Estimating the Heading of a Husky Mobile Robot with a LiDAR Compass based on Direction Maps

Without an absolute position sensor (e.g., GPS), an accurate heading estimate is necessary for proper localization of an autonomous unmanned vehicle or robot. This paper introduces direction maps (DMs), which represent the directions of only dominant surfaces of the vehicle’s environment and can be created with negligible effort. Given an environment with reoccurring surface directions (e.g., walls, buildings, parked cars), lines extracted from laser scans can be matched with a DM to provide an extremely lightweight heading estimate that is shown, through experimentation, to drastically reduce the growth of heading errors. The algorithm was tested using a Husky A200 mobile robot in a warehouse environment over traverses hundreds of metres in length. When a simple a priori DM was provided, the resulting heading estimation showed virtually no error growth.

Towards Extracting Absolute Roughness From Underground Mine Drift Profile Data

The purpose of this paper is to demonstrate a technique to utilize underground mine drift profile data for estimating absolute roughness of an underground mine drift in order to implement the Darcy-Weisbach equation for mine ventilation calculations. This technique could provide mine ventilation engineers with more accurate information upon which they might base their ventilation systems designs. This paper presents preliminary work suggesting that it is possible to estimate the absolute roughness of drift-like tunnels by analyzing profile data (e.g., collected using a scanning laser rangefinder). The absolute roughness is then used to estimate the friction factor employed in the Darcy-Weisbach equation. The presented technique is based on an analysis of the spectral characteristics of profile ranges. Simulations based on real mine data are provided to illustrate the potential viability of this method. It is shown that mining drift roughness profiles appear similar to Gaussian profiles