Applying polytopic uncertainty in the vehicle-following problem with lossy networks

In this paper, we propose a control model based on NCS theory to solve the vehicle-following problem, which presents both issues: the difficulty in the identification process and a lossy network. To deal with network losses, we use a mode dependent Markov jump linear filter coupled with a statefeedback controller. To handle parameter uncertainty, a polytopic representation of the car model is presented. The performance to be optimized is a guaranteed cost of the H
∞
norm. The proposal is compared with a full-order mode dependent H
∞
controller (also a Markovian controller), which is optimal for the network losses but cannot handle the polytopic uncertainties. Results show that our approach outperforms the controllers that are not designed to cope with parameter uncertainties.