Traffic sign relevancy

Abstract

Systems and methods are provided for vehicle navigation. In one implementation, a system for navigating a vehicle may include at least one processor configured to receive a first image frame; detect in the first image frame a representation of a traffic light and determine a color state associated with lamps included on the traffic light. The at least one processor may receive an additional image frame includes a representation of the at least one traffic light; and determine, based on a comparison of the first image frame and the additional image frame, whether the at least one traffic light includes a blinking lamp. If the at least one traffic light includes a blinking lamp, the processor may cause the vehicle to implement a navigational action relative the traffic light in accordance with the determination and also based on a detected color state for the blinking lamp.

1 - 71 . (canceled) 
     
     
         72 . A system for generating a crowd-sourced map for use in vehicle navigation, the system comprising:
 at least one processor comprising circuitry and a memory, wherein the memory includes instructions that when executed by the circuitry cause the at least one processor to:
 receive drive information collected from a plurality of vehicles that traversed a road segment, wherein the road segment is associated with a plurality of traffic signs; 
 generate two or more drivable paths for the road segment based on aggregated motion characteristics of the plurality of vehicles as they traversed the road segment; 
 store the generated two or more drivable paths in the crowd-sourced map; 
 determine, based on analysis of the drive information collected from the plurality of vehicles, whether a particular traffic sign among the plurality of traffic signs is relevant to each one of the two or more drivable paths; 
 store in the crowd-sourced map indications of which ones of the plurality of traffic signs are relevant to each one of the two or more drivable paths; and 
 make the crowd-sourced map available to at least one host vehicle for navigation relative to the plurality of traffic signs. 
   
     
     
         73 . The system of  claim 72 , wherein the two or more drivable paths are represented in the crowd-sourced map as three-dimensional splines. 
     
     
         74 . The system of  claim 73 , wherein each of the two or more drivable paths is associated with a different travel lane along the road segment. 
     
     
         75 . The system of  claim 72 , wherein the determination of whether a particular traffic sign is relevant to each one of the two or more drivable paths is based on motion characteristics of the plurality of vehicles as they traversed the road segment. 
     
     
         76 . The system of  claim 75 , wherein the motion characteristics indicate that the plurality of vehicles slowed in a vicinity of the particular traffic sign. 
     
     
         77 . The system of  claim 75 , wherein the motion characteristics indicate that the plurality of vehicles changed heading direction in a vicinity of the particular traffic sign. 
     
     
         78 . The system of  claim 72 , wherein the determination of whether a particular traffic sign is relevant to each one of the two or more drivable paths is based on a proximity of the particular traffic sign to one or more road features represented in the crowd-sourced map. 
     
     
         79 . The system of  claim 78 , wherein the one or more road features includes a roundabout, a stop line, a merge point, a lane split, or a curvature associated with a drivable path. 
     
     
         80 . The system of  claim 72 , wherein the determination of whether a particular traffic sign is relevant to a drivable path from the two or more drivable paths is based on whether a traffic sign similar to the particular traffic sign appears on an opposing side of the drivable path from the particular traffic sign. 
     
     
         81 . The system of  claim 72 , wherein the determination of whether a particular traffic sign is relevant to a particular one of the two or more drivable paths is based on a lateral distance between the particular drivable path and the particular traffic sign. 
     
     
         82 . The system of  claim 81 , wherein the determination of whether a particular traffic sign is relevant to a particular one of the two or more drivable paths is further based on a lateral distance between the particular traffic sign and a road edge. 
     
     
         83 . The system of  claim 72 , wherein the determination of whether a particular traffic sign is relevant to a particular one of the two or more drivable paths is based on a detected height of the particular traffic sign relative to the particular drivable path. 
     
     
         84 . The system of  claim 72 , wherein the determination of whether a particular traffic sign is relevant to a particular one of the two or more drivable paths is based on a determination of whether a semantic portion of the particular traffic sign is visible when traveling along a particular drivable path in a direction associated with the particular drivable path. 
     
     
         85 . The system of  claim 72 , wherein the determination of whether a particular traffic sign among the plurality of traffic signs is relevant to each one of the two or more drivable paths is based on output provided by at least one machine learning model trained to predict relevancy of traffic signs to drivable paths based on received input including at least one of: traffic sign geometry relative to a drivable path, traffic sign visibility relative to a drivable path, traffic sign geometry relative to one or more road features associated with the road segment, or motion characteristics of one or more of the plurality of vehicles as it navigated along the road segment. 
     
     
         86 . The system of  claim 72 , wherein the drive information includes at least a portion of images captured by one or more cameras associated with each of the plurality of vehicles as it navigated relative to the road segment. 
     
     
         87 . The system of  claim 72 , wherein the drive information includes motion characteristics of each of the plurality of vehicles as it navigated relative to the road segment. 
     
     
         88 . The system of  claim 72 , wherein the drive information includes indicators of paths followed by each of the plurality of vehicles as it navigated relative to the road segment. 
     
     
         89 . The system of  claim 72 , wherein the drive information includes locations and type indicators of traffic signs detected as each of the plurality of vehicles navigated relative to the road segment. 
     
     
         90 . The system of  claim 72 , wherein the drive information includes vehicle stop locations relative to the road segment. 
     
     
         91 . The system of  claim 72 , wherein the drive information includes location and type information for one or more landmarks detected as each of the plurality of vehicles navigated relative to the road segment. 
     
     
         92 . The system of  claim 91 , wherein the recognized landmark includes one or more of a pole, a lamp a post, or a lane marking. 
     
     
         93 . The system of  claim 72 , wherein the navigation relative to the plurality of traffic signs includes a warning issued to a vehicle operator. 
     
     
         94 . The system of  claim 72 , wherein the navigation relative to the plurality of traffic signs includes autonomous braking relative to a detected traffic sign. 
     
     
         95 . A method for generating a crowd-sourced map for use in vehicle navigation, the method comprising:
 receiving drive information collected from a plurality of vehicles that traversed a road segment, wherein the road segment is associated with a plurality of traffic signs;   generating two or more drivable paths for the road segment based on aggregated motion characteristics of the plurality of vehicles as they traversed the road segment;   storing the generated two or more drivable paths in the crowd-sourced map;   determining, based on analysis of the drive information collected from the plurality of vehicles, whether a particular traffic sign among the plurality of traffic signs is relevant to each one of the two or more drivable paths;   storing in the crowd-sourced map indications of which ones of the plurality of traffic signs are relevant to each one of the two or more drivable paths; and   making the crowd-sourced map available to at least one host vehicle for navigation relative to the plurality of traffic signs.   
     
     
         96 . A non-transitory computer readable medium containing instructions that when executed by at least one processor, cause the at least one processor to perform operations comprising:
 receiving drive information collected from a plurality of vehicles that traversed a road segment, wherein the road segment is associated with a plurality of traffic signs;   generating two or more drivable paths for the road segment based on aggregated motion characteristics of the plurality of vehicles as they traversed the road segment;   storing the generated two or more drivable paths in a crowd-sourced map;   determining, based on analysis of the drive information collected from the plurality of vehicles, whether a particular traffic sign among the plurality of traffic signs is relevant to each one of the two or more drivable paths;   storing in the crowd-sourced map indications of which ones of the plurality of traffic signs are relevant to each one of the two or more drivable paths; and   making the crowd-sourced map available to at least one host vehicle for navigation relative to the plurality of traffic signs.   
     
     
         97 . A system for navigating a host vehicle, the system comprising:
 at least one processor comprising circuitry and a memory, wherein the memory includes instructions that when executed by the circuitry cause the at least one processor to:
 receive an image acquired by at least one camera onboard the host vehicle as the host vehicle traverses a road segment; 
 detect a representation of at least one traffic sign in the acquired image; 
 access a crowd-sourced map generated based on drive information collected from a plurality of vehicles that previously traversed the road segment, wherein the crowd-sourced map stores indicators of relevancy of mapped traffic signs to certain drivable paths stored in the crowd-sourced map; 
 determine, based on the accessed crowd-sourced map, whether the at least one traffic sign detected in the acquired image is relevant to a drivable path along which the host vehicle is traveling; and 
 in response to a determination that the at least one traffic sign detected in the acquired image is relevant to a drivable path along which the host vehicle is traveling, cause the host vehicle to take at least one navigational action relative to the at least one traffic sign. 
   
     
     
         98 . The system of  claim 97 , wherein the memory further includes instructions that when executed by the circuitry cause the at least one processor to: in response to a determination that the at least one traffic sign detected in the acquired image is not relevant to a drivable path along which the host vehicle is traveling, cause the host vehicle to forego a navigational response relative to the at least one traffic sign. 
     
     
         99 . The system of  claim 97 , wherein the at least one traffic sign is a stop sign, and the at least one navigational action includes braking the host vehicle. 
     
     
         100 . The system of  claim 97 , wherein the at least one traffic sign is a yield sign, and the at least one navigational action includes braking the host vehicle and changing a heading direction of the host vehicle. 
     
     
         101 . The system of  claim 97 , wherein the drivable paths are stored in the crowd-sourced map as three-dimensional splines. 
     
     
         102 . The system of  claim 97 , wherein the drivable paths stored in the crowd-sourced map are determined by aggregating driving paths followed by the plurality of vehicles that previously traversed the road segment. 
     
     
         103 . A method for navigating a host vehicle, the method comprising:
 receiving an image acquired by at least one camera onboard the host vehicle as the host vehicle traverses a road segment;   detecting a representation of at least one traffic sign in the acquired image;   accessing a crowd-sourced map generated based on drive information collected from a plurality of vehicles that previously traversed the road segment, wherein the crowd-sourced map stores indicators of relevancy of mapped traffic signs to certain drivable paths stored in the crowd-sourced map;   determining, based on the accessed crowd-sourced map, whether the at least one traffic sign detected in the acquired image is relevant to a drivable path along which the host vehicle is traveling; and   in response to a determination that the at least one traffic sign detected in the acquired image is relevant to a drivable path along which the host vehicle is traveling, causing the host vehicle to take at least one navigational action relative to the at least one traffic sign.   
     
     
         104 . A non-transitory computer readable medium containing instructions that when executed by at least one processor, cause the at least one processor to perform operations comprising:
 receiving an image acquired by at least one camera onboard a host vehicle as the host vehicle traverses a road segment;   detecting a representation of at least one traffic sign in the acquired image;   accessing a crowd-sourced map generated based on drive information collected from a plurality of vehicles that previously traversed the road segment, wherein the crowd-sourced map stores indicators of relevancy of mapped traffic signs to certain drivable paths stored in the crowd-sourced map;   determining, based on the accessed crowd-sourced map, whether the at least one traffic sign detected in the acquired image is relevant to a drivable path along which the host vehicle is traveling; and   in response to a determination that the at least one traffic sign detected in the acquired image is relevant to a drivable path along which the host vehicle is traveling, causing the host vehicle to take at least one navigational action relative to the at least one traffic sign.