Warehouse Design – Autonomous Mobile Robots AMR’s - Goods to Man

Updated: Sep 4, 2020


Earlier this year I presented on 2 webinars. In the first webinar the panel discussed how logistics operations would be impacted by the Covid 19 pandemic. In the second webinar we discussed how and where Warehouse Automation could be applied to warehouses in the future:

The general consensus of opinion drawn from the 2 webinars are summarised as follows:

  • There is and will continue to be a reduction in the availability of labour – due to Covid 19 and Brexit

  • Labour costs are increasing – the minimum wage has increased and there is competition amongst employers to attract staff with increased rates (note: due to the downturn in the economy workers this may have changed)

  • Social distancing – traditional zoned picking of fast-moving items and how we operate in warehouses will need to be reconsidered due to social distancing

  • Many sites are now introducing staggered start times, one-way systems, additional health screening measures, all of which are having a negative effect on the efficiency of operations. Operational costs are increasing.

  • Ecommerce volumes had increased significantly, and this is likely to continue

  • We concluded that the use automation would continue grow.

Following the webinars, we have seen many retailers closing stores and making staff redundant or putting jobs at risk. There has been a continued trend towards online ordering. We are witnessing the demise of High Street retail and this is not likely to recover in the short term, and the trend for more orders fulfilled directly from a warehouse is likely to increase.

The picking operation in a warehouse generally accounts for more than 60% of the total labour hours. This area is now under scrutiny more than ever before, and as a result we are seeing efficient picking solutions such as “Goods to Man” increasing in popularity.

There are numerous Goods to Man systems available, including systems with conveyors and Automated Storage and Retrieval Systems (ASRS), and systems such as Autostore are proving to be highly efficient and popular.

This blog will focus on AMR’s – specifically Goods to Man systems, my next blog will consider collaborative robots.

Autonomous Mobile Robots (AMR)

When considering picking operations, the two most popular Autonomous Mobile Robots types are:

  • Goods to Man Systems

  • robots pick up and deliver shelf units (pods) to operators who work at static receipt and pick stations.

  • Collaborative Robots

  • robots that are self-guided around a warehouse, the pickers work in zones and the robots move between the zones

What is an AMR?

An AMR is a robot that moves around a warehouse independently and takes instructions from Warehouse Control / Execution Software (WCS / WES). The robots are fitted with sensors, scanners and 3d cameras. They navigate through buildings using digital maps routing from point A to B. The robots can avoid obstacles and can interact safely with humans.

Goods to Man – AMR’s

Goods to Man AMR’s will try to drive under the standing racks as much as possible. This leaves “highways” between pods open for robots carrying pods. When the robot unit reaches the target pod, it slides underneath the pod and lifts it off the ground. The robot then carries the pod to the designated location and completes the cycle. The operating areas for the robots are usually segregated from humans.

The technology is now maturing at a fast rate, with many suppliers coming to the market with new solutions. The pioneers of this technology were a company called Kiva. Goods to Man AMRs are now available from suppliers such as Swisslog, OWR Robots, Hikrobotics, Körber, Conveyor Networks Ltd, to name but a few.

Amazon Kiva Robots

Kiva Systems are considered to be the pioneers of AMR’s and had deployed them at numerous retail warehouses in the US including Amazon. Amazon bought the Kiva company in 2012 for $775 million. The Kiva system was previously used by companies including: The Gap, Walgreens, Staples, Gilt Groupe, Office Depot, Crate & Barrel, and Saks 5th Avenue. Amazon allowed the contracts to expire and did not renew them, effectively monopolising the solution. Kiva's assets now work only for Amazon's warehouses. The Amazon robots are regularly shown in action in Amazon warehouses on TV or You Tube videos.

Amazon states that fulfilment centres (FC’s) with robots are 3 times more efficient and 20% faster than traditional, less high-tech FC’s. Amazon currently has over 200,000 robots operating worldwide. Amazon has 175 fulfilment centres worldwide, there are 40 fulfilment centres across the EU including 17 fulfilment centres in the UK. The robots are not deployed at every location.

See the Kiva Amazon Robots in Action

In the UK there 7 locations that the general public can visit as part of a warehouse tour, (currently on hold due to Covid 19). Amazon fulfilment centre tours are available at the following locations:

  • Doncaster, Dunstable, Manchester, Peterborough, Rugeley, Tilbury, England, plus

  • Dunfermline in Scotland

The Tilbury fulfilment centre is the largest in the UK. This centre is classed as a “Sortable” fulfilment centres, customer orders such as books, toys and housewares in the warehouse the Amazon associates unpack and receipt items, pick, pack, sort and ship.

Summary Key Statistics for Tilbury

  • 2,000,000 sq. ft – over 3 floors

  • 2,000 employees – rising to 4,000 to cover Black Friday and Christmas Peaks

  • 6,000 plus Kiva robots

  • 48,000 shelf units (pods)

  • 18km of conveyor

  • Largest product has a max dimension of 45cm

The Tilbury facility operates with a relatively low inventory level and has been specifically designed as a fulfilment centre, with a large SKU range held in stock. Other bulky, large items, seasonal items, or fast-moving large order items are processed at other centres.

Warehouse Design

For a anyone that is considering Goods to Man AMR’s they will need to critically analyse the design of the warehouse.

  • Receipt – docks / door – the movements to receipt stations, overflow storage if required and subsequent replenishment

  • Storage – grid size, future expansion opportunities, grid movement paths and highways

  • Shelf units the number of shelf units required, the shelf unit design - the number of shelves per unit and the size of each shelf. Note: the utilisation and effective storage capacity of the shelf units may be reduced when compared to conventional shelving units.

  • Picking stations – systems used – volumes processed

  • Packing stations – and application of appropriate automation

  • Sortation of packed items ready for despatch

  • Number of robots – weight capacity, operating speeds, charging requirements, maintenance

Future growth projections. This will include assumptions relating to

  • SKU range growth

  • Stock depth

  • Customer base

  • Order forecasts – lines per order, orders per day, etc

  • Order Type / delivery method

In the picture of the Amazon Warehouse below you will notice is the amount of “fresh air” in the warehouse, the vertical space above the shelf units that is not being used. Without doubt, AMR’s can offer highly efficient picking solutions, at the possible expenses of warehouse space utilisation. Ideally AMR’s would operate on mezzanine floors in a high bay warehouse, however, this may cause challenges in transporting goods between floors.

Amazon Warehouse

Operational Processes - Overview

Stock Receipt into the System

Inbound Station

The high-level process is that received items are checked and manually located into shelf locations. The items are scanned, and the system associate’s the SKU’s and quantities to each individual shelf locations. When the instructed by the control system the shelf unit is returned to the grid storage area.

Comment: with the level of investment required for AMR’s, and in order to maximise efficiency it is commonplace to have “Pick to Light” and “Put to Light” systems deployed at both the inbound and picking workstations. This results in highly efficient operations as only the SKU’s are scanned to ensure product integrity, all other instructions are easily followed by the operator in the form of observing lights or the pushing of buttons.

Put to Light

Picking Station

Shelf units are sequenced to the pick station where the picker is directed by the screen to the shelf locations, the items are scanned, and the items are located to the destination tray. These trays when full are located to the conveyor system and proceed to the packing area. This results in very high pick rates when compared to a traditional picking operation as the operator is not having to travel between pick locations.

Completed orders can then be packed efficiently without the need to consolidate order lines.


Packing can be performed either at the pick station or in a separate packing area, possibly fed using a conveyor system. The packing process could also be automated, especially in the high-volume retail operation.


The brains behind the operation is the Control / Execution system. These systems sit in the background making decisions as to which robots perform which tasks based upon algorithms.

  • Routing: the control system will navigate the robots around the warehouse to complete tasks, AMRs are able to dynamically create their own efficient pathways within a facility.

  • Picking: the system decides which pods to present to the pick station and the sequence in which orders are picked. The overall aim to maximise the overall efficiency of the total operation.

  • Grid positioning: the system will also decide based upon what inventory is stored in the pod where it should be returned to in the grid. Pods with fast moving SKU’s will be located to the most accessible locations.

  • Receipt: the system will monitor inventory levels and available space on each shelf unit and sequence them to the receipt area for topping up.

  • Charging: the system will monitor the battery capacity of each robot and route them to the charging stations as required

The systems will also integrate with other WMS / ERP systems as required.

Case Study: Clipper Superdry Burton on Trent

In the UK, the clothing retailer Superdry has an ecommerce return rate around 25% of shipped units, the efficient processing of returns and making them available for sale again was an area that Superdry considered suitable for automation. Superdry and their logistics provider Clipper worked with Hikrobot together with their European partner Invar – a UK based warehouse solutions provider, and conducted a pilot project in 2018 using 6 robots in an area of 8,000 sq. ft. This trial was considered to be a great success and the concept has been relocated into a larger dedicated area of the warehouse on a mezzanine floor.

Hikrobots at the Superdry Duke Facility in Burton on Trent

The Superdry Facility operated by Clipper Logistics now has

  • 46 Robots

  • 80,000 sq. ft

  • 1,000 pods

  • 12 Pick to Light workstations

Reported Productivity Rates –

  • Receipt was 100 units / hour – now 300

  • Picking was 90 units / hour – now reported as 180.

In addition, a fleet of 20 Robots are processing returns at their DC in Belgium operated run by Bleckmann Logistics

Swisslog Carry Pick AMR

Körber Geek+ AMR


The overall efficiency of warehouse operations has never before been under such scrutiny. This coupled with a need to process increased volumes and a requirement to socially distance the operators has required a re-think as to how tasks are performed.

Operator confidence in robotic systems has increased, and in systems such as AMR’s where there is no single point of failure, if a robot breaks down or the battery is depleted another robot takes over the tasks. There are now many respected automation suppliers developing their own solutions and the control systems are now maturing.

As with any project of such magnitude, the analysis of the current operation and assessment of alternative concepts should be assessed by impartial independent experts. Please do not hesitate to ask for any advice. I would be very happy to provide a quotation to provide any assistance required.

363 views0 comments