Abstract
The analysis of sensor range data and its application to mo bile robot navigation are of crucial importance in the field of mobile robotic research.
We analyze the range data produced by an amplitude- modulated continuous wave (AMCW) light detection and ranging sensor and show that by physically modeling such sensors, we not only can produce reliable range estimates, but can also quantify our certainty in each range data point. We discuss the noise in the sensor and show the importance of using both phase and intensity data for calibration and data interpretation.
We consider in detail the phenomenon of "mixed pixel points" whereby false range measurements occur when the light beam transmitted is split between two or more surfaces of differing range and/or reflectivity. We describe a new algorithm capable of detecting sudden changes in surface reflectance and/or range in order to identify these "spurious" data points. We quantify the regions over which the detection method will work, as we consider its sensitivity to changes in range and surface reflectance, while also quantifying the possibility of falsely predicting a discontinuity.
Get full access to this article
View all access options for this article.