Spotlight

HomePage
Mapping
MissionReplay
Intelligence
ML Predictions
Reports

## Inspiration

Power plants operate with complex machinery and infrastructure that require continuous monitoring to maintain safety, reliability, and efficiency. Boston Dynamics Spot can autonomously inspect these environments and collect valuable sensor readings, but the amount of information collected can be difficult for operators to interpret and act upon.

This inspired us to develop Spotlight, a platform designed to bridge the gap between autonomous robotic inspections and human decision-making.

Our goal was to create a system that doesn't just display what Spot detects, but helps power plant operators understand what happened, where it happened, why it matters, and what actions to take next.

What It Does

Spotlight is an intelligent monitoring and visualization platform that transforms Boston Dynamics Spot's inspection data into meaningful, actionable insights.

The platform combines several features into a unified experience:

Interactive 3D Digital Twin: A detailed virtual power plant where operators can visualize Spot's location, movement, and surrounding equipment.
Real-Time Monitoring Dashboard: Displays sensor readings, inspection timestamps, equipment conditions, and flagged abnormalities.
Mission Replay: Allows operators to revisit previous inspection missions, visualize Spot's movements, and identify when and where potential issues occurred.
Robot Navigation and Control: Provides simulated manual control of Spot, allowing operators to intervene during missions and explore different areas.
Anomaly Detection and Analysis: Organizes abnormal readings and inspection events to help operators prioritize potential equipment issues.
Predictive Analytics: Explores equipment health trends, anomaly recurrence, and remaining useful life estimates to support preventive maintenance.
Operator-Focused Interface: Transforms technical inspection data into accessible visualizations that simplify monitoring and decision-making.

By integrating these capabilities, Spotlight aims to make robotic inspection data more understandable, accessible, and useful for plant operators.

How We Built It

We developed Spotlight as a web-based application that combines interactive 3D simulation, robotic mission visualization, and data-driven analytics.

For our visualization system, we used Three.js to construct a navigable 3D power plant environment. We incorporated industrial 3D assets and worked with Blender to improve environmental realism, including machinery, pipelines, equipment rooms, walls, and structural details.

We designed the simulation to represent Spot's movement throughout the facility, with navigation paths, collision detection, manual controls, and mission playback capabilities.

Alongside the simulation, we developed a dashboard designed to organize inspection information, including sensor readings, location data, timestamps, and anomaly flags.

Our data architecture was designed to support integration with analytics and machine learning workflows, creating opportunities to detect equipment abnormalities and predict potential maintenance needs.

Challenges We Faced

One of our biggest challenges was creating a realistic and functional 3D power plant environment while maintaining smooth robot navigation. Industrial facilities contain complex layouts, narrow passageways, and densely packed equipment, requiring careful attention to environmental modeling and collision detection.

Another challenge was synchronizing robot movement, inspection information, and mission replay into a cohesive experience. We wanted operators to understand the relationship between Spot's location and the inspection data it collected.

We also faced the challenge of balancing technical functionality with usability. Rather than creating another complicated engineering dashboard, we focused on presenting information in a way that plant operators could quickly understand.

Finally, developing predictive analytics from limited inspection data required us to consider how historical equipment records, anomaly patterns, and simulated readings could support meaningful predictions.

What We Learned

Developing Spotlight strengthened our understanding of 3D visualization, digital twins, robotic navigation, data analytics, and human-centered software design.

We learned how to approach the challenges of representing complex physical environments in a browser, organizing robotic inspection data, and designing tools that make technical information easier to understand.

Most importantly, we discovered that the value of autonomous inspection doesn't come solely from collecting more data. It comes from making that data useful to the people responsible for keeping critical infrastructure operating safely.

What's Next

We envision expanding Spotlight through direct integration with Boston Dynamics Spot, allowing the platform to receive live mission telemetry and inspection readings from physical robots.

Future improvements include more sophisticated machine learning models for anomaly detection, equipment failure prediction, and remaining useful life estimation, supported by real historical maintenance data.

We also hope to expand the digital twin with more accurate facility layouts, additional equipment models, and enhanced operator control.

Our vision for Spotlight is to transform autonomous robotic inspections into a more intelligent, proactive, and accessible approach to power plant operations.