An improved serial method for mobile robot SLAM

Abstract

Computational cost and real-time performance are important factors that many algorithms need to consider. For this purpose, this paper proposes an improved serial implementation strategy of mobile robot SLAM by combining Fast Fourier Transformation (FFT) and Iterative Closest Point (ICP) which is named FFT-ICP. We use FFT to localize coarsely and then use the results of FFT as the initial values of ICP to localize precisely. The experiments show that the method proposed here not only can speed up the operation, but also has higher precision through combining FFT and ICP. In addition, this method compromises the pros and cons of FFT and ICP. This technique is proven to be helpful for constructing an on-line SLAM system.