This article deals with a problem of the robot localization in the outdoor environment by using the GPS （global positioning system） data. In order to navigate the robot, it is necessary to transform the global posit...This article deals with a problem of the robot localization in the outdoor environment by using the GPS （global positioning system） data. In order to navigate the robot, it is necessary to transform the global position into the local map in the form of two-dimensional Cartesian coordinate system. The transformation is based on the model of the Earth-WGS 84 reference ellipsoid. The aim of this article is to experimentally evaluate a set of low-cost GPS receivers applicable as position sensors for small outdoor mobile robots. The evaluation is based on series of measurements executed in different times and places. The measured data is processed by given procedure and acquired positions are transformed into the local coordinate system. Accordingly the accuracy of the measured positions is statistically evaluated. The evaluation of used GPS receivers is done by comparison with data acquired by high-end geodetic GPS system Leica 1200, which is used as a reference GPS system.展开更多
WLAN mdoor location method based on artificial neural network （ANN） is analyzed. A three layer feed-forward ANN model offers the benefits of reducing time cost of the layout of an indoor location system, saving stor...WLAN mdoor location method based on artificial neural network （ANN） is analyzed. A three layer feed-forward ANN model offers the benefits of reducing time cost of the layout of an indoor location system, saving storage cost of the radio map establishment and enhancing real-time capacity in the on-line phase. According to the analysis of SNR distributions of recorded beacon signal samples and discussion about the multi-mode phenomenon, the one map method is proposed for the purpose of simplifying ANN input values and increasing location performances. Based on the simulations and comparison analysis with other two typical indoor location methods, K-nearest neighbor （KNN） and probability, the feasibility and effectiveness of ANN-based indoor location method are verified with average location error of 2.37m and location accuracy of 78.6% in 3m.展开更多
An accurate low-cost ultrasonic localization system is de- veloped for automated mobile robots in indoor environments, which is essential for automatic navigation of mobile robots with various tasks. Although ultrasen...An accurate low-cost ultrasonic localization system is de- veloped for automated mobile robots in indoor environments, which is essential for automatic navigation of mobile robots with various tasks. Although ultrasenic sensors are more cost-effective than other sensors such as Laser Range Finder （LRF） and vision, but they are inaccurate and directionally ambiguons. First, the matched filter is used to measure the distance accurately. For resolving the computational complexity of the matched filter, a new matched filter algorithm with simple compution is proposed. Then, an ultrasonic localization system is proposed which consists of three ultrasonic receivers and two or mote transmitters for improving position and orientation accuracy was developed. Finally, an extended Kalman filter is designed to estimate both the static and dynamic positions and orientations. Various simu lations and experimental results show that the proposed system is effective.展开更多
文摘This article deals with a problem of the robot localization in the outdoor environment by using the GPS （global positioning system） data. In order to navigate the robot, it is necessary to transform the global position into the local map in the form of two-dimensional Cartesian coordinate system. The transformation is based on the model of the Earth-WGS 84 reference ellipsoid. The aim of this article is to experimentally evaluate a set of low-cost GPS receivers applicable as position sensors for small outdoor mobile robots. The evaluation is based on series of measurements executed in different times and places. The measured data is processed by given procedure and acquired positions are transformed into the local coordinate system. Accordingly the accuracy of the measured positions is statistically evaluated. The evaluation of used GPS receivers is done by comparison with data acquired by high-end geodetic GPS system Leica 1200, which is used as a reference GPS system.
文摘WLAN mdoor location method based on artificial neural network （ANN） is analyzed. A three layer feed-forward ANN model offers the benefits of reducing time cost of the layout of an indoor location system, saving storage cost of the radio map establishment and enhancing real-time capacity in the on-line phase. According to the analysis of SNR distributions of recorded beacon signal samples and discussion about the multi-mode phenomenon, the one map method is proposed for the purpose of simplifying ANN input values and increasing location performances. Based on the simulations and comparison analysis with other two typical indoor location methods, K-nearest neighbor （KNN） and probability, the feasibility and effectiveness of ANN-based indoor location method are verified with average location error of 2.37m and location accuracy of 78.6% in 3m.
基金supported by the MKE(The Ministry of Knowledge Economy,Korea)the ITRC(Information Technology Research Center)support program(ⅡTA-2009-(C1090-0902-0007))
文摘An accurate low-cost ultrasonic localization system is de- veloped for automated mobile robots in indoor environments, which is essential for automatic navigation of mobile robots with various tasks. Although ultrasenic sensors are more cost-effective than other sensors such as Laser Range Finder （LRF） and vision, but they are inaccurate and directionally ambiguons. First, the matched filter is used to measure the distance accurately. For resolving the computational complexity of the matched filter, a new matched filter algorithm with simple compution is proposed. Then, an ultrasonic localization system is proposed which consists of three ultrasonic receivers and two or mote transmitters for improving position and orientation accuracy was developed. Finally, an extended Kalman filter is designed to estimate both the static and dynamic positions and orientations. Various simu lations and experimental results show that the proposed system is effective.
