Turning 2D LiDAR scans into autonomous road repair
An autonomous road-repair research system that uses 2D LiDAR to estimate pothole geometry and guide a custom filling mechanism.
To connect low-cost sensing with a physical repair action, so road defects could be measured and filled more consistently.
From the initial question to a tested result
The timeline traces how the project moved through problem definition, design, prototyping, integration, and validation.
Convert range profiles into road-surface evidence.
The pipeline identified surface anomalies, estimated depth and geometry, and reconstructed enough of the pothole to support volume analysis.
Sensing was connected to intervention.
The system linked estimated repair volume to a custom filling mechanism, making the project a complete cyber-physical loop rather than a detection demo.
Three decisions that shaped the project
Each decision connects a technical constraint to the choice I made, the analysis behind it, and the tradeoff that followed.
Sensing
2D LiDAR
The system needed affordable, repeatable road-surface measurements without the data and compute burden of a full 3D sensor. A planar scan offered enough structure for geometric estimation.
Reduced cost and processing load while requiring geometric inference instead of direct dense 3D capture.
Convert range profiles into road-surface evidence.
The pipeline identified surface anomalies, estimated depth and geometry, and reconstructed enough of the pothole to support volume analysis.
Pipeline
Geometry before filling
Estimating shape and volume before dispensing connects sensing directly to material use, producing a controlled repair command rather than a simple detect-and-fill response.
Added computation up front to estimate repair material and guide a more controlled fill.
Sensing was connected to intervention.
The system linked estimated repair volume to a custom filling mechanism, making the project a complete cyber-physical loop rather than a detection demo.
Mechanism
Purpose-built dispenser
The task demanded predictable placement of repair material more than broad manipulation ability, so a dedicated mechanism concentrated complexity on the actual road-repair operation.
Improved task fit at the expense of general-purpose manipulation.