Among other things, techniques are described for predicting how an agent (e.g., a vehicle, bicycle, pedestrian, etc.) will move in an environment based on prior movement, the road network, the surrounding objects and/or other relevant environmental factors. One trajectory prediction technique involves generating a probability map for an agent's movement. Another trajectory prediction technique involves generating a trajectory lattice, for an agent's movement. In addition, a different trajectory prediction technique involves multi-modal regression where a classifier (e.g., a neural network) is trained to classify the probability of a number of (learned) modes such that each model produces a trajectory based on the current input.

An information generation device includes: a measurement unit configured to obtain a plurality of measurement results by performing a measurement of a vehicle size with respect to a same traveling vehicle a plurality of times; a detection unit configured to detect an accuracy of each of the plurality of measurement results; and a determination unit configured to determine a vehicle size of the traveling vehicle from the plurality of measurement results on the basis of the accuracies.

A radar signal processing device includes processing circuitry configured to repeatedly acquire a beat signal having a frequency of a difference between a frequency of a radar signal and a frequency of a reflected wave of the radar signal reflected by an observation target, repeatedly calculate a distance between a radar device and the observation target using the acquired beat signal, and repeatedly calculate a relative speed between the radar device and the observation target using the acquired beat signal; calculate an incident angle of the reflected wave to an array antenna by using the acquired beat signal and an arrangement interval between a plurality of reception antennas included in the array antenna; and determine whether the observation target is a detection target or a non-detection target due to electromagnetic noise based on the calculated incident angle, the plurality of distances and the plurality of relative speeds.

There is described a terahertz (THz) imaging device and method for imaging an object hidden underneath clothing. The THz imaging device is generally configured for illuminating a region of clothing with a THz illumination beam, collecting a return optical beam reflected from the region of clothing in response to the illumination, generating a THz image based on the collected return optical beam, and tracking a position and orientation of the THz imaging device as the THz image is generated. A plurality of complementary THz images are generated by repeating the steps of illuminating, collecting, generating and tracking for a plurality of points of view of the THz imaging device relative to the clothing. By registering these THz images to one another in a common coordinate system based on the tracked position and orientation, the THz imaging device can output a composite THz image representing the hidden object, if any.

A computer-implemented method for occupancy state detection in an area for a pre-determined point in time. In aspects, the computer-implemented method includes operations carried out by computer hardware components. The operations include determining a probability distribution over a list of possible occupancy states of the area at a previous point in time, determining measurement data related to the area at the pre-determined point in time, and determining a probability distribution over the list of possible occupancy states of the area at the pre-determined point in time based on the measurement data and the probability distribution over the list of possible occupancy states of the area at the previous point in time.

Systems and methods for accommodating flexibility in sensing transmissions are provided. Wi-Fi sensing systems include sensing devices and remote devices configured to communicate through radio-frequency signals. Initially, a sensing device transmits a sensing configuration message to a remote device. The sensing device receives a sensing configuration response message in response to the sensing configuration message. In an example, the sensing configuration response message may include a transmission capability indication associated with the remote device. The transmission capability indication includes a flexibility indication that the remote device supports flexibility.

A tracking system for tracking an expanded state of an object is provided. The tracking system comprises at least one processor and a memory having instructions stored thereon that, when executed by the at least one processor, cause the tracking system to execute a probabilistic filter that iteratively tracks a belief on the expanded state of the object, wherein the belief is predicted using a motion model of the object and is further updated using a compound measurement model of the object. The compound measurement model includes multiple probabilistic distributions constrained to lie on a contour of the object with a predetermined relative geometrical mapping to the center of the object. Further, the tracking system tracks the expanded state of the object based on the updated belief on the expanded state.

This disclosure describes a radar system configured to estimate a yaw-rate and an over-the-ground (OTG) velocity of extended targets in real-time based on raw radar detections. This disclosure further describes techniques for determining instantaneous values of lateral velocity, longitudinal velocity, and yaw rate of points of a rigid body in a radar field-of-view (FOV) of the radar system.

Upon each new detection, called pivot detection, by a radar system, the method includes the steps consisting of: grouping together, with the pivot detection, grouped detections, defined as detections that belong to a sweep preceding the sweep of the pivot detection and that have a non-nil probability according to a grouping criterion; filtering the grouped detections so as to keep only detections that are kinematically strictly coherent with the pivot detection, by: initializing a histogram, each dimension of which is a temporal variation of a coordinate measured by the radar system; computing a potential value interval for each coordinate of the pivot detection and each grouped detection; computing a minimum temporal variation and a maximum temporal variation for the or each coordinate from potential value intervals of the pivot detection and each grouped detection; incrementing the set of classes of the histogram whose index along each dimension is located between the computed minimum and maximum temporal variations; and detecting a target once at least one class of the histogram reaches a predefined value.

A method for obstacle detection and navigation of a vehicle using resolution-adaptive fusion includes performing, by a processor, a resolution-adaptive fusion of at least a first three-dimensional (3D) point cloud and a second 3D point cloud to generate a fused, denoised, and resolution-optimized 3D point cloud that represents an environment associated with the vehicle. The first 3D point cloud is generated by a first-type 3D scanning sensor, and the second 3D point cloud is generated by a second-type 3D scanning sensor. The second-type 3D scanning sensor includes a different resolution in each of a plurality of different measurement dimensions relative to the first-type 3D scanning sensor. The method also includes detecting obstacles and navigating the vehicle using the fused, denoised, and resolution-optimized 3D point cloud.