Comparison of Linear Displacement Measurements Between A Mems Accelerometer and Hc-Sr04 Low-Cost Ultrasonic Sensor

In mobile robotics measuring displacement with high level of certainty is very critical to successful localization and mapping. However, sensors such as the depth cameras and LIDARS are often very expensive and difficult to obtain in Bangladesh. To circumvent this problem researchers in Bangladesh uses HC-SR04, a low-cost ultrasonic ranger finder for displacement measurement. MPU6050 a low-cost IMU contains a 3 axis accelerometer which enables it to calculate displacement by double integrating acceleration data. This paper compares the displacement measured using HC-SR04 to the displacement measured by MPU6050 in order to determine whether the displacement calculated from the accelerometer data alone is sufficiently accurate for robot localization. Single step trapezoidal integration method was used and the accelerometer was calibrated using 7000 data sample requiring 1604ms. A convolution type Moving Average Filter took acceleration data 64 times in each time instance for removing noise signals. All data were updated in 15ms intervals. A novel multiplier called Scale Factor (S.F) was introduced to compensate for the resolution issue without amplifying noise signals. Experimental result showed that displacement calculated from accelerometer data can estimate displacement in a mobile robot with reasonable accuracy, provided some assumptions are taken which are not representative of the real world scenarios.