In any of the large manufacturing, mining, shipping, or retail and consumer packaged goods (CPG) industries, the business rationalizes manufacturing costs by optimizing production and warehousing operations. Since several warehouse functions continue to be human-intensive, intelligent next-generation enterprise warehouse management is now a business imperative.  

Enterprises need to mitigate disruption in production by enhancing efficiencies in warehouse operations for minimal time and effort in receiving, slotting, and issuing goods as well as optimal stock inventory. The infographic below illustrates a typical warehouse management process.

warehouse management process Infosys Cobalt

Whether it’s the food industry shipping packed and perishable goods or a mining company with large shipyards, warehouse activities are traditionally manual processes at indoor warehouses or outdoor/open yards. While mechanical equipment is used to move goods between locations or other facilities, activities such as pick-and-pack, inventory updates, and finding the right item for the right job involves significant human intervention. As human involvement increases, without efficient standard operating processes (SOPs), these activities become time consuming and error prone. Moreover, in multi-level warehouses, handling hazardous materials increases the risk of injury and accidents.

Enterprises need to adopt smart technologies in warehouse operations. For instance, warehousing operations can become more efficient with artificial intelligence (AI) and machine learning (ML) for quick routing of parts and items, while technologies like the Internet of Things (IoT), vision APIs, and image analytics can be employed for predictive maintenance and hazards in the warehouses.

Challenges in warehouse operations

The number one priority for enterprises is to adopt automation to ensure safe working conditions. In a manual warehouse, with constant movement of people and equipment, the possibility of accidents is high. According to a report by the U.S. Occupational Safety and Health Administration Agency (OSHA), forklifts in warehouses cause 100,000 accidents annually, with 80% of accidents involving pedestrians.  

In 2019, warehouses in the United States recorded 4.8 cases of injuries for every 100 full-time employees, resulting in absenteeism, work restrictions or job transfers, according to the U.S. Bureau of Labor Statistics. Unfortunately, injuries cause losses for workers, businesses, and industries. In 2019, work-related injuries and deaths in the United States caused US$171 billion in losses, according to the National Safety Council.

Enterprises should enhance warehouse efficiencies by eliminating redundant tasks. When a warehouse manager does the same task repeatedly every day, complacency contributes to human error. When repetitive tasks such as stock counting are undertaken daily by a human workforce, there is a high probability of inaccurate stock counts, which cause issues downstream in production.

Warehousing functions such as put away and picking, when undertaken manually, are prone to human error. Although rack and aisle placement for specific goods are typically assigned in advance, identifying those locations and arriving at them can be a slow process.

Infosys Intelligent Warehousing solution powered by AWS and Infosys Cobalt

We leverage Infosys Cobalt repository for blending AI, ML, and IoT to reduce the complexity of warehouse operations. A suite of smart technologies simplifies daily warehouse operations and reduces the time and effort in warehousing.

Infosys has partnered with Amazon Web Services (AWS) to enhance automation and intelligence in warehouse management by leveraging AWS IoT Edge Services, Amazon Rekognition (machine learning image and video analysis), and Amazon SageMaker (machine learning).

The infographic below illustrates our vision of an intelligent enterprise warehouse management process.

intelligent enterprise warehouse management Infosys Cobalt

Machine learning for picking and put away

Optimal warehouse management begins with intelligent put away processes by harnessing historical data. Algorithms created in AWS SageMaker identify the parts that are commonly issued together, and in standardized quantities and frequency. Based on the frequency of the issued part from the warehouse, an aisle and rack closer or farther away — from associated items, components, or the picking machine or individual — are assigned. Parts commonly issued together are assigned a rack next to each other on the same aisle. It makes picking activity more efficient and saves the robot or human a significant amount of time when a new request of parts is issued.

Amazon Rekognition provides a production and maintenance order by allowing a warehouse manager to click a photograph and find the right part. The manager learns about the part’s status related to the quantity, aisle, and rack on one screen or device.

The use of blockchain

While warehousing is important, it is equally important to win the trust of the customer, may it be a retail or corporate customer. For certain critical or high-cost parts, a system is created where the customer can scan the package barcode on the company’s website and validate whether the packaging is genuinely from the company it says it is from. This will also help companies deter counterfeits in the market. To do this, enterprises can use Amazon Managed Blockchain, a fully managed service that makes it easy to join public networks or create and manage scalable private networks using the popular open-source frameworks Hyperledger Fabric and Ethereum.

Drones for accurate recognition and parts counting

Warehousing can be further enhanced by using autonomous drones and leveraging AWS IoT Edge Services and Amazon Rekognition. A drone with a mounted camera and/or RFID scanner flown in a predefined path across aisles can record items (such as handling units) to assess availability and accelerate stock counting. The AWS IoT solution can run multiple times without incremental cost, a task that can take days to perform manually and often produces inaccurate results. Such solutions can operate across multiple warehouses, or multiple drones can be used for massive warehouses. When the drone operation is complete, it returns to the charging station without human intervention.

In a retail/CPG context, an external customer can create a sales order through a conversational bot by taking a photograph and uploading it through a messenger service such as WhatsApp or web chat. Amazon Lex — a fully managed AI service with advanced natural language models for building conversational interfaces into applications along with Amazon Rekognition — processes the image, identifies the part, and verifies its availability to simplify the process and avoid delays. When an internal customer, such as a plant manager, creates a maintenance order, they take a photograph that is processed to find the parts required for maintenance. The information about the parts along with their availability is shared with the maintenance engineer to save critical operations time.

Our vision: Reduce operational costs of managing a warehouse and the time required to execute processes in warehouse operations. Intelligent warehousing powered by AWS also increases speed, enhances precision, reduces risks, and improves uptime of warehousing processes.

About the authors:

  • Nitin Mahajan, Senior Technology Architect – SAP Intelligent Technologies, Infosys.
  • Prashant Gupta – Director and GTM Lead for SAP on Cloud, Europe, Infosys
  • Jignesh Desai, is the Infosys WW PSA Migration Lead at AWS

Copyright © 2021 IDG Communications, Inc.

https://www.cio.com/article/3640577/next-generation-warehouse-management-powered-by-aws.html