SBIR Phase II: Advanced Artificial Intelligence for Robotic E-Commerce Pick-and-Pack Automation

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase II project is to improve the resiliency of the supply chain by implementing flexible robotic systems for materials handling. The robotic systems that are controlled by artificial intelligence. E-commerce sales are increasing 20% year over year. During the COVID-19 pandemic additional retail volume shifted online and many customers became accustomed to sourcing essentials using e-commerce. This shift has put a greater burden on asupply chain infrastructure that has traditionally relied on human labor to pick, sort, pack, and process items for delivery. These manual processes are monotonous, error-prone, and sometimes dangerous, have extremely high worker turnover. The automation of these processes elevates worker roles and brings greater consistency to the processes. The innovation developed during this Phase II project may enable broader automation of complex materials handling processes by creating novel training systems for artificial intelligence-enabled robotic systems that are configured specifically for individual customer needs. This innovation may increase US supply chain resilience, enabling citizens to rapidly and reliably obtain necessities such as food, medicine, and health supplies without needing to leave their homes. The commercial opportunity is large, with over $20B spent on US pick and pack wages annually.This Small Business Innovation Research (SBIR) Phase II project seeks to develop new methods for rapidly training artificial intelligence (AI)-enabled robotic systems built for object identification and manipulation. Warehouse object manipulation tasks are variable and automating them often requires custom solutions for each customer and facility. These custom solutions are often prohibitively expensive. To solve these problems, an industrial operating system that can be deployed across many configurations of materials handling processes is required. This project aims to develop modules critical to scaling commercial deployments, such as quality control vision systems, automated assessments of item pickability, and enhanced AI systems for robotic picking. The anticipated result of this project is an industrial AI-enabled robotic operating system that allows rapid configuration of robotic systems to implement highly-optimized processes for picking and packing individual items in e-commerce logistics.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.