Learning-based vehicle suspension controller design: A review of the state-of-the-art and future research potentials

In this paper, a review of the literature on vehicle suspension control with an emphasis on learning based algorithms is presented. An elaborated discussion on the potentials of learning based controllers is also given. Some of the most well-known active and semi-active suspension controllers are elicited from the literature, and their pros and cons are reported. By categorizing the existing suspension control techniques and considering their functionalities, it is tried to make a ground for indicating the high potential of learning strategies for this task. In this context, several advantageous features of statistical learning and computational intelligence methods are enumerated, which can play a key role in the future of vehicle suspension control. In the authors’ view, the necessity of considering learning strategies for suspension control lies in the fact that, over the past two decades, tremendous effort has been exerted on developing smart and autonomous vehicles to reduce the need for human-machine interaction. Given the fact that the final goal of automotive industrialists is to design efficient and safe autonomous vehicles, it is impossible to neglect the pivotal role of probabilistic artificial intelligence and statistical learning algorithms. In short, by reading this review paper, one can find out (1) the state-of-the-art of the conducted researches on suspension controllers, (2) the potential of learning strategies and their applicability to vehicle suspension control, and (3) the important open questions which deserve further investigation to come up with robust, stable and accurate learnable controllers. The outcome of this paper can be of use for practitioners working on designing smart and autonomous vehicles.