This document describes techniques for using occlusion constraints for resolving tracks from multiple types of sensors. In aspects, an occlusion constraint is applied to an association between a radar track and vision track to indicate a probability of occlusion. In other aspects, described are techniques for a vehicle to refrain from evaluating occluded radar tracks and vision tracks collected by a perception system. The probability of occlusion is utilized for deemphasizing pairs of radar tracks and vision tracks with a high likelihood of occlusion and therefore, not useful for tracking. The disclosed techniques may provide improved perception data more closely representing multiple complex data sets for a vehicle for preventing a collision with an occluded object as the vehicle operates in an environment.

Deep learning to improve or gauge the performance of a surface-penetrating radar (SPR) system for localization or navigation. A vehicle may employ a terrain monitoring system including SPR for obtaining SPR signals as the vehicle travels along a route. An on-board computer including a processor and electronically stored instructions, executable by the processor, may analyze the acquired SPR images and computationally identify subsurface structures therein by using the acquired image as input to a predictor that has been computationally trained to identify subsurface structures in SPR images.

Disclosed is an obstacle detection system of a vehicle. The obstacle detection system includes a driving information unit configured to calculate driving position information of the vehicle, a determiner configured to anticipate whether or not the vehicle will enter a joining point where the vehicle meets a target road to be joined based on the driving position information calculated by the driving information unit, a sensing unit configured to sense obstacles located beside the vehicle, and a controller configured to change a sensing range of the sensing unit so as to detect an obstacle moving on the target road to be joined, when the determiner anticipates that the vehicle will enter the joining point.

The described aspects and implementations enable fast and accurate object identification in autonomous vehicle (AV) applications by combining radar data with camera images. In one implementation, disclosed is a method and a system to perform the method that includes obtaining a radar image of a first hypothetical object in an environment of the AV, obtaining a camera image of a second hypothetical object in the environment of the AV, and processing the radar image and the camera image using one or more machine-learning models MLMs to obtain a prediction measure representing a likelihood that the first hypothetical object and the second hypothetical object correspond to a same object in the environment of the AV.

A method is provided for determining a road condition by using a radar system having transmitter and receiving units for transmitting and receiving radar waves having two different polarizations and providing transmit and receive signals indicating an intensity of the transmitted and received radar waves. Co-polarized backscattering coefficients and at least one cross-polarized backscattering coefficient are determined based on the transmit and receive signals. If the cross-polarized backscattering coefficient is greater than or equal to a threshold, the road condition is determined based on a ratio of the co-polarized backscattering coefficients and based on a difference of one of the co-polarized backscattering coefficients and the cross-polarized backscattering coefficient. If the cross-polarized backscattering coefficient is smaller than the threshold, the road condition is determined based on the ratio and a difference of the co-polarized backscattering coefficients.

Embodiments herein can determine an optimal route for an autonomous electric vehicle. The system may score viable routes between the start and end locations of a trip using a numeric or other scale that denotes how viable the route is for autonomy. The score is adjusted using a variety of factors where a learning process leverages both offline and online data. The scored routes are not based simply on the shortest distance between the start and end points but determine the best route based on the driving context for the vehicle and the user.

Route planning is provided to a vehicle (or vehicle driver) based at least in part on the type of vehicle and the terrain conditions between the originating location and destination. The vehicle may employ a terrain monitoring system including a surface-penetrating radar (SPR) system for obtaining SPR signals as the vehicle travels along a route; the obtained SPR signals may be used for navigation against reference images associated with the route. In some embodiments, a navigation server bases route selection in part on the terrain associated with various routes and characteristics of the vehicle.

According to an aspect of the present disclosure, a collision warning apparatus of a vehicle may include an information acquisition device that obtains surrounding object information and vehicle information and a controller that generates collision prediction information of a surrounding object based on the surrounding object information and the vehicle information and provides a warning to an outside of the vehicle or generates control information for controlling braking of the vehicle while providing the warning to the outside of the vehicle based on the collision prediction information.

A polarimetric radar system for object classification and road condition estimation includes a radar transmitter unit for transmitting radar waves of different polarizations, a radar receiving unit for receiving radar waves of different polarizations, a radar signal generating unit for generating and providing the radar waves to be transmitted, a signal processing circuitry for processing the generated and received radar waves, and a signal evaluation unit. The signal evaluation unit receives processed signals from the signal processing circuitry, estimates values for a set of predetermined object parameters on the basis of the received processed signals, and selects an object class from a plurality of predetermined object classes upon detecting a match of the estimated values with one out of a plurality of predetermined sets of object parameters. The signal evaluation unit is configured to provide information that is indicative of the at least one classified object.

A method and apparatus with vehicle radar control is disclosed. An apparatus with vehicle radar control includes a radio frequency (RF) transceiver including a transmitting antenna array and a receiving antenna array, and at least one processor configured to collect environmental information of the vehicle, determine a radar mode of the vehicle based on the collected environmental information, generate one or more control signal configured to control one or more of the transmitting antenna array and the receiving antenna array based on the determined radar mode, and provide the generated one or more control signals to the RF transceiver, wherein one or more of the transmitting antenna array and the receiving antenna array operate according to the one or more generated control signals.