ASLAM ROC Analysis Report

PDF

Authors
  1. Ghazel, M.
  2. Se, S.
Corporate Authors
Defence R&D Canada - Suffield, Ralston ALTA (CAN);MacDonald Dettwiler and Associates Ltd, Richmond BC (CAN)
Abstract
The objective of the Appearance-based Simultaneous Localization And Mapping (ASLAM) project is the research and development of an Appearance-Based Simultaneous Localization and Mapping system for day/night operations in indoor and outdoor environments. These algorithms would perform place recognition based on sensor data gathered from an Unmanned Ground Vehicle (UGV) as it travels through the environment. When the vehicle returns to a previously visited scene, the ASLAM algorithm would recognize the scene, update its internal representation, report this to the UGV, and finally provide information to aid in closing the loop with geometric Simultaneous Localization And Mapping (SLAM). The key objectives of Task 4.4 are to characterize the performance of the ASLAM system under different settings and to determine the optimal system parameters, with detailed Receiver Operator Characteristic (ROC) analysis. This report describes the findings of the ROC analysis. The ROC curves illustrate graphically how the probability of detection and the false alarm rate vary at different thresholds, and how the key system parameters affect the performance.

Il y a un résumé en français ici.

Keywords
Appearance-based Simultaneous Localization And Mapping;ASLAM;UGV;imultaneous;Localization And Mapping;SLAM
Report Number
DRDC-SUFFIELD-CR-2013-070 — Contract Report
Date of publication
01 Mar 2013
Number of Pages
70
DSTKIM No
CA045997
CANDIS No
806467
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: