Abstract
This paper presents a stochastic and predictive working-risk-assessment algorithm for excavators based on a one-layer laser scanner. The one-layer laser scanner is employed to detect objects and to estimate an object's dynamic behaviors such as the position, velocity, heading angle, and heading rate. To estimate the state variables, extended and linear Kalman filters are applied in consideration of laser-scanner information as the measurements. The excavator's working area is derived based on a kinematic analysis of the excavator's working parts. With the estimated dynamic behaviors and the kinematic analysis of the excavator's working parts, an object's behavior and the excavator's working area such as the maximum, actual, and predicted areas are computed for a working risk assessment. The four working-risk levels are defined using the predicted behavior and the working area, and the intersection-area-based quantitative-risk level has been computed. An actual test-data-based performance evaluation of the designed stochastic and predictive risk-assessment algorithm is conducted using a typical working scenario. The results show that the algorithm can evaluate the working-risk levels of the excavator during its operation.