State Observers for Terrain Mobility Controls: A Technical Analysis
2019; Springer Nature (Netherlands); Linguagem: Inglês
10.1007/978-3-030-20131-9_363
ISSN2211-0992
AutoresVladimir V. Vantsevich, David Gorsich, Andriy Lozynskyy, Lyubomyr Demkiv, Taras Borovets,
Tópico(s)Robotic Locomotion and Control
ResumoSome sensors in electronic controls may be not available or not feasible for terrain vehicle applications due to severe environmental conditions. The ultimate goal of this paper is to study and demonstrate a possibility to reduce the dependence of terrain mobility controls from sensors. The approach is based on designing reliable observers that can provide robust information for mobility estimation and control systems. A comparative computational analysis is presented for the following four observers: Extended Kalman filter (EKF), unscented Kalman filter (UKF), particle filter (PF), and Luenberger observer (LO). The observers' models were integrated in a closed-loop of a real-time feedback controller, which effectiveness was tested in the presence of unknown terrain disturbances and a noise of sensors. The observer-based control system with the state feedback controller is less sensitive to sensor noise than the control system with the sensors that produce states values. EKF, UKF, PF showed more robust results to disturbances than LO. In particular, PF is more robust to sensor noise than the other observers at high levels of noise.
Referência(s)