Abstract
The Simultaneous Localization and Mapping (SLAM) problem is of great significance within the modern field of unmanned systems. However, many current methodologies have high cost implications, utilising expensive Light Detection and Ranging (LIDAR) or Charged Coupled Devices (CCD) sensors to obtain information pertaining to the local topology of the device. The objective of this paper is to reduce the inherent cost of SLAM by generating a motion model which is suitable for use with the low cost Microsoft Kinect sensor system. A novel filtering methodology is developed which can separate the static and dynamic accelerations in order to compute a full 6 DOF pose estimate from a 3 axis–accelerometer, suitable for application as a SLAM motion model. The filter is seen to operate in constant time, at a frequency sufficient for on–line implementation to a suitable level of accuracy for use with SLAM.
| Original language | British English |
|---|---|
| Pages (from-to) | 70-79 |
| Number of pages | 10 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 9246 |
| DOIs | |
| State | Published - 2015 |
| Event | 8th International Conference on Intelligent Robotics and Applications, ICIRA 2015 - Portsmouth, United Kingdom Duration: 24 Aug 2015 → 27 Aug 2015 |
Keywords
- Filtering
- Kinect sensor
- SLAM problem
- Unmanned Aerial Systems (UAS)