Industrial Forklift Optimization using AWS Services
Smarter Material movement is a key component of manufacturing operations. The advent of Industry 4.0 and smart factory adoption around the globe has accelerated this initiative even further. This article discusses a lighthouse use case for forklift fleet tracking and route optimization in a typical manufacturing plant.
The blog provides the framework for the initial step of reaching forklift fleet optimization, by providing building blocks to ingest and analyze forklift location data, it provides visibility (that were not possible before this) to a Material Planning and Logistics (MP&L) manager on the plant floor, who could use that to optimize his/her daily operation (scheduling of forklift operators).
Currently factory operators utilize traditional methods to track forklifts. For example, the plant manager tracks forklift usage based on time sheets and manual supervision. This leads to poor utilization of human labor and forklift resources thus impacting operations and increasing costs. Low visibility in forklift operations leads to underutilization of resources in the plants, resulting in inefficient manufacturing operations. Another common problem is forklift operators utilize excessive long pathways to drop off the load leading to energy waste.
Typically manufacturing customers rent forklifts on a contract for the fiscal year. Any underestimation or overestimation of forklifts required for the year impacts the business. Utilization of metrics and data will improve forklift visibility and location tracking. This will result in annual cost savings for manufacturing customers along with better utilization of labor.
Summary of the solution
Before we jump into the solution a brief background on UWB technology. UWB technology was invented in 1895 but was stagnant for a long duration due to dependence on legacy radio technologies. The technology evolved in the 20th century as radar and communication technology improved. Military reserved the technology for itself due to its application in detection and tracking of personnel and vehicles at the perimeter of critical areas such as military installations. In 2002 the Federal Communication Commission (FCC) allowed the unlicensed use of UWB systems in radar, public safety and data communication applications. Recently it gained adoption in smartphone technology and industrial asset tracking due to its superior accuracy and low energy requirements as compared to Bluetooth and Wi-Fi.
This solution involves a system of UWB tags, sensors and anchors integrated with AWS IoT services. Collected location data from the sensors are sent to AWS Cloud for analysis of forklift usage. Analyzed data provides visibility to the forklift operations for the plant manager and enables routing recommendations. The diagram below provides the architecture of our solution.
Technical architecture and steps for the solution
The above image shows UWB tags located on the forklifts aggregating location data and transferring the data from customers manufacturing plant to AWS Cloud. The data processed in the cloud transitions through phases of data transformation and analytics discussed in the steps below.
1. Collecting and migrating location data on-premises
The forklifts have Ultrawide band tags attached to them for the collection of location data. The tags emit short duration (2-3ns) Ultra-wideband (UWB) pulses. The UWB sensors attached on the forklift detect these pulses. UWB Tag transmissions are short in duration, providing excellent real-time location accuracy. The short-duration UWB pulses also result in long battery life. The UWB Anchors located at strategic locations in the plant (placed where forklifts often pass by) collect data from the tags sending the data to the UWB Gateway for aggregation. The Gateway has AWS IoT Greengrass installed on it to process the data locally and send the data securely to the cloud.
The Lambda function running on the gateway ingests the data and passes it to stream manager. Stream manager is designed to work in environments with intermittent or limited connectivity to send data to the cloud. The customer can define bandwidth use, timeout behavior, and how stream data is handled when the core is connected or disconnected. For critical data transfer, AWS Cloud leverages priorities to control the order of the streams. For connectivity customer registers Greengrass core running on the hub on IoT Core in the cloud. Devices in AWS IoT Greengrass environments use X.509 certificates for authentication and AWS IoT policies for authorization. The devices use MQTT protocol to send location data to the cloud since the payload for location data will be low (in KBs) and formatted in JSON format to be sent over to the cloud.
2. Data Store
As data gets ingested in AWS Cloud IoT Core sends the data to S3 for storage. S3 stores location data in unique namespaces depending on the timestamps and asset id, this helps in differentiating individual forklift data. Data lifecycle policy is added to the bucket so that the raw data can be transferred to S3 Glacier for archival purposes and cost optimization. This raw data once processed is utilized for analytics and visualization of Forklift locations.
3. Data Transformation
In the Data Transformation phase a Glue Job runs to identify the schema of the data stored in S3 and converts it into parquet format. This optimizes the data for query. AWS Glue automates the task of data discovery, conversion, mapping, and job scheduling. The solution has two Data Catalogs one for metadata which includes forklift id and timestamp and the other one for raw data stored in S3. Glue crawlers is defined to discover the data and add data to the catalog, the Data Catalog runs ETL jobs to refine the data. This data is sent to the Amazon Athena for querying and analyzing.
4. Analytics and Visualization
In the next phase the data is sent to Athena for analysis. With SQL support in Athena customer queries on the table generated from Glue Catalog. Analytics performed on this data is utilized to determine the location of individual forklifts in relation to specific date and time. Additionally standby time of the forklifts and utilization of forklifts over a period of time is calculated. Athena is used as a data source for Amazon QuickSight to visualize the data. Charts prepared on forklift ID, distance covered by the individual forklift, and standby time of forklifts provides deep understanding of forklift usage. These charts provide drilled down feature on the time of the day, period of the month and period of the year to identify patterns to optimize forklift rentals.
Benefits for manufacturers
Improving efficiency to achieve cost saving is always important for the success of manufacturing, which is often a low margin business. The solution above would:
· Enable plant manager to find more efficient scheduling and routing mechanism
· Reduce the total number of the forklifts (often rental) and operators required
· Reduce the down time that is caused by forklifts operations
· Open doors for further automation of overall material movement operations
· A pathway to real-time visibility to the forklift movements inside the plant floor, potentially on the tablet device used by M & L (Material Planning ans Logistics) manager, so that it would
The visible outcome would help the customer to get closer to the goal that “No forklift is ever empty”, and “No production process is ever waiting for forklifts”.
Forklift tracking and optimization lays foundation for larger scale Material, Planning, and Logistics (MP & L) optimization in a manufacturing environment. This capability can be further integrated into a MES (Manufacturing Execution Systems) and provide data for ERP (Enterprise Resource Planning) systems improve manufacturing operations to help achieve smart-factory goals.
For readers who are interested to know more about AWS IoT services can check out the product page. For information on AWS Manufacturing partners specializing in implementing UWB solutions check out our manufacturing partner page.