Atrenta, Inc.

Synopsys, Inc.

Waymo LLC

Maher Mneimneh

Mohammad Homayoun Movahed-Ezazi

Paras Mal Jain

Mohamed Shaker Sarwary

Namit Gupta

Shaker Sarwary

Ashish Bansal

Kai Ding

Wei Chai

Haoyu Chen

Fahim Rahim

Sudeep Mondal

Deepak Ahuja

Minfa Wang

Malay K Ganai

Scott Cotton

Anne Hobbs Dorsey

Qiaojing Yan

Pronay Kumar Biswas

Nishant Gupta

Mohamad Shaker Sarwary

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing knowledge distillation for autonomous vehicles. One of the methods includes obtaining sensor data characterizing an environment, wherein the sensor data has been captured by one or more sensors on-board a vehicle in the environment; processing, for each of one or more surrounding agents in the environment, a network input generated from the sensor data using a neural network to generate an agent discomfort prediction that characterizes a level of discomfort of the agent; combining the one or more agent discomfort predictions to generate an aggregated discomfort score; and providing the aggregated discomfort score to a path planning system of the vehicle in order to generate a future path of the vehicle.

Knowledge distillation for autonomous vehicles

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, that determine yield behavior for an autonomous vehicle. An agent that is in a vicinity of an autonomous vehicle can be identified. An obtained crossing intent prediction characterizes a predicted likelihood that the agent intends to cross a roadway during a future time period. First features of the agent and of the autonomous vehicle are obtained. An input that includes the first features and the crossing intent prediction is processed using a machine learning model to generate an intent yielding score that represents a likelihood that the autonomous vehicle should perform a yielding behavior due to the intent of the agent to cross the roadway. From at least the intent yielding score, an intent yield behavior signal is determined and indicates whether the autonomous vehicle should perform the yielding behavior prior to reaching the first crossing region.

Pedestrian intent yielding

A system and method for managing and composing verification engines and simultaneously applying such compositions to verify properties with design constraints allocates computing resources to verification engines based upon properties to be checked and optionally a user-specified budget. The verification engines are run in order to verify a received register transfer level (RTL) design description of a circuit according to user-specified assertions and constraints received by the system. The particular verification engines to be run are selected from a database of such engines and a run order is designated in sequential, parallel and distributed flows.

System and method for managing and composing verification engines

A system and method for netlist clock domain crossing verification leverages RTL clock domain crossing (CDC) verification data and results. The netlist clock domain crossing verification system (NCDC) migrates RTL-level constraints and waivers to the netlist design so that the user does not have to re-enter them. The NCDC checks the netlist and generates a report that compares RTL-level CDC checking results to the netlist-level CDC checking results to make it easy to see new issues. The NCDC receives and stores netlist corrections from user input or automatically corrects certain CDC violations, in the netlist.

System and method for netlist clock domain crossing verification

A method of hybrid clock domain crossing (CDC) verification includes receiving a design or an integrated circuit (IC) design constraints. Static CDC verification is performed, including structural and functional verification. The result is checked and explicit or implicit assumptions are made to signoff verification. Incomplete formal analysis results are discarded after review. Assertions and monitors are generated by this process to capture the assumptions and check partially covered properties by formal analysis. A dynamic simulation is run using a testbench, the generated assertions and the monitors. The static verification and dynamic verification processes may be repeated until a satisfactory coverage is obtained. A system, such as a computer aided design (CAD) system, is configured to perform CDC verification of the IC design. The system may generate assertions and monitors to then run a simulation and determine coverage. Results are then reiterated through the system back to the static CDC verification.

System and method for a hybrid clock domain crossing verification

A system and methods for metastability verification of an integrated circuit design are provided. An IC design is received and the source-to-destination paths of the IC design are determined. For each of the determined source-to-destination paths, it is determined whether the corresponding source is synchronized. For each source its respective synchronized or unsynchronized result is stored and a report is generated for each source describing whether it is synchronized or unsynchronized. The system may be at least a portion of a computer aided design (CAD) system.

System and method for metastability verification of circuits of an integrated circuit

A system and methods for reasonable formal verification provides a user with coverage information that is used for verification signoff. The coverage is calculated based on formal analysis techniques and is provided to the user in terms of design-centric metrics rather than formal-centric metrics. Design-centric metrics include the likes of a number of reads from or writes to memories and a number of bit changes for counters, among many others. Accordingly, a setup for failure (SFF) function and a trigger the failure (TTF) function take place. During SFF formal analysis is applied in an attempt to reach a set of states close enough to suspected failure states. During TTF formal analysis is applied, starting from the SFF states, to search for a state violating a predetermined property. If results are inconclusive the user is provided with a design-centric coverage metric that can be used in signoff.

System and methods for reasonable functional verification of an integrated circuit design

Clock-domain crossing (CDC) verification for system on chip (SoC) integrated circuits (IC) can be time consuming and complex, especially as the size of the SoC and the complexity of the modules of which it comprises increase. A bottom-up verification process includes the replacement of a CDC verified module by an abstracted model of the module with constraints defined on the boundaries of the module. Performing the process in a hierarchic manner from bottom upwards allows for faster verification of modules higher in the hierarchy as at least portions thereof are replaced with the abstracted modules.

Hierarchical bottom-up clock domain crossing verification

A method and system for metastability verification of an integrated circuit design are provided. An IC design is received and the source-to-destination paths of the IC design are determined. For each of the determined source-to-destination paths, it is determined whether the corresponding source is synchronized. For each source its respective synchronized or unsynchronized result is stored and a report is generated for each source describing whether it is synchronized or unsynchronized.

Growing community of inventors

San Jose, CA, United States of America

Maher Mneimneh