Total automation of the warehouse picking process – are we there yet?

With the ever-increasing popularity of shopping online, ecommerce businesses are keen to make the customers’ shopping experience as straightforward and efficient as possible. Order fulfilment, particularly the picking of products, after purchase by the customer, remains a part of the process vulnerable to delays.

Ecommerce giants, Amazon recognised the need to improve their warehouse picking process and have invested in robotic technology. It’s investment in robotic technology from Kiva Systems, reversed the picking process, when 15,000 robots were introduced into Amazon’s warehouses.

Instead of pickers accessing the warehouse shelves to pick products and taking them to the packing station, now the shelves come to the pickers at the packing station. The robotic technology highlights which product to pick and confirms the bar code. Robots reduce the time spent physically walking in the warehouse to retrieve goods and minimise the wrong products being picked.

Whilst this partial introduction of robots into the warehouses has speeded up the picking process, the manual picking from the shelves and packaging still presents a potential bottleneck in the order fulfilment process.

This perceived need prompted Amazon to issue a picking challenge at the recent International Conference on Robotics and Automation. The challenge was to replace the human picker with robots and highlighted the need for multi skilled robots to:
Recognise objects on the picking list and the shelves
Decide how to remove the objects from the shelf.
Execute the manoeuvre.
Move to and from the shelves to the packing station
Be aware of errors and rectifying them once recognised.

The challenge highlighted the importance of computer modelling and sensors. It also used existing robotic technology successfully.

The challenge involved picking twelve products in twenty minutes from a stationary shelf pod consisting of twelve cubby holes, holding twenty four items.

Observations from the picking challenge:
Transitioning computer modelling to real time situation
Manoeuvrability of manipulator robots picking items from small cubby holes at various angles.
Robots were faster than anticipated

Robots were scored on how many items they picked correctly and lost points for incorrect items, as well as damaged and dropped items.

Total automation of the warehouse picking process is not here yet but the recent Amazon picking challenge demonstrates it may be soon.