Graph Representation Querying of Machine Learning Models for Traffic or Safety Rules

Abstract

A graph representation of a tactical map representing a plurality of static components of an environment of a vehicle is generated. Nodes of the graph represent static components, and edges represent relationships between multiple static components. Different edge types are used to indicate respective relationship semantics among the static components. Individual nodes are represented as having the same number and types of edges in the graph. Using the graph as input to a neural network based model, a set of results is obtained. A motion control directive based at least in part on the results is transmitted to a motion-control subsystem of the vehicle.

1 .- 20 . (canceled) 
     
     
         21 . A method, comprising:
 generating a graph representation of one or more components of an environment of a vehicle;   obtaining, from one or more machine learning models, a response to a query pertaining to safety or traffic rules of a particular path of the vehicle, wherein input to the one or more machine learning models includes the graph representation; and   causing, based at least in part on the response to the query, the vehicle to move along the particular path.   
     
     
         22 . The method as recited in  claim 21 , wherein the one or more components of the environment include at least one static component. 
     
     
         23 . The method as recited in  claim 22 , wherein the static component comprises one of: (a) a lane segment of a road, (b) an intersection, (c) a traffic sign, (d) a traffic signal or (e) a pedestrian walkway. 
     
     
         24 . The method as recited in  claim 21 , wherein the graph representation comprises an edge representing one of: (a) a geometric constraint associated with at least a pair of static components of the environment, (b) a topological constraint associated with at least a pair of static components of the environment or (c) one or more attributes associated with at least a pair of static components of the environment. 
     
     
         25 . The method as recited in  claim 21 , wherein the one or more machine learning models include a neural network-based machine learning model. 
     
     
         26 . The method as recited in  claim 21 , further comprising:
 analyzing, using the one or more machine-learning models, a representation of one or more moving objects in the environment of the vehicle, wherein causing the vehicle to move along the particular path is based at least in part on a result of said analyzing.   
     
     
         27 . The method as recited in  claim 21 , wherein the graph comprises (a) a first node representing a first component of the environment, (b) a second node representing a second component of the environment, (c) a first edge belonging to a first edge type indicative of a relationship of a first semantic category between the first node and the second node, and (d) a second edge belonging to a second edge type indicative of a relationship of a second semantic category between the first node and the second node. 
     
     
         28 . A system, comprising:
 one or more processors; and   one or more memories;   wherein the one or more memories store program instructions that when executed on or across the one or more processors perform a method comprising:
 generating a graph representation of one or more components of an environment of a vehicle; 
 obtaining from one or more machine learning models, a response to a query pertaining to safety or traffic rules of a particular path of the vehicle, wherein input to the one or more machine learning models includes the graph representation; and 
 causing, based at least in part on the response to the query, the vehicle to move along the particular path. 
   
     
     
         29 . The system as recited in  claim 28 , wherein the one or more components of the environment include at least one static component. 
     
     
         30 . The system as recited in  claim 29 , wherein the static component comprises one of: (a) a lane segment of a road, (b) an intersection, (c) a traffic sign, (d) a traffic signal or (e) a pedestrian walkway. 
     
     
         31 . The system as recited in  claim 28 , wherein the graph representation comprises an edge representing one of: (a) a geometric constraint associated with at least a pair of static components of the environment, (b) a topological constraint associated with at least a pair of static components of the environment or (c) one or more attributes associated with at least a pair of static components of the environment. 
     
     
         32 . The system as recited in  claim 28 , wherein the one or more machine learning models include a neural network-based machine learning model. 
     
     
         33 . The system as recited in  claim 28 , wherein the method further comprises:
 analyzing, using the one or more machine-learning models, a representation of one or more moving objects in the environment of the vehicle, wherein causing the vehicle to move along the particular path is based at least in part on a result of said analyzing.   
     
     
         34 . The system as recited in  claim 28 , wherein the graph comprises (a) a first node representing a first component of the environment, (b) a second node representing a second component of the environment, (c) a first edge belonging to a first edge type indicative of a relationship of a first semantic category between the first node and the second node, and (d) a second edge belonging to a second edge type indicative of a relationship of a second semantic category between the first node and the second node. 
     
     
         35 . One or more non-transitory computer-accessible storage media storing program instructions that when executed on or across one or more processors cause the one or more processors to perform a method comprising:
 generating a graph representation of one or more components of an environment of the vehicle;   obtaining from one or more machine learning models, a response to a query pertaining to safety or traffic rules of a particular path of the vehicle, wherein input to the one or more machine learning models includes the graph representation; and   instructing, based at least in part on the response to the query, the vehicle to move along the particular path.   
     
     
         36 . The one or more non-transitory computer-accessible storage media as recited in  claim 35 , wherein the one or more components of the environment include at least one static component. 
     
     
         37 . The one or more non-transitory computer-accessible storage media as recited in  claim 36 , wherein the static component comprises one of: (a) a lane segment of a road, (b) an intersection, (c) a traffic sign, (d) a traffic signal or (e) a pedestrian walkway. 
     
     
         38 . The one or more non-transitory computer-accessible storage media as recited in  claim 35 , wherein the graph representation comprises an edge representing one of: (a) a geometric constraint associated with at least a pair of static components of the environment, (b) a topological constraint associated with at least a pair of static components of the environment or (c) one or more attributes associated with at least a pair of static components of the environment. 
     
     
         39 . The one or more non-transitory computer-accessible storage media as recited in  claim 35 , wherein the one or more machine learning models include a neural network-based machine learning model. 
     
     
         40 . The one or more non-transitory computer-accessible storage media as recited in  claim 35 , wherein the method further comprises:
 analyzing, using the one or more machine-learning models, a representation of one or more moving objects in the environment of the vehicle, wherein causing the vehicle to move along the particular path is based at least in part on a result of said analyzing.