Identificação e controle de um veículo elétrico com diferencial eletrônico

Author(s)

Alexandre Monteiro Ribeiro

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Mecânica
• Defense date: 2021-11-18

Examining board members

André Ricardo Fioravanti; Marcelo Becker; Ubirajara Franco Moreno; Ricardo Coração de Leão Fontoura de Oliveira

• Advisor: André Ricardo Fioravanti; Ely Carneiro de Paiva; Alexandra Bento Moutinho

Abstract

In the context of road vehicles, loss of stability may have critical safety implications. From this fact, we raise the need of studying vehicle stability for secure and reliable maneuvers execution. Simply determining whether the car, in a particular maneuver, is stable does not offer a comprehensive portrait of how far the system is from the stability boundary. By finding the region of attraction around the equilibrium point of the vehicular system, on the other hand, we can determine the direction and dimension of disturbances that will cause the vehicle to cross into the unstable region. Within this field, this dissertation focuses upon ground vehicle stability analysis through the region of attraction (RoA) estimation. This region is defined by the set of initial conditions for which the system trajectories converge to the equilibrium. With the aid of sum-of-squares (SOS) programming techniques, Lyapunov functions, whose level sets are inner-bounds of the RoA, are found. The optimization approach is explored algorithmically and areas of local stability analysis and control synthesis are covered. A discussion around the decision variables to obtain larger inner bounds on the RoA is provided. For local stability analysis, we present the SOS program that estimates the region of attraction for polynomial systems. The tire forces are approximated using both polynomial and rational functions and the lateral dynamics of a nonlinear vehicle model is written as a set of polynomial ordinary differential equations. The RoA is then estimated for the vehicle under various constant speed cornering and straight-line forward motion. For controller synthesis, we present the SOS program that searches for a state feedback polynomial control law with input saturation for the objective of not only estimating, but also expanding the RoA. One major difficulty in this design is that the SOS generalizations assume affine-input systems, in which the vehicle model does not belong. The issue is addressed using the first-order Taylor expansion. A detailed discussion of such approximation is regarded. The controller developed in this dissertation is evaluated in a scaled vehicle platform. To show that this vehicle is a valid and reliable test-bed platform whose lateral dynamics are similar to those of a full-sized vehicle, a thorough dynamic characterization is performed. With a persistent agreement between the theoretical and measured responses, the SOS-based analyses are confidently performed. The SOS optimization-based methods in this dissertation complement the existing nonlinear analysis and design methods in the context of ground vehicles (AU)

• FAPESP's process: 18/05712-2 - Identification and control of a robotic electric vehicle with electronic differential
• Grantee: Alexandre Monteiro Ribeiro
• Support Opportunities: Scholarships in Brazil - Doctorate