Image: Nextera Robotics

Didge

Nextera RoboticsProgress tracking & quality inspection, Safety & Environmental Monitoring

Wheeled robot that drives site interiors on its own, capturing 360-degree photos and point clouds

4
Deployments: 3
Companies that ran it

How it works on your site

The Didge is a compact wheeled inspection robot that autonomously navigates active construction site interiors to capture 360-degree photography, generate point clouds, and run AI-based safety and progress analysis. It covers six subtasks across daily walkthrough documentation, safety monitoring, 3D scanning, BIM-to-field comparison, and defect tracking; manual walkthroughs, exterior inspection, and trades coordination stay with the crew. The Didge has been shipping since 2021 and is deployed on commercial, residential, and healthcare projects in the United States, with a Gilbane Building Company joint venture providing market access.

Progress tracking & quality inspection

BIM-to-Field Comparison (Scan-vs-BIM)

The BIM Compare Tool overlays captured site data with the uploaded design model to flag deviations, missing elements, and code violations. This requires a BIM model upload to the Didge AI platform; projects without a BIM model are limited to photo documentation and safety monitoring.

Daily Walkthrough and Progress Documentation

The robot deploys from its charging station, navigates the floor autonomously, and captures 360-degree imagery from consistent daily locations without a human operator walking the site. AI processes the imagery to track construction progress by trade, recognizing hundreds of object types and generating percent-in-place metrics. The platform produces virtual walkthroughs with remote navigation, and post-completion documentation provides a behind-the-walls record for future maintenance.

Defect Identification and Punch List / Closeout Tracking

The Didge captures the floor in 360-degree imagery on its daily patrol, and the project team walks that capture remotely to log issues: tag an item, take a screenshot, open an RFI or push it to Procore. Spotting the defect and assigning it to a trade stays with the project team.

3D Scanning and Point Cloud Capture

The robot captures LiDAR and camera data during its autonomous patrol, generating point clouds that feed the BIM Compare Tool and spatial documentation. The crew does not operate a tripod scanner or walk a capture route; the robot handles capture from its daily patrol.

Safety & Environmental Monitoring

Hazard Identification and Unsafe Behavior Detection

The robot detects fall hazards, uncovered holes, and unsafe conditions across its patrol routes using the same vision AI stack as PPE monitoring. Custom detection tasks can be programmed for site-specific hazards; one confirmed example is pooled water detection on a healthcare project.

PPE Compliance Monitoring

The robot identifies PPE violations, uncovered holes, guardrails, ladders, and lighting conditions using onboard cameras and cloud-based AI. It positions itself near high-risk activity areas pulled from the project schedule for targeted monitoring. The safety team reviews automated weekly reports and a Construction Site Safety Index score rather than relying on manual walk-arounds.

Image: Nextera Robotics

Who has run it

4 deployments · 3 companies · 1 country

Will it work on your site

Spec: Value
Environment: Indoor only
Surface: Concrete (finished and unfinished), crushed gravel
Power: Battery, 6-hour runtime (maker-stated)
Site preparation: Mesh network (3-4 radio units per building) and charging base stations required before deployment
Connectivity: High-power mesh network, works through walls (maker-stated)

Show all 11

Navigation: SLAM-based autonomous navigation
Autonomy: Autonomous single-floor navigation (wheeled); autonomous stair climbing (tracked variant); teleoperated floor changes (wheeled variant)
Control interface: Web dashboard
BIM/CAD: BIM model required for progress tracking and scan-vs-BIM features; not required for photo documentation or safety monitoring
Fleet management: Cloud-based; shared neural networks across fleet (maker-stated)

As the maker states it.