Bringing Visual Intelligence to AMRs: Peer Robotic’s Vishrut Kaushik

Roboflow · Beginner ·👁️ Computer Vision ·4d ago
In this video, we chat with Vishrut Kaushik, Senior Robotics Engineer at Peer Robotics, to discuss how they are using computer vision to build collaborative autonomous mobile robots (AMRs) that move heavy pallets and trolleys in industrial environments. Vishrut explores how computer vision unlocks new possibilities for giving robots the context needed to seamlessly work alongside humans. We also dive into the technical hurdles of training custom vision models to handle extreme edge cases, such as detecting worn-out wooden pallets in pitch-black trailers or identifying assets completely covere…
Watch on YouTube ↗ (saves to browser)

Chapters (12)

Introduction to Peer Robotics & Collaborative AMRs
3:36 Demo: Autonomous Pallet Unloading & Outbound Operations
9:01 The Role of LiDAR vs. Computer Vision in Robotics
14:09 Managing Unpredictable and Chaotic Warehouse Environments
17:57 Edge Cases: Dark Environments and Stretch-Wrapped Pallets
23:28 Vision Models: Oriented Bounding Boxes and Instance Segmentation
29:10 The Future: Semantic Scene Understanding and Safety Monitoring
31:49 Connecting the Robot to Roboflow APIs for Continuous Learning
33:53 Career Advice: Obsess Over the Problem Statement
39:33 Onboard Hardware: Local CPU and GPU Processing
42:38 Deployment Timelines: From Hours to Weeks
44:33 Accelerating Model Training with Roboflow
I Gave This Fish $10,000 to Trade Stocks
Next Up
I Gave This Fish $10,000 to Trade Stocks
Coding with Lewis