Accurate Dead Reckoning for Omni-Directional Autonomous Mobile Robot

对移动的机器人当前位置和姿态的识别称为推算，是机器人实现自主运动的重要功能之一。计算是通过外部信号或内部传感器进行的。后者被称为航位推算。移动的机器人的航位推算最常用的方法是将编码器安装在车轮或电机轴上，而引起定位误差的主要原因是车轮的滑移。这种误差在运动过程中不断积累。因此，当机器人在大范围内运动时，它是不可忽视的。为了减小这种误差，有必要研制与车轮转动无关的速度传感器，在许多情况下，为了精确推算，内部和外部传感器同时使用。然而，最重要的是提高每个传感器的精度。为此，研制了一种高精度的内置位置传感器，利用机器人自带的小型速度传感器测量机器人的二维速度，并根据测量的速度数据计算出机器人的二维位移，从而提高了航位推算的精度，减少了外部标记的数量。此外，即使在外部传感器因障碍物而无用的区域中，也可以实现精确的自定位。

To recognize the current position and posture of a mobile robot is called reckoning, which is one of the most important function for autonomous movement of robot. Reckoning is processed by means of an outer sign or internal sensor. The latter is called dead reckoning. The most usual method for dead reckoning of mobile robot is to use encoders which are attached to wheels or motor shafts.The main cause of the error in the positioning is slippage of the wheels. This error keeps accumulate while moving. Therefore, it cannot be neglected when the robot moves in extensive area. It is necessary to develop the velocity sensor which is not related to rotation of wheels to reduce this error.In many cases, internal and external sensors are used simultaneously for accurate reckoning. However, the most important thing is to progress the accuracy of each sensor. In this investigation, an accurate internal position sensor was developed.2-dimensional velocity of the robot was measured using the compact velocity sensor onboard the robot, and 2-dimensional displacement was calculated from the measured velocity data.By using this method, the accuracy of the dead reckoning is improved and the number of outer sign can be decreased. Furthermore, accurate self-positioning can be achieved even in the area where external sensor is useless by obstacles.