Download as PDF, PPTX
Uploaded byinside-BigData.com
Nvidia needsfor atonomousvehicles
The document discusses the comprehensive end-to-end system required for developing autonomous vehicles, emphasizing the importance of data collection, model training, and simulation. It highlights NVIDIA's technological advancements in artificial intelligence and deep learning, explaining how extensive data labeling and simulation are crucial for improving self-driving capabilities. Key takeaways include understanding the full workflow of AI, the need for robust data center design, and the significance of following best practices for autonomous vehicle infrastructure.
More Related Content
Similar to Nvidia needsfor atonomousvehicles
![Coded Agents – with UiPath SDK + LangGraph [Virtual Hands-on Workshop]](https://cdn.slidesharecdn.com/ss_thumbnails/codedagentsdeck-251215155422-5497c599-thumbnail.jpg?width=640&height=640&fit=bounds)
Nvidia needsfor atonomousvehicles
- 1. NORM MARKS |SEPT.6, 2018 END TO END NEEDS FOR AUTONOMOUS VEHICLES
- 2. 2 THE MOST EXCITINGTIME IN TECH HISTORY NVIDIA GPU GAMING $100B Industry ARTIFICIAL INTELLIGENCE $3T IT Industry AUTONOMOUS VEHICLES $10T Transportation Industry
- 3. 3 END-TO-END SYSTEM FORAV COLLECT DATA TRAIN MODELS SIMULATE RE-SIMULATE MAPPING Lanes Lights Path Signs PedestriansCars Sources: NVIDIA & RAND Corporation
- 4. 4 Validate/ Verify Train on NVIDIA DGX Libraryof Labeled Data Test Data Data Factory DRIVE Pegasus NVIDIA PERCEPTION INFRASTRUCTURE LARGE-SCALE DEEP LEARNING MODEL DEVELOPMENT Workflow, Tools, Supercomputing Infrastructure Data Ingest, Labeling, Training, Validation, Adaptation Automation, Best Model Discovery, Traceability, Reproducibility Purpose-built for Safety Standards of Automotive “Data is the new source code”
- 5. 5 DATA COLLECTION ANDLABELING FOR AI Raw Data Total Images Useful Data Labeled Images # DNNs 100’s of petabytes of data from test vehicles 10’s of billions of total images from test vehicles 20% to 50% of data may not be useful 1,500 workers Label up to 1M images per month 10+ DNNs for self-driving vehicles Source: Data from test fleets of 50-100 cars
- 6. 6 AI FOR SELF-DRIVINGWORKFLOW DNN Development Exploration Development Model Selection Simulate Re-Simulate Labeled Data Trained Model Fine Tune Model Export Model Inference at Edge Train & Test Adjust DeployGet Data Test & Validate
- 7.
- 8.
- 9. 9 Customer Application DRIVE OS DRIVEAV Object, Freespace, Path / Lane, Path Planning, Wait, Map, Sign, Lights, Road Markings, Surround DRIVE IX Gaze, Eye Openness, Head Pose, Gestures, Emotions Facial Recognition, Voice Recognition & Lip Reading Exterior Driver Recognition Automatic Personalization Device usage detection Cyclist Alert Distracted Driver Alert Driver/Passenger Recognition AI OUTSIDE AND INSIDE THE VEHICLE
- 10. 10 MANY THINGS TOLEARN
- 11. 11 “ Autonomous vehiclesneed to be driven more than 11 billion miles to be 20% better than humans. With a fleet of 100 vehicles, 24 hours a day, 365 days a year, at 25 miles per hour, this would take 518 years.” Rand Corporation, Driving to Safety
- 12. 12 SIMULATION THE PATH TOBILLIONS OF MILES World drives trillions of miles each year. U.S. has 770 accidents per billion miles. A fleet of 20 test cars cover 1 million miles per year.
- 13. 13 NVIDIA DRIVE SIM ANDCONSTELLATION AV VALIDATION SYSTEM Virtual Reality AV Simulator Same Architecture as DRIVE Computer Simulate Rare and Difficult Conditions, Recreate Scenarios, Run Regression Tests, Drive Billions of Virtual Miles 1,000’s Constellations Drive Billions of Miles per Year
- 14. 14 NVIDIA DRIVE SIM ANDCONSTELLATION AV VALIDATION SYSTEM Virtual Reality AV Simulator Same Architecture as DRIVE Computer Simulate Rare and Difficult Conditions, Recreate Scenarios, Run Regression Tests, Drive Billions of Virtual Miles 1,000’s Constellations Drive Billions of Miles per Year
- 15. 15 NVIDIA DRIVE SIM ANDCONSTELLATION AV VALIDATION SYSTEM Virtual Reality AV Simulator Same Architecture as DRIVE Computer Simulate Rare and Difficult Conditions, Recreate Scenarios, Run Regression Tests, Drive Billions of Virtual Miles 1,000’s Constellations Drive Billions of Miles per Year
- 16. 16 NVIDIA DRIVE SIM ANDCONSTELLATION AV VALIDATION SYSTEM Virtual Reality AV Simulator Same Architecture as DRIVE Computer Simulate Rare and Difficult Conditions, Recreate Scenarios, Run Regression Tests, Drive Billions of Virtual Miles 1,000’s Constellations Drive Billions of Miles per Year
- 17.
- 18. 18 NVIDIA DRIVE END-TO-ENDPLATFORM COLLECT & PROCESS DATA TRAIN MODELS PedestriansCars Lanes Path LightsSigns SIMULATE DRIVE PedestriansCars Lanes Path LightsSigns
- 19. 19 370 PARTNERS DEVELOPING ON NVIDIADRIVE CARS TRUCKS MOBILITY SERVICES SUPPLIERS MAPPING LIDAR CAMERA / RADAR STARTUPS
- 20. 20 KEY TAKEAWAYS 1. Understandend-to-end requirements of autonomous vehicle development 2. AI demands data center design built on dense GPU compute-at-scale 3. Consider the complete workflow of AI from experimentation to training to inference 4. Carefully weigh cost of productivity vs hardware cost alone = true TCO of DL 5. NVIDIA best practices leads to TSTADI reference platform (Training, Simulation, Testing for Autonomous Driving Infrastructure)
- 21.