Calibrating Range Measurements of Lidars Using Fixed Landmarks in Unknown Positions

Area of Science:

• Robotics and Sensor Technology
• Geospatial Measurement Systems
• Signal Processing

Background:

• Light Detection and Ranging (lidar) sensors are crucial for 3D mapping and environmental sensing.
• Existing lidar calibration methods often require expensive additional hardware or complex procedures.
• Sensor nonlinearity and range-dependent errors degrade the accuracy of raw lidar measurements.

Purpose of the Study:

• To develop a hardware-independent calibration method for lidar sensors.
• To address sensor nonlinearity and heteroskedastic, range-dependent measurement errors.
• To improve the accuracy of lidar range measurements without specialized equipment.

Main Methods:

• Exploiting environmental assumptions: calibrating the lidar in a 2D plane parallel to the ground with fixed, orthogonal features.
• Utilizing the principle that relative distances and angles between fixed points remain constant despite sensor movement.
• Developing mathematically tractable models for analytically solvable equations suitable for embedded systems.

Main Results:

• A novel calibration procedure that does not require additional hardware.
• Demonstrated tenfold improvement in the accuracy of raw lidar measurements in field tests.
• Analysis of Mean Squared Error (MSE) performance dependency on sensor noise levels.

Conclusions:

• The proposed method offers an effective and economical solution for lidar sensor calibration.
• The approach successfully corrects for sensor nonlinearity and range-dependent errors.
• The technique is suitable for deployment in embedded systems, enhancing the reliability of lidar-based applications.