The estimation of tire/road contact forces is a classic problem in Automotive Engineering. Its importance, mainly in control applications, have grown in the last years due to the evolution that cars experimented. Active Safety Systems, such as ABS (Antilock Braking System), TCS (Traction Control System) and Stability Control fully rely on the knowledge of such forces. Their presence in almost any commercial vehicle, over the years, have considerably increased safety and driving comfort. On top of these basic control systems, more complex ones such as ADAS (Advanced Driving Assistance Systems) are developed. Their diffusion and complexity increased in the last years, laying the foundation of Autonomous Vehicles control systems. In the present huge body of literature, a lot of different estimation techniques have been proposed. Most of them suffer from a number of common problems, which limit their applicability. This work aims to overcome the major issues, following a different approach. Both the longitudinal and lateral forces are estimated online for each one of the wheels, basing only on measurements provided by standard cars sensors. No contact forces model is used, so that the resulting scheme relies only on a nonlinear model of the vehicle dynamics. Five Second Order Sliding Mode observers, arranged in a particular configuration, guarantee a fast convergence and robustness. An adaptive version of the Sub-Optimal (SSOSM) algorithm is implemented, to guarantee a wide range of application while keeping the chattering low. A theoretical discussion is given about the boundedness of the resulting estimation errors and some simulations, carried out with professional simulation software, are proposed to validate the work.
La stima delle forze di contatto ruota/strada è un problema classico in ambito Automotive Engineering. La sua importanza, principalmente in applicazioni di controllo, è aumentata notevolmente negli ultimi anni a causa dell'evoluzione dei veicoli. I sistemi di sicurezza attivi, come ABS (sistema di frenata antibloccaggio), TCS (controllo di trazione) e controllo di stabilità si basano interamente sulla conoscenza di tali forze. La loro presenza nella maggior parte dei veicoli commerciali ha permesso, negli anni, di aumentare considerevolmente la sicurezza e il comfort di guida. Sulla base di questi sistemi di controllo, ne sono sviluppati di più complessi, come i sistemi di aiuto alla guida (ADAS, Advanced Driving Assistance Systems). La loro diffusione e complessità sono aumentate considerevolmente negli ultimi anni, gettando le basi dei sistemi di controllo per veicoli a guida autonoma. Nella vastissima letteratura presente, è proposto un gran numero di differenti tecniche di stima. La maggior parte di esse è soggetta a problemi che ne limitano l'effettiva applicabilità. Questo lavoro ha l'obiettivo di superare tali limitazioni, seguendo un diverso approccio. Le forze longitudinali e laterali vengono stimate online per ciascuna delle ruote, partendo solo da misure fornite da sensori standard, presenti di serie nella maggior parte delle automobili. Non vengono utilizzati modelli per le forze di contatto: lo schema risultante si basa unicamente su un modello non lineare della dinamica del veicolo. Cinque osservatori Sliding Mode del secondo ordine, disposti in una particolare configurazione, garantiscono robustezza e una veloce convergenza. Viene proposta una versione adattativa dell'algoritmo Sub-Optimal, per garantire un ampio range di applicabilità mantenendo basso il chattering. Una discussione teorica sulla limitatezza degli errori di stima e diverse simulazioni, effettuate con software di simulazione professionali, sono infine proposte a validazione del lavoro.
Tire/Road Contact Forces Estimation in Cars via Adaptive Second Order Sliding Mode Observers
ZAMBELLI, MASSIMO
2016/2017
Abstract
The estimation of tire/road contact forces is a classic problem in Automotive Engineering. Its importance, mainly in control applications, have grown in the last years due to the evolution that cars experimented. Active Safety Systems, such as ABS (Antilock Braking System), TCS (Traction Control System) and Stability Control fully rely on the knowledge of such forces. Their presence in almost any commercial vehicle, over the years, have considerably increased safety and driving comfort. On top of these basic control systems, more complex ones such as ADAS (Advanced Driving Assistance Systems) are developed. Their diffusion and complexity increased in the last years, laying the foundation of Autonomous Vehicles control systems. In the present huge body of literature, a lot of different estimation techniques have been proposed. Most of them suffer from a number of common problems, which limit their applicability. This work aims to overcome the major issues, following a different approach. Both the longitudinal and lateral forces are estimated online for each one of the wheels, basing only on measurements provided by standard cars sensors. No contact forces model is used, so that the resulting scheme relies only on a nonlinear model of the vehicle dynamics. Five Second Order Sliding Mode observers, arranged in a particular configuration, guarantee a fast convergence and robustness. An adaptive version of the Sub-Optimal (SSOSM) algorithm is implemented, to guarantee a wide range of application while keeping the chattering low. A theoretical discussion is given about the boundedness of the resulting estimation errors and some simulations, carried out with professional simulation software, are proposed to validate the work.È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: unitesi@unipv.it.
https://hdl.handle.net/20.500.14239/18885