Systems and methods of determining trajectories of an actor in an environment in which a vehicle is operating are provided. The method includes detecting an actor that may move within a scene in the environment by an object detection system of a vehicle in the environment, determining a kinematic history of the actor, and using context of the scene and the kinematic history of the actor to determine a plurality of reference polylines for the actor. The method further includes generating a contextual embedding of the kinematic history of the actor to generate a plurality of predicted trajectories of the actor, in which the generating conditions each of the predicted trajectories to correspond to one of the reference polylines. The method additionally includes using, by the vehicle, the plurality of predicted trajectories to plan movement of the vehicle.

An object detection system in a surrounding environment of a vehicle. The detection system comprising, a radar, and a processing unit. The radar comprising a transmitter, a receiver, and an ultra-low phase-noise frequency synthesizer. The detection system gathers electromagnetic data of the objects from radio signal received by the receiver; classify each of the objects by analyzing the gathered electromagnetic data; continuously compare classifications and object detections to immediate past classifications and object detections, and to previously classified and detected objects; continuously validate current, immediate past, and past object detections; generate a three-dimensional electromagnetic-map of the surrounding environment by utilizing the electromagnetic signatures of each of the classified objects; and reclassify objects and combine the generated three-dimensional electromagnetic map with one of a geographical-map, a physical map, or a combination thereof to determine a direction and a distance of each of the one or more classified objects from the system.

A user identity determining method includes: acquiring target multidimensional feature information of a target user, wherein the target multidimensional feature information includes at least two types of feature information in at least one of biometric feature information or non-biometric feature information; comparing the target multidimensional feature information with multidimensional feature information of a plurality of designated users, respectively, to obtain a comparison result; and determining an identity of the target user based on the comparison result.

The present disclosure relates to a gait recognition method based on deep learning, which comprises recognizing an identity of a person in a video according to the gait thereof through dual-channel convolutional neural networks sharing weights by means of the strong learning capability of the deep learning convolutional neural network. Said method is quite robust to gait changes across a large view, which can effectively solve the problem of low precision in cross-view gait recognition existing with the prior art gait recognition technology. Said method can be widely used in scenarios having video monitors, such as security monitoring in airports and supermarkets, person recognition, criminal detection, etc.

A computer-implemented method for gait analysis of a subject includes obtaining visual data from an image capture device positioned in front of or behind the subject, the visual data comprising at least two image frames of the subject over a period of time walking toward or away from the image capture device, the at least two image frames capturing at least a portion of the gait of the subject, detecting within the at least two images body parts as two-dimensional landmarks using a pose estimation algorithm on each of the at least two frames, generating a joint model depicting the location of the at least one joint in each of the at least two frames, using the joint model to segment a gait cycle for the at least one joint, and comparing the gait cycle to a threshold value to detect abnormal gait.

In the multi-target detection and tracking method, lidar (2D laser scanner) scans point cloud data of surroundings and transfers the collected data to the edge server. Then, the edge server uploads the data to the cloud. After obtaining the lidar data, point clouds of footsteps are extracted through dynamic point extraction, point clustering, and random forest model, respectively. Footsteps are matched to form human tracking trajectory by using trajectory matching. After the tracking process, the walking information is published to the users, in a visual form. Meanwhile, the gait parameters are saved into files, including walking speed and step length, when human is detected. Comparing to the visual sensor based human tracking methods, the present invention employs lidar to avoid the interference of ambient light, which leads to easier implementation and larger universality, especially for multi-target scenarios.

A device, method and system provide a wellness management process and/or an exercise management process for use with a smartphone or other mobile computing device. Various data about the user is obtained and used for determining and recommending an action or exercise to the user to improve the user's wellness/physique/health. The action determined can be based on: (1) current biometric and/or motion data about the user (from the sensors), and (2) current physical condition(s), such as health/medical information or condition about the user (from the user's personal information, e.g., health library or programmed into the smartphone). Specific information about the user is taken into consideration when recommending user action or exercise, such as the user's specific physical, health or medical conditions.

An example operation includes one or more of receiving, by a computer associated with a transport, a gait of an individual from at least one camera associated with the transport, validating, by the computer, the gait, receiving, by the computer, a gesture of the individual from the at least one camera, validating, by the computer, the gesture, and performing, by the computer, one or more functions based on the validated gait and the validated gesture.

A system for analysis of user gait and to provide correction in form of haptic and visual feedback. This system comprises a motion and force sensors and a haptic actuator embedded in the user shoe insoles in communication with a smart-phone based analysis application, configured to calculate motion and orientation of the user feet in relation to the value, location and distribution of ground reaction forces measured by sensors located in the shoe insoles and after analysis of said forces and motion, to provide haptic feedback to the user foot instructing about the location (and timing) of pressure the user must apply to achieve an optimal gait.

A system and method for a machine learning-based automated electronic time recording for personnel includes identifying, via a scene capturing device, a representation of a time recording space; identifying a body having a time recording pose within the time recording space based on an assessment of the representation of the time recording space; extracting a plurality of distinct features from the representation of the time recording space based on identifying the body having the time recording pose; executing automated user-recognition based on the extracting of the plurality of distinct features; executing automated time recording recognition based on the extracting of the plurality of distinct features; and executing automated electronic time recording, via a time recording application based on the automated user-recognition and the automated time recording recognition.