IMU-Based End-Effector Position Estimation on Agriculture and Construction Road Vehicles

Estimating the position of the bucket or tool on an agricultural/construction vehicle has become increasingly important for operator assistance (automating repetitive movements, virtual ceiling/wall, tip-over prevention). Conventional approach instrument individual actuators within the kinematic linkage with linear position sensors. This paper develops a low-cost, retrofit-friendly alternative using inertial measurement units (IMUs) on a base and rotating link to estimate linkage angles. The method combines (i) zero-phase prefiltering (8 Hz cutoff), (ii) adaptive noise cancellation (ANC, NLMS) to reject terrain-induced vibration, and (iii) a nonlinear H_∞ observer that fuses accelerometer/gyroscope data to reject residual dynamic-acceleration effects and estimate the rotational angle. The system can run in real time at 120 Hz with one-sample latency (~8.33 ms). Tests on a laboratory rig and a full-scale tracked skid-steer loader show < 1° RMSE (and < 1.5° peak error) in linkage orientation under significant vehicle acceleration and vibration, and the approach outperforms a standard EKF baseline on the same datasets. These results demonstrate a practical, accurate, low-cost alternative to conventional linkage sensing for end-effector position estimation.