top of page

SkyOps Mission Planner

Air war operations planning software solution that enables decision-makers to quickly and confidently plan a mission using machine learning and AI.

Permission given by customer to share this work, Do No Distribute




Naval Information Warfare Center Pacific

Mile Two Project Lead of 3 person design team, UX/UI Designer



Adobe XD · Miro

Rapid Ideation & Concept Generation ·  Iterative Design · Research/Knowledge Elicitation · User Interviews · Information Architecture · Scenario Development ·

Data Visualization · Wireframing · Rapid Prototyping · Cognitive System Engineering Principles · Project Management · Customer Collaboration & Communication

Problem Characterization

Air Task Order (ATO) planning is a multi-dimensional, multi-perspective problem. The current process is very complex and time-consuming, resulting in:

  • Frequent, manual updates to maintain a fresh understanding of data

  • Limited plan comparison, testing, and optimization

  • Coordination challenges

  • Inefficient use of resources, time, and effort

Approach and Methodology

The Mile Two design team had a tight three-month deadline to create a proof of concept. To keep up with the pace, we had to devise a plan to rapidly gain knowledge of the mission planning process while making significant progress in concept design.


We adopted a weekly sprint process, building visual concepts based on existing knowledge and making assumptions where information was lacking. Regular meetings with subject matter experts (SMEs) helped us fill those knowledge gaps while simultaneously increasing design fidelity.

This approach led to an effective process of using critiques and validating/correcting assumptions to enhance our subject knowledge. Additionally, our strong collaboration with customers and SMEs fostered a sense of partnership in the design process.

Early Research Insights  

During our initial research phase, we unearthed multiple factors that were effecting the efficiency of the mission planning processes.

  • MAAP Chiefs make use of an existing AI software to narrow down the top 3-5 plans. However, they encounter difficulties in discerning the subtle differences between these plans. They must toggle between plan details to locate key metrics, hindering the ability to make direct comparisons.

  • They encounter challenges in visualizing how adjustment in one metics might positively or negatively effect other metics

  • The process lacks a visual representation of time, making it difficult to see how plans change over time. 

Domain Modeling: Aircraft Packages

The term “package” refers to a group of aircraft that are assigned to carry out a specific mission or task. This model represents all the elements that are incorporated and considered by a planner in the creation of a package.

Fight Tonight M2 Internal - Frame 26.jpg

Domain Modeling: Concept Map

Simplifying the complex mission planning process down the the core sequence of how the intended concept works. This established a foundation for the overall architecture.

Fight Tonight M2 Internal - Frame 27.jpg

Domain Modeling: Functional Decomposition

A functional decomposition includes the objective(s) and functions of a process within a work domain. We use this technique to understand the "why" and "how" of the process being studied.

Fight Tonight M2 Internal - Goal Means Decomp .jpg

Understanding the objectives and functions of a process can help us identify alternative functions and evaluate them in terms of how well they support the desired objectives. This functional decomp is broken down by each key player to better comprehend their unique goals and needs.

Scenarios and Wireframes

Multiple versions of scenarios were created to explore how the mission planner might interact with the SkyOps tool. By augmenting these scenarios with wireframes we were able to more effectively communicate our ideas to both the customer and subject matter experts (SMEs), while confirming the feasibility of our concepts and validating key assumptions. This pivotal phase helped to confirm that our knowledge of the mission planning process was robust enough to begin shifting designs to a higher visual fidelity.

Design Process

Design progression involved moving from low-fidelity wireframes and basic prototypes to higher-fidelity mockups and more refined visual representations. Throughout this iterative process, the design fidelity consistently reflected our growing domain knowledge. Initially, we emphasized rapid concept iteration, which provided valuable insights. As we gathered knowledge, we narrowed down concepts and channeled it into more refined designs, allowing us to make efficient use of our limited time.

1) Screen Zone Modeling
     Tools: Miro

Fight Tonight M2 Internal - Frame 17.jpg

2) Low Fidelity Wireframing
     Tools: Miro

Fight Tonight M2 Internal - Frame 18.jpg

3) High Fidelity Wireframing
     Tools: Miro

Fight Tonight M2 Internal - Frame 19.jpg

4) Low Fidelity UI Representation
     Tools: Adobe XD

Fight Tonight M2 Internal - Frame 20.jpg

5) Mid Fidelity UI Representation
     Tools: Adobe XD

Fight Tonight M2 Internal - Frame 21.jpg

6) High Fidelity UI Representation
     Tools: Adobe XD

Fight Tonight M2 Internal - Frame 22.jpg

Key Solution Functions

Planning & Re-Planning

  • Improves coordination through shared data perspectives

  • Visualizes alternative plans for comparison to improve decision making

  • Enables decision makers efficiently select the best plan and initiate it

Foundations for Autonomous Teammates

  • Designed for aggregating large amounts of planning data in one location

  • Human planner feedback loop to support machine learning model development and training

Final Proof of Concept

The Fight Tonight UI consists of five main UI components: 

  • Comparison View

  • Chart

  •  Plan Timeline

  • Package Timeline

  • Insights

Adjustments or to a plan can be executed in the Comparison View, Plan Timeline, and Package Timeline. Any adjustments to the plan within these components will be displayed in all the other components. This enables experimenting and investigating capabilities to improve a plan overall while providing understanding of what constraints drive a plan's scores. The UI provides two main ways to make plan adjustments: Nudging and Proactive AI.

Fight Tonight M2 Internal - Slide 1.jpg

Final Proof of Concept: Importing Plans

The MAAP Chief imports the top 3 plans from the plan generator into SkyOps. Plan summaries are shown in the Comparison View.

Fina Frame A.1_2x.jpg

Left Monitor (Plan Summaries & Comparison View)

Final Proof of Concept: Selecting & Nudging Plans

When selecting a plan the plan summary expands to show more data and the others minimize. Once you have this view, you can begin to nudge and adjust the plan.

Plan Summary: Nudge sub-metrics, examine sortie details

Plan Timeline: Add/remove targets from packages, adjust strike time, show overlays of fuel, escort, SEAD missions, and view insights.

The effects of any of these adjustments will be immediately displayed in the other panels as well. This example demonstrates the adjustment of a sub-metric by moving a slider. The red highlighting shows constraint limits in the slider. The effects of the slider change moves targets in the timeline.

Fina Frame B.1_2x.jpg

Left Monitor (Plan Summaries & Comparison View)

Fina Frame B.2_2x.jpg

Right  Monitor (Selected Plan Details - Chart + Timeline)


Full Dual  Monitor View

Final Proof of Concept: Package Details

The MAAP Chief can also drill down into the package timeline details. Here they an add, remove, and adjust targets and package elements. They can also add or remove aircraft from the package itself by utilizing slack resources or adjust the timing of additional resource elements.

Final Miro5_2x.jpg

Final Proof of Concept: Proactive AI

An alternative method of plan improvement involves using the Proactive AI function.

Fina Frame D_2x.jpg
bottom of page