A Predictor-based Attitude and Position Estimation for Rigid Bodies Moving in Planar Space by using Delayed Landmark Measurements

TL;DR

This paper introduces a globally convergent predictive observer for rigid body attitude and position estimation in planar space, effectively handling delayed landmark measurements through a novel PDE-based approach.

Contribution

It presents a new observer design that guarantees exponential convergence despite measurement delays, including a PDE-free implementation for practical use.

Findings

Observer achieves exponential convergence of estimation errors.

Maximum tolerable sensor delay for convergence is derived.

Simulation confirms effectiveness on a wheeled mobile robot.

Abstract

This paper proposes a globally and exponentially convergent predictive observer for attitude and position estimation based on landmark measurements and velocity (angular and linear) readings. It is assumed that landmark measurements are available with time-delay. The maximum value of the sensor delay under which the estimation error converges to zero is calculated. Synthesis of the observer is based on a representation of rigid-body kinematics and sensor delay, formulated via ordinary and partial differential equations (ODE-PDE). Observability condition specifies necessary and sufficient landmark configuration for convergence of attitude and position estimation error to zero. Finally, for implementation purposes, a PDE-free realization of the predictive observer is proposed. Simulation results are presented to demonstrate performance and convergence properties of the predictive observer in case of a wheeled mobile robot.