Filtering vehicle radar returns for vehicle sensor calibration

Abstract

One or more radar sensors coupled to a vehicle receive readings during a calibration time period. Each radar sensor receives data covering a field of view of the sensor, which may be split into angle-based bins. A noise-reducing filter (e.g., median filter) may be applied. A function is generated by processing raw radar cross section (RCS) returns into values plotted against angles compared to the direction that the sensor is facing. The function may be smoothed. Radar sensor measurements captured after calibration are corrected using the function, for example by automatically subtracting or dividing amounts corresponding to the function.

What is claimed is: 
     
         1 . A system for vehicle sensor calibration, the system comprising:
 a range sensor coupled to a vehicle that is located in a calibration environment, wherein the calibration environment also includes one or more sensor targets that are detectable by the range sensor, wherein the range sensor captures a plurality of sensor calibration datasets during a calibration time period;   a memory storing instructions; and   a processor that executes the instructions, wherein execution of the instructions by the processor causes the processor to:
 receive the plurality of sensor calibration datasets from the range sensor during the calibration time period, wherein the plurality of sensor calibration datasets include data corresponding to the one or more sensor targets and also include noise, 
 determine one or more filters such that application of the one or more filters to the plurality of sensor calibration datasets modifies the plurality of sensor calibration datasets by reducing noise while maintaining the data corresponding to the one or more sensor targets; and 
 calibrate interpretation of subsequent data captured by the range sensor after the calibration time period such that interpretation of the subsequent data includes application of the one or more filters to the subsequent data. 
   
     
     
         2 . The system of  claim 1 , wherein the range sensor includes a radio detection and ranging (RADAR) sensor. 
     
     
         3 . The system of  claim 1 , wherein the range sensor includes at least one of a light detection and ranging (LIDAR) sensor or a laser rangefinder. 
     
     
         4 . The system of  claim 1 , wherein the range sensor includes at least one of a sound navigation and ranging (SONAR) sensor or a sound detection and ranging (SODAR) sensor. 
     
     
         5 . The system of  claim 1 , wherein the one or more filters discard data falling outside of a defined range of distances from the vehicle. 
     
     
         6 . The system of  claim 1 , wherein the one or more filters discard data falling outside of a defined range of angles within a field of view of the range sensor. 
     
     
         7 . The system of  claim 1 , wherein the one or more filters discard data falling below a defined minimum return strength threshold. 
     
     
         8 . The system of  claim 1 , wherein the one or more filters discard data that does not match known information about the one or more sensor targets. 
     
     
         9 . The system of  claim 8 , wherein the one or more filters discard data that does not match known positions of the one or more sensor targets. 
     
     
         10 . The system of  claim 8 , wherein the one or more filters keep data that matches one or more known distances between targets of two or more of the one or more sensor targets, and discard data that does not match the one or more known distances between targets of two or more of the one or more sensor targets. 
     
     
         11 . The system of  claim 1 , wherein the one or more filters apply a clustering algorithm to generate one or more clusters from data from the range sensor, and wherein the one or more filters discard data that is not clustered by the clustering algorithm. 
     
     
         12 . The system of  claim 11 , wherein the clustering algorithm is a density-based spatial clustering of applications with noise (DBSCAN) algorithm, and wherein the one or more filters keep data corresponding to one or more densest clusters of the one or more clusters, and discard remaining data. 
     
     
         13 . A method for vehicle sensor calibration, the method comprising:
 receiving a plurality of sensor calibration datasets from a range sensor of a vehicle during a calibration time period, the vehicle located in a calibration environment that also includes one or more sensor targets that are detectable by the range sensor, wherein the plurality of sensor calibration datasets include data corresponding to the one or more sensor targets and also include noise;   determining one or more filters such that application of the one or more filters to the plurality of sensor calibration datasets modifies the plurality of sensor calibration datasets by reducing noise while maintaining the data corresponding to the one or more sensor targets; and   calibrating interpretation of subsequent data captured by the range sensor after the calibration time period such that interpretation of the subsequent data includes application of the one or more filters to the subsequent data.   
     
     
         14 . The method of  claim 13 , wherein range sensor includes a radio detection and ranging (RADAR) sensor. 
     
     
         15 . The method of  claim 13 , wherein range sensor includes at least one of a light detection and ranging (LIDAR) sensor, a laser rangefinder, a sound navigation and ranging (SONAR) sensor, or a sound detection and ranging (SODAR) sensor. 
     
     
         16 . The method of  claim 13 , wherein the one or more filters discard data falling outside of a defined range, the defined range describing one of distances from the vehicle or angles within a field of view of the range sensor. 
     
     
         17 . The system of  claim 1 , wherein the one or more filters discard data falling below a defined minimum return strength threshold. 
     
     
         18 . The system of  claim 1 , wherein the one or more filters discard data that does not match known positions of the one or more sensor targets. 
     
     
         19 . The system of  claim 8 , wherein the identifying the data that does not match known positions of the one or more sensor targets is performed based on application of a density-based spatial clustering of applications with noise (DBSCAN) algorithm. 
     
     
         20 . A non-transitory computer readable storage medium having embodied thereon a program, wherein the program is executable by a processor to perform a method of vehicle sensor calibration, the method comprising:
 receiving a plurality of sensor calibration datasets from a range sensor of a vehicle during a calibration time period, the vehicle located in a calibration environment that also includes one or more sensor targets that are detectable by the range sensor, wherein the plurality of sensor calibration datasets include data corresponding to the one or more sensor targets and also include noise;   determining one or more filters such that application of the one or more filters to the plurality of sensor calibration datasets modifies the plurality of sensor calibration datasets by reducing noise while maintaining the data corresponding to the one or more sensor targets; and   calibrating interpretation of subsequent data captured by the range sensor after the calibration time period such that interpretation of the subsequent data includes application of the one or more filters to the subsequent data.