Denso International America, Inc.

University of Michigan

Wei Zhang

Joseph Lull

Rajesh Kumar Malhan

Xueding Wang

Kayvan Najarian

Paul Langford Carson

Jonathan Gryak

Chris Mi

Neda Masoud

Ibrahim Oraiqat

Jeff C White

Heming Yao

Ethan Zhang

Issam El Naqa

Yannis Paulus

Jean Moran

Kyle Cuneo

A system for predicting aggressive driving behavior for a driver of a vehicle includes a first edge computing device that can acquire spatial-temporal data for the vehicle from one or more sensors that are part of traffic infrastructure. The first edge computing device includes a processor and instructions executable by the processor that execute deep learning methods on the data from the sensors to cluster the data as a driving score. A trained model is applied to the driving score to determine an aggressive driving behavior risk level, and the first edge computing device is configured to predict the aggressive driving behavior based on the aggressive driving behavior risk level.

System for predicting aggressive driving

A method and system for performing online adapted radiotherapy are provided using combined ultrasound and ionizing radiation induced acoustic imaging (iRAI) computed tomography imaging techniques that can be used for measurement of low to ultrahigh dose deliveries (>40 Gy/s). Multiplexed transducers detect US and iRAI signals allowing for anatomical/functional imaging and radiation mapping with absolute dosimetry measurements of a region of interest during a radiotherapy session. Corrections to radiation dosage intensities and locations is determined and provided as feedback to a radiation source to improve the accuracy of applied radiation dosages intra- or inter-radiotherapy treatment sessions preventing the irradiation of healthy tissues and ensuring the accurate delivery of radiation to a tumor or region of interest.

Combined radiation acoustics and ultrasound for radiotherapy guidance and cancer targeting

The present invention relates to systems for low-energy (e.g., 1.0 nJ-7.0 nJ) photoacoustic microscopy and methods for employing such systems. In certain embodiments, such systems employ a low-energy nanosecond pulsed laser beam (NPLB), at least two amplifiers, and a data acquisition system with at least three channels to generate at least three digital signals (e.g., which are averaged and normalized to the energy of the NPLB). In other embodiments, provided herein are systems for combined use of photoacoustic microscopy, dye-based microscopy (e.g., with fluorescein), and optical coherence tomography.

Low energy photoacoustic microscopy (PAM) and combined PAM, dye-based microscopy, and optical coherence tomography

A method for guiding a user operating a requesting device includes identifying requesting parameters of the requesting device, where the requesting parameters include a location of a requesting device, a trajectory of a requesting device, or a combination thereof. The method includes generating one or more safety scores associated with one or more roadways based on vehicle behavior information associated with one or more vehicles, pedestrian behavior information associated with one or more roadside unit devices, or a combination thereof. The method includes determining a route of the requesting device based on the one or more safety scores and the requesting parameters and broadcasting the route to the requesting device.

Methods and systems for guiding road users

A system for predicting and interpreting bad driving behavior of a vehicle includes a first edge computing device that can acquire spatial-temporal data for the vehicle from one or more sensors that are part of traffic infrastructure. The first edge computing device includes a processor and instructions executable by the processor that execute unsupervised deep learning methods on the data from the sensors to cluster the data into segments and integrate a language model with the deep learning method to output the bad driving behavior in a natural language. The instructions further include normalizing the data, processing the data with a first artificial neural network (ANN) to output a first vector, processing the clustered data segments with a second ANN to output a second vector, concatenating the vectors into a single vector, and processing the single vector with a third ANN to output a predicted bad driving behavior of the vehicle.

System and method for predicting and interpreting driving behavior

A system for predicting and interpreting driving behavior of a vehicle includes a first edge computing device that can acquire spatial-temporal data for the vehicle from one or more sensors that are part of traffic infrastructure. The first edge computing device includes a processor and instructions executable by the processor that execute unsupervised deep learning methods on the data from the sensors to cluster the data into segments and integrate a language model with the deep learning method to output driving behavior in a natural language. The instructions further include normalizing the data, processing the data with a first artificial neural network (ANN) to output a first vector, processing the clustered data segments with a second ANN to output a second vector, concatenating the vectors into a single vector, and processing the single vector with a third ANN to output a predicted driving behavior of the vehicle.

An automated pruning technique is proposed for reducing the size of a convolutional neural network. A large-sized network is trained and then connections between layers are explored to remove redundant parameters. Specifically, a scaling neural subnetwork is connected to the neural network and designed to infer importance of the filters in the neural network during training of the neural network. Output from the scaling neural subnetwork can then be used to remove filters from the neural network, thereby reducing the size of the convolutional neural network.

Automatic filter pruning technique for convolutional neural networks

In an example, an image classification system is disclosed. The image classification system modifies an image having a first height and a first width to be input to a convolutional neural network for image classification. The image classification system includes an image resizing module that is configured to resize the image so that the resized image comprises a second height and a second width. An aspect ratio of the resized image corresponds to an aspect ratio of the image having the first height and the first width. The image classification system also includes an alignment module that is configured to modify pixels of a feature map corresponding to the resized image based upon a comparison of a desired feature map size and an actual feature map size.

Image classification system for resizing images to maintain aspect ratio information

A wireless charging system may be used to charge a battery in a vehicle via a receiving coil. The wireless charging system may include a coil charge device, a linear track, a linear motor, and a charge control module. The coil charge device includes a carriage and a transmitting coil positioned on the carriage. The linear track extends across a designated path. The coil charge device is positioned on and moveable along the linear track. The linear motor is operable to move the coil charge device along the linear track. The charge control module controls a position of the coil charge device along the designated path via the linear motor.

Growing community of inventors

Novi, MI, United States of America

Wei Zhang