AI summary by TwinMind
A friendly, autonomous robot that delivers your food | Ali Kashani
TED
8m 52s
Food Delivery Problem Analysis
- Americans order 20+ million restaurant deliveries daily
- Over half deliveries within walking distance
- 9 out of 10 delivered by cars
- Moving 2-pound burrito in 2-ton car 20 million times/day
- Energy inefficiency: bringing 2-ton metal cage with heated seats for small food items
Urban Car Dependency Issues
- America: 4 parking spots per car
- Downtown areas: 50%+ real estate dedicated to cars
- Cities designed around cars for all distances (2-200 miles)
- Same SUV used for coffee or coffee table purchases
- Opportunity: freeing streets/parking lots → more housing, social spaces, parks
Future City Vision
- 5 blocks: summon bike/scooter
- Rush situations: passenger drone pickup
- Food delivery: autonomous delivery without human drivers
- Potential impact: reduce 1.5 million cars in US (2x San Francisco size)
- Global implications: cities like Delhi, Tehran where car pollution kills thousands yearly
Delivery Robot Development (3-year project)
- Small self-driving robots navigating alleys/sidewalks
- Walking pace movement
- Secured cargo compartments
- Building blocks for future city infrastructure
Robot Design Philosophy
- Avoiding Hollywood dystopia expectations
- Creating familiar, community-friendly robots
- Surprise and delight strategy for first impressions
- Design: shopping cart crossed with Minions
- Active deployment: San Francisco, Los Angeles
Human-Robot Communication Development
Visual Communication
- Eyes as universal language
- Show robot direction/confusion
- Make robots more human-like
Audio Communication
- Running sounds with frequent gaps
- Doppler effect for visually impaired users
- Location identification system
Movement-Based Language
- Intersection behavior: gentle forward movement before crossing
- Wheelchair users: robots yield by pointing away from sidewalk
- Small directional adjustments causing pedestrian confusion
Social Interaction Challenges
Historical Context
- 2015: Canadian hitchhiking robot destroyed in US
- Carnegie Mellon research: small toy robots play dead when tampered
Delivery Robot Strategies
- Show awareness vs playing dead
- Acknowledge tampering situations
- Get people to step away
- Kid interaction: run toward closest adults (kids love harassing robots)
Future Applications Beyond Food Delivery
Food Waste Reduction
- Gather excess food for shelters nightly
- Address: 30% food waste, 10% food insecurity in America
Emergency Medical Response
- Hundreds of robots carrying emergency medications
- Allergic reaction/asthma attack response
- 1-2 minute response time (faster than humans)
Pandemic Infrastructure
- Essential needs delivery during emergencies
- Community support during lockdowns/quarantines
Technology Evolution Perspective
- Current limitation: objects need human help for A-to-B movement
- 3D world complexity challenges
- New sensors + AI enabling autonomous navigation
- Technology metaphor: baby learning object recognition, words, conversation but not walking yet
- Teaching technology 3D world navigation without human assistance
Vision for Sentient Objects Future
- New era: sentient objects moving freely
- Importance: avoiding alien-like appearance
- Goal: objects come to life with joy
- Reference: Toy Story vs Terminator aesthetic approach
