基於多重假設追蹤之多機器人同時定位與追蹤

Abstract

在環境中定位對於自主移動式機器人來說是不可或缺的能力。過去的研究已顯示，協作型定位(Cooperative Localization)對於多機器人定位的應用非常有幫助。然而，當環境中的移動物體位於機器人周遭時，可能會影響協作型定位的效能。本論文提出協作型同時定位與追蹤(Cooperative Simultaneous Localization and Tracking)方法，展現定位與追蹤之間能夠互相幫助。在本研究的實驗中，當機器人處於定位資訊缺乏的環境中，利用追蹤所獲得的資訊，仍能有效地協助估計其位置。在資料連結(data association)錯誤的情況下，結合定位與追蹤反而可能造成彼此的錯誤估測。因此，透過結合協作型同時定位與追蹤以及多重假設追蹤(Multiple Hypothesis Tracking)兩方法，來保留正確的資料連結。此外，兩方法之結合甚至能夠在靜止物體對稱之環境中協助估測機器人位置。在此研究的實驗中採用機器人足球賽(RoboCup)之環境設定，使用之人形機器人僅具有低精確性之鏡頭及里程表(odometer)。實驗結果顯示，基於多重假設追蹤之多機器人同時定位與追蹤能提供穩健及精確的定位以及追蹤資訊。

Localization is one of the most essential capabilities of autonomous robots. Cooperative localization has been proved to be effective in multi-robot localization. However, nearby moving objects could degrade the cooperative localization performance. In this thesis, we demonstrate that the cooperative simultaneous localization and tracking approach is superior in challenging scenarios. Localization and moving object tracking are mutually beneficial. We also illustrate the disadvantage of jointly estimating the states while the data associations are incorrect. The cooperative localization and tracking approach is integrated with the multiple hypothesis tracking (MHT) framework in order to maintain correct data associations. By integrating with the MHT framework, the proposed method is even able to correctly estimate the robots’ poses while the static features are symmetrically distributed. The proposed approach is evaluated using humanoid robots in the RoboCup environment in which only uncertain data from onboard cameras and odometry are usded. Ample experimental results with ground truthing from laser scanners demonstrate the accuracy and feasibility of the proposed MHT based multi-robot simultaneous localization and tracking algorithm.