Fully autonomous trajectory estimation with long-range passive RFID

ó We present a novel approach which enables a mobile robot to estimate its trajectory in an unknown envi- ronment with long-range passive radio-frequency identication (RFID). The estimation is based only on odometry and RFID measurements. The technique requires no prior observation model and makes no assumptions on the RFID setup. In particular, it is adaptive to the power level, the way the RFID antennas are mounted on the robot, and environmental characteristics, which have major impact on long-range RFID measurements. Tag positions need not be known in advance, and only the arbitrary, given infrastructure of RFID tags in the environment is utilized. By a series of experiments with a mobile robot, we show that trajectory estimation is achieved accurately and robustly. I. INTRODUCTION Radio-frequency identication (RFID) is a technology that allows for the contactless identication of goods. Its growing use in economy makes it attractive also for robotics applica- tions. Whenever a mobile robot is already equipped with an on-board RFID reader, it can cost-efciently exploit remote RFID transponders (also called tags) as uniquely identiable landmarks for navigation. Long-range passive RFID (860- 915 MHz), the prevailing technology in industry with a read range of up to 10 m, suffers from frequent nondetections of tags in read range. Moreover, passive RFID almost entirely lacks distance and bearing information between RFID reader and transponders, and measurements are inuenced by ma- terials such as water and metal surrounding the tags. In this paper, we present a novel method for trajectory estimation that actually exploits the high degree of location- specicity of those sometimes undesirable characteristics. In densely tagged environments with tags on walls and in shelves (supermarkets, for example), each single RFID measurement can contain enough information to roughly estimate the pose of the robot. This is although tags are only detected, but not localized during trajectory estimation. Here, we even treat the case that the tags are located in unknown positions, making our approach suitable for autonomous mapping without prior map. In brief, our technique works as follows: We observe RFID measurements and odometry readings while the robot is exploring an unknown environment. The sensor readings are used to rst derive an observation model for the given hardware conguration and the explored environment. This