CAJ | 학술논문

为了提高室内实时定位中定位精度不高，并解决定位结果跳动过快的问题，提出了一种基于稀疏表示的室内定位算法。针对基于无线信号的定位方法精度不高，采用稀疏表示多分类算法提高定位精度。针对定位结果过快的跳动，利用卡尔曼滤波对位置坐标进行滤波处理，使定位结果更加接近真实的运动轨迹。结合稀疏表示与卡尔曼滤波算法，并通过仿真实验和实验台实验验证了算法的性能，算法的定位精度比传统的KNN算法有明显提高，并且定位结果更加接近真实的运动轨迹。
위료제고실내실시정위중정위정도불고，병해결정위결과도동과쾌적문제，제출료일충기우희소표시적실내정위산법。침대기우무선신호적정위방법정도불고，채용희소표시다분류산법제고정위정도。침대정위결과과쾌적도동，이용잡이만려파대위치좌표진행려파처리，사정위결과경가접근진실적운동궤적。결합희소표시여잡이만려파산법，병통과방진실험화실험태실험험증료산법적성능，산법적정위정도비전통적KNN산법유명현제고，병차정위결과경가접근진실적운동궤적。
In this paper, a multi-classification algorithm using sparse representation for indoor positioning was proposed, to improve the positioning accuracy and smooth the positioning results.Sparse representation was introduced to reduce the positioning error. Kalman filter was introduced to smooth the positioning results,which was designed to use the trajectory information. A novel algorithm combining sparse representation and Kalman filter was proposed. Both simulation experiments and test bed experiments proved its performance.The positioning accuracy was improved significantly compared to traditional KNN algorithm,and the positioning results were much closer with real trajectory.