Limited memory predictors based on polynomial approximation of periodic exponentials

The demands for indoor positioning in location-based services (LBS) and applications grow rapidly. It is beneficial for indoor positioning to combine attitude and heading information. Accurate attitude and heading estimation based on magnetic, angular rate, and gravity (MARG) sensors of micro-electro-mechanical systems (MEMS) has received increasing attention due to its high availability and independence. This paper proposes a quaternion-based adaptive cubature Kalman filter (ACKF) algorithm to estimate the attitude and heading based on smart phone-embedded MARG sensors. In this algorithm, the fading memory weighted method and the limited memory weighted method are used to adaptively correct the statistical characteristics of the nonlinear system and reduce the estimation bias of the filter. The latest step data is used as the memory window data of the limited memory weighted method. Moreover, for restraining the divergence, the filter innovation sequence is used to rectify the noise covariance measurements and system. Besides, an adaptive factor based on prediction residual construction is used to overcome the filter model error and the influence of abnormal disturbance. In the static test, compared with the Sage-Husa cubature Kalman filter (SHCKF), cubature Kalman filter (CKF), and extended Kalman filter (EKF), the mean absolute errors (MAE) of the heading pitch and roll calculated by the proposed algorithm decreased by 4–18%, 14–29%, and 61–77% respectively. In the dynamic test, compared with the above three filters, the MAE of the heading reduced by 1–8%, 2–18%, and 2–21%, and the mean of location errors decreased by 9–22%, 19–31%, and 32–54% respectively by using the proposed algorithm for three participants. Generally, the proposed algorithm can effectively improve the accuracy of heading. Moreover, it can also improve the accuracy of attitude under quasistatic conditions.

Download Full-text

On-Shelf Utility Mining of Sequence Data

ACM Transactions on Knowledge Discovery from Data ◽

10.1145/3457570 ◽

2021 ◽

Vol 16 (2) ◽

pp. 1-31

Author(s):

Chunkai Zhang ◽

Zilin Du ◽

Yuting Yang ◽

Wensheng Gan ◽

Philip S. Yu

Keyword(s):

High Efficiency ◽

Sequence Data ◽

Real Life ◽

Search Space ◽

Upper Bounds ◽

Utility Mining ◽

Limited Memory ◽

Time Periods ◽

High Utility ◽

Synthetic Datasets

Utility mining has emerged as an important and interesting topic owing to its wide application and considerable popularity. However, conventional utility mining methods have a bias toward items that have longer on-shelf time as they have a greater chance to generate a high utility. To eliminate the bias, the problem of on-shelf utility mining (OSUM) is introduced. In this article, we focus on the task of OSUM of sequence data, where the sequential database is divided into several partitions according to time periods and items are associated with utilities and several on-shelf time periods. To address the problem, we propose two methods, OSUM of sequence data (OSUMS) and OSUMS + , to extract on-shelf high-utility sequential patterns. For further efficiency, we also design several strategies to reduce the search space and avoid redundant calculation with two upper bounds time prefix extension utility ( TPEU ) and time reduced sequence utility ( TRSU ). In addition, two novel data structures are developed for facilitating the calculation of upper bounds and utilities. Substantial experimental results on certain real and synthetic datasets show that the two methods outperform the state-of-the-art algorithm. In conclusion, OSUMS may consume a large amount of memory and is unsuitable for cases with limited memory, while OSUMS + has wider real-life applications owing to its high efficiency.

Download Full-text