A research team from RMIT University is combining smartphone technology with artificial intelligence (AI) techniques to improve the productivity of warehouse operations.
The Victorian government's Department of Business and Innovation Digital Futures Fund has put $500,000 towards the two-year project. The university will also collaborate with CSIRO, Shiny Pty Ltd, Federation Logistics, Shiny Embroidery and the Australasian Production and Inventory Control Society.
The project will use advanced data mining and dynamic optimisation techniques to increase warehouse productivity by around 15 to 20 per cent and reduce human error, which can be as high 5 per cent, to almost 0 per cent, said Dr Andy Song, the project head and lecturer at RMIT's School of Computer Science and Information Technology.
"The data from the phone will integrate with the [warehouse] system to mine useful data and to improve the business process. We will use AI technology to make the user's phone more intuitive,” he said.
“For example, the person is walking down the aisle or is performing a picking – the phone would know what the operator is doing, such as walking from shelf to shelf, or standing [and] performing an operation like replenishment.
"The phone, therefore, would know the context and interact with the person accordingly. As the result, the traditional operations/interactions such as button click, scroll down can be minimised.”
Song aims to offer a more cost-effective way for small to medium businesses (SMEs) to streamline their operations and gain further business insights through combining data mining and smartphones.
"The process will be much more simplified and at low cost. There are expensive enterprise-level solutions available out there so for SMEs; they are not really suitable and the initial investment for them is too great so it kind of prohibits them. An industry-level device costs at least several thousand dollars each, and for a smartphone it’s at the most about $800."
A prototype has been developed and Song plans to test part of project in a warehouse next month. The project has been developed to work on both iOS and Android devices, he said.
Some of the challenges that Song and the team have had to face are dealing with the complex nature of dynamic, changeable data and ensuring that the solution can be implemented across a broad range of business constraints.
“We have to create a predictive model to cope with future changes, and also we have a few techniques to handle the dynamic nature of the data and try to compensate both short term and long term variations. For constraints, we will establish a general framework to anticipate a wide range of possible constraints.”
Follow Rebecca Merrett on Twitter: @Rebecca_Merrett