This paper considers the platooning problem: we aim to steer a train of vehicles along an unknown path generated by the first vehicle, which is human driven. Among existing approaches, we study a decentralised local approach to avoid robustness issues due to communication failure which are common to centralised approaches. However, decentralised control rises up the problem of lateral deviation which is accumulated along the platoon. The work presented in this paper proposes such an approach, which avoids this well known issue. Each robot uses its perceptions to compute the position of its preceding vehicle, which is memorized with those of previous time step to form a path. It then tends to follow this path instead of aiming the actual preceding vehicle's position. In other terms, the lateral controller will have as input a position on this path which is closer than the preceding vehicle's position. Experimentations were performed on Khepera III differential wheeled robots. They highlight a drastic reduction of the lateral error in the platoon motion, comparing to the motion obtained with usual controllers.