attribute. The attribute application then generates an attribute based on the content and the spatial reference identifier. Finally, the attribute application transmits the road segment attribute to a navigation system that performs at least one navigation operation based on a road database and the attribute. Because the attribute is specified based on spatial referencing, the attribute application requires fewer resources to generate attributes than conventional approaches that generate different attributes for different versions and formats of road databases.

Generating road segment attributes based on spatial referencing

In one embodiment, a match application identifies an edge corresponding to a road segment based on an observational point specifying an estimated position and an estimated heading. The match application performs search operations on a road network graph based on the estimated position to identify multiple candidate edges. For a first candidate edge, the match application computes a first quality level based on at least a distance between the estimated position and the first candidate edge and an angle between the estimated heading and a direction associated with the first candidate edge. Subsequently, the match application determines that the first quality level indicates that the first candidate edge has a higher quality than any other candidate edge. Finally, the match application transmits the first candidate edge to an application that associates content specified in relation to the estimated position with the first candidate edge.

Matching observational points to road segments represented as edges in graphs

In one embodiment, a layer application generates navigation-related data based on observations associated with vehicles. In operation, the layer application generates an attribute based on a first observation associated with a vehicle and an observation type. The layer application then compares the attribute to an existing road database to generate a dynamic layer. Subsequently, the layer application causes the dynamic layer to be transmitted to a navigation subsystem. After receiving the dynamic layer, the navigation subsystem performs navigation operation(s) based on the road database and the dynamic layer. Because the layer application may receive and operate on real-time observations, providers of road databases may produce dynamic layers at a frequency that enables navigation applications to continually provide accurate navigation data.

Dynamic layers for navigation database systems

In one embodiment, a speed limit application associates speed limits with road segments based on a road graph. In operation, the speed limit application selects a source road segment that meets a target road segment at an intersection based on the road graph. The source road segment is associated with a speed limit. Subsequently, the speed limit application determines a confidence value associated with extrapolating the first speed limit to the target road segment based on the first road graph. The speed limit application then determines that the confidence value indicates that a confidence in the extrapolation satisfies a minimum confidence requirement. Consequently, the speed limit application generates an attribute that associates the first speed limit with the target road segment. Finally, the speed limit application causes a navigation-related operation to be performed based on the attribute.

Extrapolating speed limits within road graphs

In one embodiment, a cluster application generates navigation-related data based on observations received from vehicles. In operation, the cluster application computes an oriented distance between two observed object positions based on a heading, where each observed object position is associated with a different one of two observations. The cluster application then generates a cluster that includes the two observations based on the oriented distance. Subsequently, the cluster application computes an object position that is associated with the cluster based on the two observations. The cluster application transmits the object position and at least one characteristic associated with the observations to a update application that generates an update to a road database. Because the cluster application computes the object position based on multiple observations that are likely of a single object, the object position associated with the cluster may be more reliable than the observed object positions.

Growing community of inventors

Odelzhausen, Germany

Volodymyr Ivanov