Gm Cruise Holdings LLC

Nvidia Corporation

Sandeep Gangundi

Karthikeyan Natarajan

Sarthak Sahu

Abdelrahman Elogeel

Manjunatha Gowda

Kaushik Narayanun

Lijuan Luo

Harbinder Sikka

Nathan Harada

Phil Ferriere

Kyle Vogt

Raluca Musaloiu-Elefteri

Cornelis Jacob Lekkerkerker

Mayank Ketkar

Wanlin Yang

Zachary Witte

Avinash Nair

Craig Quiter

Siddhartho Bhattacharya

The disclosed technology provides methods for training of a convolutional neural network (CNN) to identify or predict its own errors, and then those errors are used as inputs with feature visualization to generate images of scenes associated with those errors. This allows adjustment of a set of labeled training images, and then the adjusted set of labeled training images are used to retrain or further train the CNN. Systems and machine-readable media are also provided.

System and method for feature visualization in a convolutional neural network

The subject disclosure relates to techniques for managing the configuration of vehicles. A process of the disclosed technology can include steps of receiving a first vehicle identifier for a first vehicle, wherein the first vehicle is associated with a first vehicle platform, identifying a first configuration for the first vehicle platform based on the first vehicle identifier, receiving a second vehicle identifier for a second vehicle, wherein the second vehicle is associated with a second vehicle platform, and wherein the first vehicle platform is different from the second vehicle platform, and identifying a second configuration for the second vehicle platform based on the second vehicle identifier, wherein the first vehicle and the second vehicle run on an identical software stack. Systems and machine-readable media are also provided.

Configuration management system for autonomous vehicle software stack

Distance measurements are received from one or more distance measurement sensors, which may be coupled to a vehicle. A three-dimensional (3D) point cloud are generated based on the distance measurements. In some cases, 3D point clouds corresponding to distance measurements from different distance measurement sensors may be combined into one 3D point cloud. A voxelized model is generated based on the 3D point cloud. An object may be detected within the voxelized model, and in some cases may be classified by object type. If the distance measurement sensors are coupled to a vehicle, the vehicle may avoid the detected object.

Object detection based on three-dimensional distance measurement sensor point cloud data

The subject disclosure relates to ways to identify self-hit data collected by autonomous vehicle (AV) sensors. In some aspects, a method of the disclosed technology includes steps for generating a geometric model of an autonomous vehicle (AV), wherein the geometric model specifies physical boundaries of the AV in three-dimensional (3D) space, collecting sensor data for an environment around the AV, and identifying one or more data points, from among the collected sensor data, that correspond with a surface of the AV. Systems and machine-readable media are also provided.

Eliminating sensor self-hit data

Techniques to improve the accuracy and speed for detection and remediation of difficult to test nodes in a circuit design netlist. The techniques utilize improved netlist representations, test point insertion, and trained neural networks.

Deep learning based identification of difficult to test nodes

Growing community of inventors

Milpitas, CA, United States of America

Sandeep Gangundi