No abstract

With the rapid growth and popularity of wireless LANs (WLANs), a demand of Voice over IP (VoIP) is also increasing. In recent years, dominant trend of development in Mobile Ad hoc Networks (MANET) technology makes VoIP services more appealing. However, it is still quite challenging to ensure quality voice communication in MANET. Some of these challenges include packet loss due to network congestion, and variable packet latency. Not only traditional single-path ad-hoc routing but also traditional voice codec fails to perform well under these conditions. In this paper, we depict a framework for VoIP over multipath multihop wireless network, in which we propose multipath ad-hoc routing and traffic allocation approach.

This paper describes a localization method based on Monte Carlo Localization approach for a mobile robot. The method uses range data which are measured from ultrasound transmitting beacons whose locations are given a priori. The ultrasound receiver on-board a robot detects the range from the beacons. The method requires several beacons, theoretically over three. The method proposes a sensor model for the range sensing based on statistical analysis of the sensor output. The experiment uses commercialized beacons and detector which are used for trilateration localization. The performance of the proposed method is verified through real implementation. Especially, it is shown that the performance of the localization degrades as the sensor update rate decreases compared with the MCL algorithm update rate. Though the method requires exact location of the beacons, it doesn't require geometrical map information of the environment. Also, it is applicable to estimation of the location of both the beacons and robot simultaneously.

No abstract

This paper describes implementation of functions for mobile robot localization, which is one of the vital technologies for autonomous navigation of a mobile robot. There are several function libraries for mobile robot navigation. Some of them have limited applicability for practical use since they can be used only for simulation. Our research focuses on development of functions which can be used for localization of indoor robots. The functions implement deadreckoning and motion model of mobile robots, measurement model of range sensors, and frequently used calculations on angular directions. The functions encompass various types of robots and sensors. Also, various types of uncertainties in robot motion and sensor measurements are implemented so that the user can select proper ones for their use. The functions are tested and verified through simulation and experiments. 키워드mobile robot, localization library, autonomous navigation, differential drive robot, bicycle robot model 이동 로봇, 위치 추정 라이브러리, 자율주행, 차륜 구동 로봇, 이륜차 로봇 모델 * 조선대학교