CAJ | 학술논문

为改善传统人体行为识别方案需要专门硬件、成本较高且不易推广等缺陷，提出了一种基于WiFi的室内人体行为识别方法。该方法无需专门硬件，无需被测人体配合，基于被动采集的室内RSS特征值，通过融合算法，即可有效识别四种不同人体行为：无人、走、坐、站。实验结果表明，融合算法可以获取92.58％的识别准确度。
위개선전통인체행위식별방안수요전문경건、성본교고차불역추엄등결함，제출료일충기우WiFi적실내인체행위식별방법。해방법무수전문경건，무수피측인체배합，기우피동채집적실내RSS특정치，통과융합산법，즉가유효식별사충불동인체행위：무인、주、좌、참。실험결과표명，융합산법가이획취92.58％적식별준학도。
Previous work on the human activity recognition mainly relied on special hardware devices , which are less scalable due to raised cost .To this end ,it introduced an indoor human activity recognition system using WiFi signals in this paper .The system can be integrated into any existing WLAN networks without additional hardware supports .Also it does not need the subjects to be cooperative during the rec-ognition process .By applying a novel fusion algorithm to the passively collecting radio signal strength (RSS) data,it was shown that four diversified actions can be recognized namely ,empty,walking,sitting and standing .Experimental results showed that the fusion algorithm can achieve the average recognition accuracy rate of 92.58%.