A unified H∞ dynamic observer for vehicle state and road profile estimation in half-vehicle models

Vehicle state variables are essential to accurately estimating to enhance control performance, ride comfort and overall vehicle safety, particularly in the contemporary intelligent and autonomous vehicles. The observer-based estimation is an efficient alternative to expensive or unreliable sensors, which rebuilds important dynamic quantities based on limited measurements. Here, an integrated H ∞ dynamic observer has a major benefit in its dynamic gain structure, which gives better estimation accuracy and is effective in reducing the effect of disturbances, without the need to explicitly measure them. This observer has a better robustness than classical proportional observers, which have steady-state errors, and proportional–integral observers, whose performance is worse in the case of unknown inputs, due to the use of linear matrix inequality-based optimization. The given approach is evaluated based on MATLAB simulations of a half-vehicle suspension model that is exposed to disturbances and unmeasured road inputs. Findings indicate that it has a better ability to estimate states and reconstruct the road excitation in a single framework, which proves its advantages compared to traditional observer designs.