Abstract
In this paper, we propose a novel online learning algorithm, SpCoSLAM 2.0 forspatial concepts and lexical acquisition with higher accuracy and scalability.In previous work, we proposed SpCoSLAM as an online learning algorithm based onthe Rao--Blackwellized particle filter. However, this conventional algorithmhad problems such as the decrease of the estimation accuracy due to theinfluence of the early stages of learning as well as the increase of thecomputational complexity with the increase of the training data. Therefore, wefirst develop an improved algorithm by introducing new techniques such asrejuvenation. Next, we develop a scalable algorithm to reduce the calculationtime while maintaining a higher accuracy than the conventional algorithm. Inthe experiment, we evaluate and compare the estimation accuracy and calculationtime of the proposed algorithm, conventional online algorithm, and batchlearning. The experimental results demonstrate that the proposed algorithm notonly exceeds the accuracy of the conventional algorithm but also capable ofachieving an accuracy comparable to that of batch learning. In addition, theproposed algorithm showed that the calculation time does not depend on theamount of training data and becomes constant for each step with the scalablealgorithm.