Robotic order fulfillment in retail and e-commerce

Situation & problem

Our client, a retail company, wanted a fleet of mobile manipulators that can navigate a supermarket or warehouse and do order picking for e-commerce. The solution needs to be able to pick a large number of different product categories: boxed, canned, bagged, bottled products of different shapes and sizes. Some of them stacked on shelves in different patterns, others hanging on rails.

Intermodalics’ solution

We created a full innovative robotic solution for our customer in retail. Starting from a proof of concept using off the shelf hardware to a deployable product in less than 9 months. Combining our expertise in advanced robotic systems, software engineering and a lean approach to product development. As long term users and contributors of both licenseable and open source robotics software a first prototype could be set up in less than 2 months. Our deep understanding of the underlying algorithms allowed us to build custom capabilities in 3D-perception, navigation and motion planning, while leveraging the 3D engine of Pickit3D.

Outcome

We delivered to the client an end-to-end solution of a fully autonomous mobile manipulator that is able to locate itself in a warehouse, able to navigate through it, detect and locate individual products in 3D, select a tool and approach to grasp and manipulate the object, and fill the tote with the customer’s order in a smart way. The solution includes a backend processing customer orders to operational robot tasks and software tools to simulate, visualize, deploy and monitor the operation of the overall system in production.