Team PLAID’s objective is to develop an autonomous robot that can pick various objects in a warehouse environment. The system was designed to pick up at least 12 items from a shelf and drop them inside designated target totes, within 15 minutes. The system must also report the items picked.
Picking is the last remaining link to be automated in an Amazon warehouse. After 2009, Kiva robots changed the face of Amazon warehousing, moving shelves to workers without human intervention. However, manpower is needed in order to pick items and to place them in the designated box with reference to work order. With our robot, the cost of operation in a warehouse can be reduced.
This project is a continuation of a previous project completed by last year’s team HARP which competed in the 2016 Amazon Picking Challenge. However, there is very little carryover in terms of reused hardware or code. This year’s competition includes items that are more challenging to grasp and identify. Moreover, this is the first year that Amazon is allowing participants to design and use their own shelving system which warranted large system re-designs.
Throughout the project, Team PLAID focused on tackling challenges faced by autonomous picking system such as accuracy and speed of picking items. Functional and nonfunctional requirements and milestones were based on the Amazon Robotics Challenge rules. Major subsystems were perception, grasping, planning, and storage system. This report will cover system description, project management, risk and testing performance.