School of Infocomm (SOI)

Improving Supply Chain Traceability with Blockchain-enabled Smart Contract & IoT

Smooth Workflow Process 

Food tracking and traceability systems can utilise Distributed Ledger Technology (DLT) and Internet-of-Thing (IoT) sensors to streamline the workflow of food supply chain management. Such a traceability system mitigates the complex business processes and speeds up the workflow of working through multiple business owners along the nodes in the value chain.

Main Role Of DLT Technology: 

DLT Technology can be used to manage the food supply chain more effectively. It aims to obtain transparency, ease of transactions and speed of delivery, synchronisation, tracking and security. 

DLT is important in the global food trade as it ensures food safety and underpins trust to the end consumers. This technology can be used by supply chain management to trace the food from farm to fork. 

Final Software Features: 

The software developed serves as a reference implementation for a food supply chain application.

• It can be used by supply chain management to trace the food from the farm to fork. The target users are the process owners and consumers. 

• This technology is targeted at corporations and supply chain management that deals with livestock.

• This system can also be customised for other food industries such as halal, organic, agri-food and aquaculture industries.

Currently, proof-of-concept stage has been completed, and this technology is available to industry partners for potential licensing.
 

Principal Investigator:
Mr David Leong

Estimated Time of Completion Prediction for Last-Mile Logistics

Instead of just using the Estimated Time Of Arrival (ETA), this project aims to calculate Estimated Time Of Completion (ETC) values using more parameters such as time of day, weather and public holiday, among others, on top of the traditional approach that mainly relies on route distance.

ETC Prediction Model: 

It is a systematic Machine Learning Solution to predict ETC that overcomes the drawbacks of the existing methods. The system learns from historical data as well as additional features to improve the predictive model. 

Key Advantages of ETC Solution: 

• Low cost 

• Simple to use 

• Small set of input parameters are required 

• Predict the stop duration with reasonable accuracy 

• Conclude that factors such as vehicle, public holiday, rainfall are insignificant in the prediction

• Apart from being a productive tool for route planning system, the software can also be used in various situations such as:

  • Customer Services/Call Centres 

  • Fleet Management 

  • Loading Bay Assignment

Principal Investigator: 
Ms Deborah Zhou

Deep Learning for Warehouse Process Analysis

A time and motion study (or time-motion study) is a business efficiency technique combining the Time Study work of Frederick Winslow Taylor with the Motion Study work of Frank and Lillian Gilbreth.

One major application of time-motion study is to improve efficiency of a warehouse operation. The current manual method employed by the industry is time-consuming and prone to human error. 

Main Aim of the Project: 

In this project, we aim to provide an automated system that can perform the time-motion analysis of any given video. This is achieved using Deep Learning Recognition Models to identify pose and its relative position to regions of interest.

By identifying the landmarks of the pose within the regions of interest, we can calculate the time taken the subject spent within the regions. 

Advantages of Time-Motion Analysis Project:  

• Improve efficiency of a process and increase productivity.

• No specialised hardware required.

• Fast implementation to monitor different regions for different processes.

• Able to process recorded videos and live images from a webcam. The time-motion analysis of this software can be deployed in warehouses and factories/manufacturing lines.

• Beside time-motion analysis, the software can be customised to perform different tasks such as monitoring of danger zone as well as inventory usage.
   

Principal Investigator (PI):
Dr Jimmy Goh

co-PI:
Mr Zack Toh, Mr Melvin Ng

Team member:
Mr Derek Ang

Automating Leave of Absence (with Artificial Intelligence)

Organisations in all industries, including education, that track attendance will need to process Leave of Absence (LOA) submissions from staff and/or students. Processing such documents has traditionally been tedious and labour-intensive, as manual review of the submitted information is required to ensure data accuracy and legitimacy.

Additionally, storing such information is typically done in an unstructured manner, as it is usually in paper form or scanned versions of paper forms. 

Proposed AI Model: 

Reaping the benefits of digitalisation and automation, an AI model is trained in this project to automatically extract relevant and related data when a LOA document is submitted and subsequently, storing the data in structural database for easy retrieval and access. 

Key Advantages of AI Model: 

This AI model can be integrated to any system whereby the extraction of information from LOA submission is required to automate the business process, hence giving rise to the below possible adaptations of the project: 

• Work/medical leave 

• Education institution attendance recording system

Benefits brought about by the system include:

• Human-in-the-Loop Design

• AI model trained with local data

• Business process automation

• Digitised content

• Reduced error in data entry 

• Enhanced user experience
   

Principal Investigator (PI):
Mr Seow Khee Wei

co-PI:
Mr Koay Seng Tian