Adaptive Memory Size Based Fuzzy Control for Mobile Pedestrian Navigation

The range of memory specifications of mobile pedestrian navigation systems poses difficulties for the developer (in terms of increased time and effort) when it comes to developing a resampling algorithm for mobile pedestrian navigation devices. Thus, a new resampling algorithm is required with a flexible capacity that would cater for a range of computing device memory specifications. This paper develops a new single distribution resampling algorithm, the Adaptive Memory Size-based Fuzzy Control (AMSFC), that integrates traditional resampling and traditional variation resampling in one architecture. The algorithm switches the resampling algorithm on the basis of the memory of the particular mobile pedestrian navigation, thus making it easier for the developer to develop a particle filter without having to consider the memory utilisation of mobile pedestrian navigation devices during different particle filter development processes. At the beginning of the operational process, the AMSFC selector is used to select a suitable resampling algorithm (for example, systematic resampling or rounding copy resampling) based on the physical memory of current computing devices. If systematic resampling is selected, the resampling algorithm samples each particle for each j cycle, while if the rounding copy resampling algorithm is selected, the resampling samples more than one particle of each j cycle. This demonstrates that the proposed method (AMSFC) can switch resampling algorithms to meet the differing physical memory requirements. The authors aim to extend this work in future by implementing their proposed method in a number of different emerging applications (in example, medical applications and real time locator systems).