Bacterial Programming Based Kinematic Chain Estimation of Construction Vehicle

Abstract

Construction vehicle automation for high accuracy applications require information about the state of the machine, resulting in a fully sensitized system with precise kinematic parameters. Since the measurement of these parameters contains uncertainties, accurate measurement of them is an expensive task. Automatic calibration of link parameters makes the task of kinematic parameter determination easier. This study reports a method for forward kinematic chain estimation of an excavator by bacterial programming (BP) based on randomly placed inertial navigation systems (INS) per segments with microelectromechanical sensors (MEMS) within. MEMS INS with fusion techniques provide increasing accuracy with outstanding resilience against harsh environment in a rigid housing. With known robot kinematic the tool orientation estimation can be made more accurate also the path can be planned. The unknown model structure and parameters are established and identified by BP without any a priori or given information about the device according to Denavit-Hartenberg (DH) transformation conventions. Fundamentals of this approach are described in detail and shown on simulated measurement results.

Keywords

• Bacterial programming
• inertial navigation system
• Denavit-Hartenberg
• MEMS
• kinematics
• extended Kalman filter