【作者】 江文峰； 陈渝； 史元春； 王碧波； 谷洪亮； 孙云峰；

【Author】 JIANG Wen-Feng1+, CHENYu1, SHI Yuan-Chun1, WANG Bi-Bo2, GU Hong-Liang1, SUN Yun-Feng2 1(Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China) 2(School of Software, Tsinghua University, Beijing 100084, China)

【机构】 清华大学计算机科学与技术系； 清华大学软件学院；

【摘要】 本文主要解决了在智能空间中进行固定点的坐标计算以及移动设备的位置跟踪。采用了主动模式对系统建模,使用扩展Kalman滤波器来计算三维坐标,最后实验证明本策略提高了系统的精度和鲁棒性。更多还原

【Abstract】 The text mainly solves the coordinate computing of the static points and the tracking of the mobile equipments in smart space. Through localization, the information of the positions of the particular devices can be obtained, and then the coordinate information can be provided to the equipments which need the information to supply services to the users in smart space. These make the interaction between human and computers better and the smart space more “smart”. So the coordinate information is necessary and essential to the pervasive computing. We provide a new kind of architecture of the localization system. From the point of the architecture of the hardware, the hardware called “cricketed” produced by MIT was used, some receiving units (used to receive the RF and ultrasonic signals) are disposed on the ceiling of the room, the devices to be localized bind with sending unit to send the RF and ultrasonic signals synchronously, through the cooperation of the RF and ultrasonic signals between the receiving units and the sending units, the distances of the sending unit to the particular receiving units can be obtained. This model is called active model, the converse configuration is called passive model, because there are more distances obtained through the active model at the same time, the precision computed under this model is higher, we chooses the active model in the system. After obtaining the distances, we should use proper arithmetic to get the coordinates. The arithmetic chosen is Extended Kalman Filter. The EKF arithmetic use the predicting-correction methods to evaluate the real value of the state (the coordinate value, here) under the interfering of the Gauss White Noise. In order to utilize the EKF arithmetic, we give the model of the system in the predicting-correction model in the text. In the end, we give the results of the experiment using the localization system in smart space. According to the above, we displace the hardware units used for receiving signals on the ceiling of the room, and then a sending unit is fastened to a small “vehicle”. The vehicle is moving slowly in a fixed route in smart space, at the same time, the coordinates of the vehicle are recorded and tracking results are given through the output of files. The results of the experiment shows this principal is good for improving the precision and the robustness of the localization system.更多还原