Using prediction models for scene difficulty in vehicle routing

Abstract

A route is selected for travel by an autonomous vehicle based on at least a level of difficulty of traversing the driving environment along that route. Vehicle signals, provided by one or more autonomous vehicles, indicating a difficulty associated with traveling a portion of a route are collected and used to predict a most favorable driving route for a given time. The signals may indicate a probability of disengaging from autonomous driving mode, a probability of being stuck for an unduly long time, traffic density, etc. A difficulty score may be computed for each road segment of a route, and then the scores of all of the road segments of the route are added together. The scores are based on number of previous disengagements, previous requests for remote assistance, unprotected left or right turns, whether parts of the driving area are occluded, etc. The difficulty score is used to compute a cost for a particular route, which may be compared to costs computed for other possible routes. Based on such information, a route may be selected.

1 . A method, comprising:
 receiving, by one or more computing devices, information from one or more vehicles configured to operate in an autonomous driving mode, the information identifying different types of difficulties encountered by the one or more vehicles while operating along one or more routes in the autonomous driving mode, wherein the different types of difficulties include at least one of a sensor difficulty or a roadway condition impacting autonomous driving;   storing the information in memory along with metadata that identifies one or more details regarding each one of the different types of difficulties; evaluating, by the one or more computing devices, the stored information and the metadata in view of a plurality of route options; and   selecting, by the one or more computing devices, a given one of the plurality of route options for use by a vehicle when operating in the autonomous driving mode.   
     
     
         2 . The method of  claim 1 , wherein the information includes a vehicle signal identifier for each one of the different types of difficulties that identifies a specific signal sent by a respective one of the one or more vehicles. 
     
     
         3 . The method of  claim 2 , wherein the vehicle signal identifier is either (i) included with the specific signal sent by the respective vehicle, (ii) assigned by the one or more computing devices, or (iii) composed from a combination of information from the respective vehicle and the one or more computing devices. 
     
     
         4 . The method of  claim 1 , wherein the metadata identifies at least one of a time each one of the different types of difficulties was encountered, or location information identifying a location at which each one of the different types of difficulties was encountered. 
     
     
         5 . The method of  claim 4 , wherein the location information includes at least one of a regional identifier, a street identifier, or a road segment identifier. 
     
     
         6 . The method of  claim 4 , wherein the location information includes geographical coordinates. 
     
     
         7 . The method of  claim 1 , wherein the metadata further includes a difficulty value for each one of the different types of difficulties that indicates a level of difficulty imposed by a driving environment. 
     
     
         8 . The method of  claim 7 , further comprising weighting the difficulty values based on temporal information associated with each one of the different types of difficulties. 
     
     
         9 . The method of  claim 1 , further comprising generating a prediction model identifying which road segments are likely to present difficulty based on the information stored in the memory. 
     
     
         10 . The method of  claim 1 , wherein the information stored in the memory includes a signal type indicating a type of difficulty encountered by each of the one or more vehicles that transmitted corresponding information. 
     
     
         11 . The method of  claim 10 , wherein the type of difficulty is one of an extended wait to make a turn or other maneuver, an unprotected turn, an occluded sensor view, a rapid brake application, a limited sensor view, an unprotected lane crossing, or a driving difficulty resulting from a driving environment. 
     
     
         12 . The method of  claim 10 , wherein the type of difficulty is one of a user complaint, measured passenger discomfort, a fault response, a need for a human driver to take manual control, or a request for assistance from a remote operator. 
     
     
         13 . The method of  claim 1 , further comprising transmitting the selected route option to a particular vehicle for operating in the autonomous driving mode. 
     
     
         14 . A method of operating a vehicle in an autonomous driving mode, the method comprising:
 controlling, by one or more processors, driving operations of the vehicle in the autonomous driving mode;   receiving, by the one or more processors, sensor information from a perception system of the vehicle regarding objects or conditions in an external environment of the vehicle;   generating, by the one or more processors, information regarding a difficulty experienced while operating in the autonomous driving mode, the information identifying at least one type of difficulty encountered by the vehicle while operating along one or more routes in the autonomous driving mode, wherein the at least one type of difficulty includes at least one of a sensor difficulty for a sensor of the perception system or a roadway condition impacting autonomous driving; and   transmitting the generated information to a remote system.   
     
     
         15 . The method of  claim 14 , wherein the generated information includes a vehicle signal identifier for each one of the at least one type of difficulty. 
     
     
         16 . The method of  claim 14 , wherein the generated information includes one or both of a time or a location at which each of the at least one type of difficulty was encountered. 
     
     
         17 . The method of  claim 16 , wherein the location includes at least one of a regional identifier, a street identifier, or a road segment identifier. 
     
     
         18 . The method of  claim 14 , wherein the at least one type of difficulty encountered by the vehicle is one of an extended wait to make a turn or other maneuver, an unprotected turn, an occluded sensor view, a rapid brake application, a limited sensor view, an unprotected lane crossing, or a driving difficulty resulting from a driving environment. 
     
     
         19 . The method of  claim 14 , wherein the at least one type of difficulty encountered by the vehicle is one of a user complaint, measured passenger discomfort, a fault response, a need for a human driver to take manual control, or a request for assistance from a remote operator. 
     
     
         20 . The method of  claim 14 , further comprising:
 receiving, from the remote system, a routing signal based on difficulty information obtained from one or more other vehicles; and   controlling, by the one or more processors, another driving operation of the vehicle in the autonomous driving mode according to the received routing signal.