In order to overcome the uncertainty of the data sampling period of the sensor due to equipment reasons, a mobile robot localization system is developed under the uncertain sampling period for the tightly-fused light detection and ranging (LiDAR), compass, and encoder data. The errors of position and velocity, the robot’s yaw, and the sampling period are chosen as state variables. The ranges between the corner feature points (CFPs) and the mobile robot measured by the LiDAR, compass, and encoder are considered as an observation. Based on the tightly-integrated nonlinear model, the extended unbiased finite-impulse response (EFIR) filter fuses the sensors’ data for the integrated localization system. The performances of the traditional loosely-coupled integration scheme, tightly-coupled integration scheme with a constant sampling interval, and tightly-coupled integration with an uncertain sampling interval are compared based on real data. It is shown experimentally that the proposed scheme is more accurate then the traditional loosely-coupled integration and the one relying on a constant sampling interval, which improves by about 10.2%.

为了克服由于设备原因而导致的传感器数据采样周期的不确定性，在不确定采样周期下开发了用于紧密融合光检测和测距（LiDAR），指南针和编码器数据的移动机器人定位系统。选择位置和速度的误差，机器人的偏航角以及采样周期作为状态变量。由LiDAR，指南针和编码器测量的拐角特征点（CFP）和移动机器人之间的范围被视为观察值。基于紧密集成的非线性模型，扩展的无偏有限冲激响应（EFIR）滤波器融合了传感器的数据，用于集成定位系统。传统的松耦合积分方案，具有恒定采样间隔的紧密耦合积分方案的性能，根据实际数据比较不确定采样间隔的紧密耦合积分。实验表明，提出的方案比传统的松耦合集成更准确，后者依赖于恒定的采样间隔，可提高约10.2％。