Detecting Spurious Objects For Autonomous Vehicles

Abstract

Aspects of the disclosure relate to detecting spurious objects. For instance, a model may be trained using raining data including a plurality of LIDAR data points generated by a LIDAR sensor of a vehicle. Each given LIDAR data point includes location information and intensity information, and is associated with waveform data for that given LIDAR data point. At least one of the plurality of LIDAR data points is further associated with a label identifying spurious objects through which the vehicle is able to drive. The model and/or a plurality of heuristics may then be provided to a vehicle in order to allow the vehicle to determine LIDAR data points that correspond to spurious objects. These LIDAR data points may then be filtered from sensor data, and the filtered sensor data may be used to control the vehicle in an autonomous driving mode.

1 . A method of controlling a vehicle, the method comprising:
 receiving, by one or more computing devices, a plurality of LIDAR data points generated by a LIDAR sensor of the vehicle, each generated LIDAR data point of the plurality of LIDAR data points (1) including location information and intensity information and (2) being associated with waveform data other than the intensity information of the generated LIDAR data point;   determining, by the one or more computing devices, that one or more LIDAR data points of the plurality of LIDAR data points correspond to a spurious object using the waveform data associated with the plurality of LIDAR data points;   determining, by the one or more computing devices, a type of the spurious object using a model with the waveform data associated with the determined one or more LIDAR data points as input; and   controlling, by the one or more computing devices, the vehicle in an autonomous driving mode based on the determination of the type of the spurious object.   
     
     
         2 . The method of  claim 1 , wherein the spurious object is distinguishable from a solid object and through which the vehicle is able to drive. 
     
     
         3 . The method of  claim 1 , wherein the determining that the one or more LIDAR data points of the plurality of LIDAR data points correspond to a spurious object is further based on one or more heuristics. 
     
     
         4 . The method of  claim 3 , wherein the one or more heuristics is based on a peak elongation of waveform data for a particular LIDAR data point. 
     
     
         5 . The method of  claim 3 , wherein the one or more heuristics is based on a number of peaks in waveform data for a particular LIDAR data point. 
     
     
         6 . The method of  claim 3 , wherein the one or more heuristics is based on a peak width in waveform data for a particular LIDAR data point. 
     
     
         7 . The method of  claim 3 , wherein a particular LIDAR data point of the plurality of LIDAR data points is further associated with a peak elongation relative to an expected peak width, and wherein the peak elongation is used to determine whether the LIDAR data point corresponds to the spurious object through which the vehicle is able to drive. 
     
     
         8 . The method of  claim 1 , wherein the waveform data associated with a particular LIDAR data point of the plurality of LIDAR data points includes a plurality of samples of light received at the LIDAR sensor for the particular LIDAR data point. 
     
     
         9 . The method of  claim 1 , wherein the frame corresponds to one 360 degree rotation of the LIDAR sensor. 
     
     
         10 . The method of  claim 1 , wherein the plurality of LIDAR data points corresponds to a single frame captured by the LIDAR sensor. 
     
     
         11 . The method of  claim 1 , wherein the type of the spurious object is vehicle exhaust. 
     
     
         12 . The method of  claim 1 , wherein the type of the spurious object is dust. 
     
     
         13 . The method of  claim 1 , wherein the type of the spurious object is fog. 
     
     
         14 . The method of  claim 1 , wherein the type of the spurious object is rain. 
     
     
         15 . The method of  claim 1 , wherein the type of the spurious object is snow. 
     
     
         16 . The method of  claim 1 , wherein, when the type of the spurious object is dust or fog, the controlling the vehicle includes slowing down the vehicle. 
     
     
         17 . The method of  claim 1 , wherein, when the type of the spurious object is rain, the controlling the vehicle includes turning on wipers of the vehicle. 
     
     
         18 . A system for controlling a vehicle, the system comprising one or more processors configured to:
 receive a plurality of LIDAR data points generated by a LIDAR sensor of the vehicle, each generated LIDAR data point of the plurality of LIDAR data points (1) including location information and intensity information and (2) being associated with waveform data other than the intensity information of the generated LIDAR data point;   determine that one or more LIDAR data points of the plurality of LIDAR data points correspond to a spurious object using the waveform data associated with the plurality of LIDAR data points;   determine a type of the spurious object using a model with the waveform data associated with the one or more LIDAR data points as input; and   control the vehicle in an autonomous driving mode based on the determination of the type of the spurious object.   
     
     
         19 . The system of  claim 18 , wherein, when the type of the spurious object is dust or fog, the system is configured to slow down the vehicle. 
     
     
         20 . The system of  claim 18 , wherein, when the type of the spurious object is rain, the system is configured to turn on wipers of the vehicle.