Physics modeling for radar and ultrasonic sensors

摘要

A machine learning module 72 generates a probability distribution (90) from training data 42 including labeled modeling data correlated with reflection data. The modeling data includes data from a LIDAR system (20a), camera system (24a), or a GPS for a target environment/object. Reflection data may be collected from the same environment/object by a radar (38a) or ultrasonic (40a) system. The probability distribution assigns reflection coefficients for radar or ultrasonic systems conditioned on values for the modeling data. The machine learning module first learns a mapping from 3D modelling information to physical modelling characteristics such as radar or ultra-sonic data. The 3D information relates to data from camera, LIDAR or local-awareness (e.g. GPS) systems. Additionally, a test bench may evaluate an algorithm, for processing reflection data to generate control signals to an autonomous vehicle (34a), with simulated reflection data from a virtual sensor engaging reflection values assigned within the virtual environment.