Artificial Intellingence in the Car
- 1. 1 Attention Back in the Driver’s Seat Dr. Ing. Roberto Sicconi CTO at TeleLingo Dr. Maggie Stys CEO at TeleLingo
- 3. 3 80,000 Accidents 8,000 Deaths +14% since 2014 Distraction Inattentive DrivingDrowsiness 15% claims costs increase since 2014 $40B cost of distracted driving claims
- 4. Insufficient Driver Attention Risk of drowsiness on long stretches and of losing control of the vehicle Drowsiness Missed opportunity to make daily commute stimulating and productive Frustration TeleLingo Proprietary & Confidential Driving accidents are on the rise US: 660,000 smart phones used by drivers any given time Distracted Driving US 2013 (NHTSA) 424k injuries 3154 deaths EU 2008 (CARE) 4200 deaths Car accidents US 2013 3.8M injuries 35,200 deaths EU 2008 1.8M injuries 43,000 deaths Insufficient Data about Drivers Distractions Insufficient attention paid to driving, inability to handle sudden emergencies Churn Repair costs and Personal Injury claims are rising as accidents become increasingly expensive Costs Fraudulent claims (staged accidents, settled bodily injury, phantom passenger claims, car thefts) are surging ($35B)* Fraud Direct insurers, online comparison tools, limited knowledge of customer reduce retention rates Driver Insurance Highway accidents US 2010 (NHTSA) Total: $277B+$594B (damages + societal harm) Distracted driving 17% ($46B+$129B) *source III 2016 4
- 5. Analyze/Model Driver Attention + Evaluate Driving Context Risks + Manage Effective Feedback AI The secret sauce?
- 6. 6 Analyze Driver Attention Model Driver Behavior Evaluate Car “Stress” Audio (Voice) Feedback to Driver Stimulating Mind Exercises Evaluate Traffic, Weather Model Driving Context Verbal Interaction w/ Driver distracted drowsy Track Improved Performance Margin in Driver Attention? Eyes on road Hands on wheel Heartbeat & Skin Evaluate Driver Stress Model Driver StateOpt. biometrics from wearable bracelet / smart watch Smart DMS1 for Self-Driving Vehicles 1: Driver Monitoring System 2: Artificial Intelligence AI2
- 7. AI CoPilot + LingoFit 1 (iOS/Android) Auto-IR HD Camera Mobile App 2017 + LingoFit 2 (Win/Linux) Auto-IR HD Camera Mini PC 2017 + LingoFit 3 (Win/Linux) 3D IR Camera NUC Mini PC 2017 + LingoFit 4 (Linux) Integrated CPU & Smart Cameras Opt. Wearable Biosensors 2018 LingoFit 5 (integrated in rearview mirror) 2019
- 8. Car Insurance 38M Attention Monitoring (UBI*) 9M Teen Drivers 100M Single Drivers Fleet Management 3M Truck Drivers Ride Sharing 1M Drivers All numbers for US market, 2019 est. * UBI: Usage Based Insurance Driver ID, Anti-theft Feedback on driving behavior Drowsy driving detection Events recording (e.g. hard braking) In-cabin monitoring (driver and passengers) Driver ID, Anti-theft Feedback on driving behavior Events recording (e.g. hard braking) Distracted driving prevention Recording of unsafe events (driver and passengers) Driver ID, Anti-theft Driving behavior scoring Events recording (e.g. extreme braking) Drowsy driving prevention Distracted driving prevention In-cabin crash, whiplash recording LingoFit 2, 3LingoFit 1, 2, 3 LingoFit 4, 5
- 9. What makes LingoFit unique? TeleLingo Proprietary & Confidential Digital Assistant with Copilot Expertise Like an expert passenger, LingoFit • Monitors driver’s attention and responsiveness • Assesses risk by evaluating traffic, speed, driving conditions and dangers ahead • Alerts driver of sudden risks • Detects drowsiness and helps restore awareness LingoFit uses AI and Computer Vision • To detect distraction or drowsiness • To learn and adapt to driving style and communication preferences • To decide when and how to use sound and verbal interactions without distracting the driver Watch driver Watch scene Attention level Risk level AI Interact with driver 9
- 10. Autonomous Vehicles = no more accidents? Year 2032 Autonomous Vehicles 32% of new cars (best case scenario) = 3% of cars on the road Google Mercedes Audi
- 11. LingoFit and Self-driving Vehicles • McKinsey estimates that by 3032 at most 30% of new vehicles (~30M) will be autonomous -> 3% of the total 1.2B vehicles on the road • LingoFit is complementary even to Level 4 self- driving vehicles, where the driver needs to stay alert and supervise • How does the car know you are watching the road and ready to take control if needed? Or asleep? • LingoFit to process biosensors data to determine stress, fatigue, health issues and provide medical advice
- 12. Thank You!