Calibration of an Inertial Measurement Unit and Its Impact on Antilock Braking System Performance

An Inertial Measurement Unit (IMU) provides vehicle acceleration that can be used in Active Vehicle Safety Systems (AVSSs). However, the signal output from an IMU is affected by changes in its position in the vehicle and alignment, which may lead to degradation in AVSS performance. Investigators have employed physics and data-based models for countering the impact of sensor misalignment, and the effects of gravity on acceleration measurements. While physics-based methods utilize parameters varying dynamically with vehicle motion, data-based methods require an extensive number of parameters making them computationally expensive. These factors make the above-explored methods practically challenging to implement on production vehicles. This study considers a 6-axis IMU and evaluates its impact on Antilock Braking System (ABS) performance by considering the IMU signal obtained with different mounting orientations, and positions on a Heavy Commercial Road Vehicle (HCRV). It then develops a computationally effective transformation that requires only two parameters for compensating the IMU sensor mis-orientation and extracting vehicle acceleration from the IMU signal. This transformation also provides physical intuition on the sensor mis-orientation for gauging the vehicle’s dynamic characteristics. Subsequently, a Kalman filter is utilized to estimate the unresolved offset in longitudinal acceleration. The IMU calibration combined with an ABS algorithm was evaluated in a Hardware-in-Loop experimental setup using IPG TruckMaker®. Improvements in longitudinal acceleration estimates by 94-98% were achieved with the calibration algorithm, when compared to the unprocessed IMU data. Moreover, the processed longitudinal acceleration estimates significantly enhanced wheel slip estimation performance by over 60% for a majority of the test cases, avoiding critical problems of wheel lock, and zero brake torque before the vehicle reaches its crawling speed. The outcomes of this study are expected to contribute as a critical block in the development of an indigenous ABS solution for HCRVs.