Robert Bosch

Naim, Iftekhar

Ramakrishnan, Naveen

Benitez, Diego

Naveen Ramakrishnan

Jayanta Kumar Dutta

Stefan Weissert

Ravi Kumar Satzoda

Yumi Kondo

Filipe Goncalves

Alexander Hirsch

Krishnan Bharath Navalpakkam

Ryan Burt

Thomas Bleile

Manuel Maier

Roland Klinnert

Diego Benitez

Marc Naumann

Jan Mathias Koehler

Michael Dambier

Rahul Kapoor

Chun-Hao Liu

Jochen Quante

Jeffrey L Irion

Sauptik Dhar

Jeff Irion

Luiz Ricardo Douat

Michael Erz

Iftekhar Naim

Karsten Thalheimer

Alexander Hagmeister

Frederik Brockmann

Gregor Wunder

Fabio Cecchi

Bernd Bandemer

Steffen Erkel

Baptiste Herve

Identifying key frames of a video for use in training a machine learning model is provided. Object detection is performed to identify frames of a video including target classes of objects of interest. Feature extraction is performed on the identified frames to generate raw feature vectors. The feature vectors are compressed into lower dimension vectors. The compressed feature vectors are compressed into a plurality of clusters. The clustered compressed feature vectors are filtered to identify the key frames from each of the plurality of clusters. The key frames may be provided as a representative data set of the video.

Automated key frame selection

A vehicle for collecting image data of a target object for generating a classifier. The vehicle includes an image sensor and an electronic processor. The electronic processor is configured to determine a plurality of potential trajectories of the vehicle, determine, for each of the plurality of potential trajectories of the vehicle, a total number of views including the target object that would be captured by the image sensor as the vehicle moved along the respective trajectory, and determine a key trajectory of the vehicle from the plurality of potential trajectories based on the total number of views including the target of the key trajectory.

Trajectory design for image data acquisition for object detection/recognition

A method of solving a graph simultaneous localization and mapping (graph SLAM) for HD maps using a computing system includes partitioning a graph into a plurality of subgraphs, each of the subgraphs having all of the vertices of the graph and a subset of the edges of the graph. Constrained and non-constrained vertices are defined for the subgraphs. An alternating direction method of multipliers (ADMM) formulation for Graph SLAM is defined using the partitioned graph. A distributed Graph SLAM algorithm is then defined in terms of the constrained and non-constrained vertices based on the ADMM formulation. The distributed Graph SLAM algorithm is then used to solve the Graph SLAM problem for HD maps.

Scalable graph SLAM for HD maps

A method and system are disclosed for monitoring passengers in within a cabin of a vehicle and determining whether the passengers are engaging in abnormal behavior. The method and system uses a novel vector to robustly and numerically represent the activity of the passengers in a respective frame, which is referred to herein as an 'activity vector.' Additionally, a Gaussian Mixture Model is utilized by the method and system to distinguish between normal and abnormal passenger behavior. Cluster components of the Gaussian Mixture Model are advantageously learned using an unsupervised approach in which training data is not labeled or annotated to indicate normal and abnormal passenger behavior. In this way, the Gaussian Mixture Model can be trained at a very low cost.

System and method for detecting abnormal passenger behavior in autonomous vehicles

Methods and systems for hierarchy detection for block diagrams. One system includes an electronic processor configured to access a block diagram. The electronic processor is also configured to identify a set of connected components in the block diagram. The electronic processor is also configured to convert a first connected component included in the set of connected components into a directed acyclic graph (DAG). The electronic processor is also configured to determine a set of candidate hierarchies included in the DAG. The electronic processor is also configured to verify the set of candidate hierarchies. The electronic processor is also configured to generate a displayable hierarchical block diagram based on the verified set of candidate hierarchies.

Hierarchy detection for block diagrams

A method of operating a community parking system includes generating sensor data with a plurality of vehicles corresponding to vehicle parking spaces and parked vehicles located in a region, transmitting the sensor data to a parking data system, and generating parking map data with the parking data system based on the sensor data, the parking map data including a location of the vehicle parking spaces. The method also includes generating parking service data with the parking data system based on the sensor data, the parking service data identifying a status of each vehicle parking space as either occupied or unoccupied, and transmitting the parking service data to a particular vehicle of the plurality of vehicles to assist an operator of the particular vehicle in locating one of the vehicle parking spaces having an unoccupied status.

System and method for distributed parking area map generation and parking area service using in-vehicle sensors

A method for event identification in video data includes identifying a feature vector having data corresponding to at least one of a position and a direction of movement of an object in video data, generating an estimated feature vector corresponding to the feature vector using a dictionary including a plurality of basis vectors, identifying an error between the estimated feature vector and the feature vector, identifying a high-interest event in the video data in response to the identified error exceeding a threshold, and displaying the video data including the high-interest event on a video output device only in response to the error exceeding the threshold.

System and method for detection of high-interest events in video data

A method distinguishes electrical signals corresponding to an activation or deactivation of a single electrical power consuming device from an electrical signal that supplies electricity to multiple electrical power consuming devices has been developed. The method includes generating signatures corresponding to activation and deactivation of a plurality of electrical devices from a transformation of an electrical power signal using sparse deconvolution to cluster and then identify the activation and deactivation of one of the plurality of electrical devices.

Method for unsupervised non-intrusive load monitoring

A method of monitoring a position of a moveable entity includes equipping a moveable entity with a position sensor that outputs position signals indicating current geographical positions of the sensor, and using a machine learning system to process the position signals in accordance with a machine learning algorithm to identify reference positions indicated by the position signals corresponding to a first type of activity performed by the entity. Rules are defined based on the identified reference positions. The computer processor then monitors the position signals and apply the rules to the position signals to identify positions that violate the rules.

Intelligent electronic monitoring system

Growing community of inventors

Campbell, CA, United States of America

Naveen Ramakrishnan