Map based annotation for autonomous movement models training

A computerized apparatus having a processor, the processor being adapted to perform the steps of:

• obtaining a first driving scenario, wherein the first driving scenario comprises a first functional map comprising at least a representation of a first non-synthetic static topology of a first road segment, whereby defining a first original driving path of a vehicle driving in the first road segment;
• receiving a first modified driving scenario and a first driving path, wherein the first driving path is applicable within the first modified driving scenario, wherein the first modified driving scenario is defined based on the first driving scenario, wherein the first modified driving scenario introduces a path altering object into the first functional map, wherein the path altering object is a synthetic object, wherein the path altering object induces the first driving path that is different than the first original driving path;
• obtaining a second driving scenario, wherein the second driving scenario comprises a second functional map comprising at least a representation of a second non-synthetic static topology of a second road segment, whereby defining a second original driving path of a vehicle driving in the second road segment;
• receiving a second modified driving scenario, wherein the second modified driving scenario introduces a path non-altering object into the second functional map, wherein the path non-altering object is a synthetic object, whereby the second original driving path is applicable in the second modified driving scenario;
• training an autonomous driving model, wherein the autonomous driving model is configured to provide a predicted driving path within a road segment based on a functional map representation of the road segment, wherein said training comprises:
• indicating to the autonomous driving model that the first driving path is applicable to the first modified driving scenario;
• and
• indicating to the autonomous driving model that the second original driving path is applicable to the second modified driving scenario,
• whereby augmenting an original training set with map based annotated modified scenarios at least one of which affects a driving path and at least one of which does not affect a driving path.

A computer program product comprising a non-transitory computer readable storage medium retaining program instructions, which program instructions when read by a processor, cause the processor to perform a method comprising:

• obtaining a first driving scenario, wherein the first driving scenario comprises a first functional map comprising at least a representation of a first non-synthetic static topology of a first road segment, whereby defining a first original driving path of a vehicle driving in the first road segment;
• receiving a first modified driving scenario and a first driving path, wherein the first driving path is applicable within the first modified driving scenario, wherein the first modified driving scenario is defined based on the first driving scenario, wherein the first modified driving scenario introduces a path altering object into the first functional map, wherein the path altering object is a synthetic object, wherein the path altering object induces the first driving path that is different than the first original driving path;
• obtaining a second driving scenario, wherein the second driving scenario comprises a second functional map comprising at least a representation of a second non-synthetic static topology of a second road segment, whereby defining a second original driving path of a vehicle driving in the second road segment;
• receiving a second modified driving scenario, wherein the second modified driving scenario introduces a path non-altering object into the second functional map, wherein the path non-altering object is a synthetic object, whereby the second original driving path is applicable in the second modified driving scenario;
• training an autonomous driving model, wherein the autonomous driving model is configured to provide a predicted driving path within a road segment based on a functional map representation of the road segment, wherein said training comprises:
• indicating to the autonomous driving model that the first driving path is applicable to the first modified driving scenario;
• and
• indicating to the autonomous driving model that the second original driving path is applicable to the second modified driving scenario,
• whereby augmenting an original training set with map based annotated modified scenarios at least one of which affects a driving path and at least one of which does not affect a driving path.