How digital twins can optimize Travel and Hospitality operations

One of the most frequent requests we hear from travel and hospitality (T&H) businesses and technology leaders is, “How can we use technology to optimize operations?” Optimizing operations is complex and if not done properly can lead to penalties, loss of yield, and the erosion of brand loyalty. In this blog, we’ll explore how travel and hospitality companies can use digital twins to monitor and optimize operations.

The complexities of travel and hospitality operations

Operations has a direct impact on growth, customer experience and employee satisfaction, so it’s important that each process works together perfectly. In the travel and hospitality industry, there are often many teams, supply chains, and logistics processes that must align in order to deliver the desired output. For example, an airline turnaround process requires multiple teams to perform at least 10 steps to ensure a plane takes off on time and safely.

In addition, travel and hospitality operations are challenging because:

Most segments of the hospitality industry (including accommodations and lodging, airlines, airports, and restaurants) are still grappling with severe staff shortages, making it difficult to scale and meet demand.
Data is located in silos across multiple organizations and IT applications. In an airport, data comes from the airport itself, airlines, ground handling agents, and so on.
Data needed to monitor operations range for formatted data from IT applications to sensor data from Internet of Things (IoT) devices to video streams from cameras. Multiple technologies and skill sets are need to ingest, parse and derive insights from these various data sets in real-time.
Visualizing the processes and their dependencies on the overall operations is quite difficult and the current forms of reports/dashboards are not intuitive to the user monitoring the operations.

How Digital Twins can help

A digital twin is a living digital representation of an individual physical system that is dynamically updated with data to mimic the true structure, state, and behavior of the physical system in order to drive business outcomes. Digital twins can be created for a physical environment, an object or even a process.

In the travel and hospitality industry, digital twins can be created (of an airport, hotel, or cruise ship) to replicate a view of operational processes, monitor various assets or key performance indicators (KPIs), and to process metrics. It is straightforward to create a single pane of data that reflects the actual behavior of the environment being monitored.

Digital twins can be used to observe operational processes, trigger alerts, and act on events that breach the threshold. The 3D visualization provides a natural user interface to interact with complex infrastructure spatially or drill down on processes to pinpoint issues and understand the impact on a complex set of connected processes. By providing decision makers current state information, historical data, and a recommended best action, travel and hospitality companies can generate a better understanding of their systems and processes. That understanding can be used to improve operational decision making and drive more effective actions.

Using digital twins, you can visualize multiple types of data including:

Geo-spatial/3D images of the asset/infrastructure being modelled
Near real-time data/events from business applications, such as a check-in application
Sensor data from Wi-Fi routers or people counters and images/video feeds from compatible camera
Image and video feeds from compatible cameras
Historic data from data stores such as data warehouses and data lakes
Predictive data and recommendations from artificial intelligence and machine learning (AI/ML) engines to anticipate problems or help resolve them

Building a Digital Twin with AWS IoT TwinMaker

Amazon Web Services (AWS) IoT TwinMaker makes it straightforward for developers to create digital twins of real-world systems. It has the ability to use existing data from multiple sources, create virtual representations of any physical environment, and combine existing 3D models with real-world data. You can harness digital twins to create a holistic view of your operations faster and with less effort.

How AWS IoT TwinMaker works

Main features:

Data connectors – AWS IoT TwinMaker provides the tools to bring in data from multiple sources simultaneously. There are native connectors for AWS services like AWS IoT SiteWise for collecting, organizing, and storing equipment and time-series sensor data. It also connects with Amazon Kinesis Video Streams for capturing, processing, and storing video data or Amazon Simple Storage Service (Amazon S3) for object files. AWS IoT TwinMaker also provides a framework for you to create custom data connectors to use with other AWS or third-party data sources, such as Amazon Timestream, Snowflake, and Siemens MindSphere.

This functionality is crucial for a complex industry like Travel and Hospitality where data is distributed throughout many systems and locations. We can, for example, bring data from sensors at the airport, and combine that with telemetry and scheduling data from airlines. We can also bring in insights from our gate video analytic workloads running in the AWS Cloud to give a comprehensive view of what is happening for an aircraft turnaround process.

Model builder – To model your physical environment, you can create entities in AWS IoT TwinMaker that are virtual representations of your physical systems, such as a check-in counter or a boarding gate. You can also specify custom relationships between these entities to accurately represent the real-world deployment of these systems. You then connect these entities to your various data stores to form a knowledge graph, that structures and organizes the information about the digital twin for effortless access and understanding of the current state with possible downstream implications.
Scene composer – With AWS IoT TwinMaker, you build a 3D digital twin by using your existing and previously built 3D visual models. Getting a 3D model of a complex space like an airport, a hotel or a cruise ship can be challenging. By leveraging tools like the Matterport integration, you can create photorealistic scans of your environments in a matter of weeks. Once you have a 3D model of your environment, you can then add overlays from your connected data sources to create a spatially aware visualization of your operations.
Applications – AWS IoT TwinMaker comes with a plug-in for Grafana Labs, a fully managed observability platform for your apps and infrastructure. The plug-in provides custom visualization panels, including a 3D scene viewer, as well as a data-source component to connect to your digital twin data, allowing you to quickly create 3D-enabled applications for your specific needs. The plug-in can also be used to build applications with Amazon Managed Grafana, which is a fully managed service for open-source Grafana.

Another option is using the AWS IoT Application Kit, an open-source UI component library which allows developers to create web applications that integrate with AWS IoT TwinMaker quickly and securely. IoT Application Kit gives developers the flexibility of the web while also allowing the security, scalability, and reliability of AWS IoT TwinMaker.

Reference architecture for a Digital Twin solution:

Now that we have discussed what AWS IoT TwinMaker is and its key functionalities, navigate to the following guidance to learn what a reference architecture for a digital twin for what an airport/airline operations would look like. This architecture can be used as a foundational block for other sub-verticals in the industry, the only change being the addition of data sources that are relevant for the use case being implemented.

Sample airport Digital Twin:

Now let’s see what a final application built with AWS IoT Application Kit could look like for a customer. In this case we have built a digital twin of an airport, which we can see through the scene viewer feature (Demo 1, panel 1). We also built surrounding features as react components to enhance the twin (Demo panels 2-5).

The application includes role-based access so that the different personas which interact with the digital twin will have customized views from the single pane of data. This is useful for being able to limit the amount of data each persona views to a needs-to-know basis.

In the case of an airport, the building management systems (BMS) will be interesting for the building operations team, but add no value to the flight dispatchers (who will just need gate data and a view into the turnaround process for each aircraft). For the purpose of this demo, we are using an admin role with access to all data.

Demo 1 – Airport digital twin dashboard

1. On the first panel (Demo 1, panel 1) we have the digital twin view. This provides a single pane view of everything that is happening at the airport. We can monitor passenger flow through the different checkpoints (check-in, security and immigration), equipment and asset tracking of mechanical escalators, solar panels or self-service check-in kiosks and gate information about the next outgoing flight. Within this view you can zoom, pan and move around the environment to better understand what is exactly going on. This can be very useful, for example, when you want to know which auto check kiosk is not working. You can see that kiosk go red or amber directly on the twin and quickly locate it geographically.

2. Panel (2) is a dashboarding panel which shows historical data about the specific asset that we are monitoring. This is where being able to access historical data from our data platform is crucial. It gives us a vision of what the trends have been for that asset in the past. We can see how unlikely an occurrence is or whether it is part of normal operation.

3. Panel (3) gives us more context into the alarm, why it was triggered, and connects to our recommendation engine to generate an action that could remediate the situation. These recommended actions are not automated because a supervisor is needed to make the best possible decision based on the data. However, some lower impact actions could be configured to send an update back to the edge with an action.

Demo 2 – Airport digital twin gate view with live video

4. This second screenshot (Demo 2) shows a zoomed in view into the gates, where we can see the state of the next outgoing flight by the icon next to the aircraft. In this case Gate-1 is on time showing a blue icon, and Gate-2 is delayed and shows an amber color, as well as having some additional information overlayed on the scene. We might want to understand exactly what is going on at Gate-2 because the aggregated data is not giving us enough insights. In this case we can open the live video panel (Demo 2, panel 4), and tap into the live view of exactly what is happening at the specific gate which it is monitoring.

Demo 3 – Airport digital twin gate process view of an aircraft turnaround

5. In the next panel (Demo 3, panel 5) we can see how the concept of a process digital twin can also be extremely helpful. In the 3D view, we have the same gate view with an overall status indicator for that next outgoing flight. This indicator is an aggregated view of the underlying process of an aircraft turnaround. By creating a process digital twin (which represents all of the steps and dependencies for a process to be considered complete), we can have a granular view into which piece is causing the delays. For this aircraft turnaround case we can see how we must go through the unloading and loading of cargo, deboarding and preparation of the aircraft for the upcoming flight, as well as boarding again. All of these steps must be completed in a specific order and within certain SLAs for a flight to leave on time. By giving flight dispatchers a deeper look into what is causing the delay of a flight, they can make informed decisions and involve the right resources to try and resolve problems in the most efficient manner.

Digital twins can be used to improve operations across a wide variety of use cases in the travel and hospitality industry as shown in the following image.

Travel and hospitality digital twin use cases

Conclusion

Optimizing operations is an ongoing complex task that requires effort and resources. In this blog we explored how by building a digital twin we can enhance our monitoring capabilities by creating a single pane data view of our environment. Making a digital twin available to operations decision makers will help give them the right information and context to efficiently optimize the plethora of complex use cases in the T&H industry. Smoother operations can translate into penalty prevention, improved yields and an increase in brand loyalty.

Interested in learning how to build digital twins? Join us at re:Invent on Thursday, the 30th of November. We’re leading an interactive chalk talk, TRV302: Optimize travel and hospitality operations with digital twins to demonstrate how to efficiently build a digital twin. Or, contact an AWS Representative to learn more.

AWS for Industries