In this paper, we have presented a lateral control system, which is capable of performing lane keeping and lane changing. This system was built based on a two level control architecture. One major difficulty in lane changing control is that the vehicle need to travel a certain distance without seeing any road references when the on-board sensors range cannot cover both the two adjacent lanes. By using the concept of Virtual Desired Trajectory (VDT), the lane changing problem is then transferred to a lane tracking task. A multi-model fuzzy controller, which includes four local controllers and a fusion block, was then proposed for the lane tracking tasks in both the lane keeping and lane changing controllers. Simulations showed that it could get good performances in the whole range of operation speed, and it could also repel the system uncertainties such as changes in vehicle load, movement inertia and wheel stiffness. Furthermore, as in each local controller, there are only 3 linguistic values for the input variables and 5 linguistic values for the output variable, and there are totaly 9 fuzzy rules. Thus the calculation procedure is not complex, and is rather rapid. It appears a promising control algorithm for realtime applications and to be embedded in our reduced scale 5-vehicle platoon platform.
Lateral control of autonomous vehicles using multi-model and fuzzy approaches
2010
7 Seiten, 10 Bilder, 3 Tabellen, 21 Quellen
Aufsatz (Konferenz)
Datenträger
Englisch
Lateral control of autonomous vehicles based on fuzzy logic
Tema Archiv | 2015
|Neural Network Approaches for Lateral Control of Autonomous Highway Vehicles
SAE Technical Papers | 1991
|Neural networks approaches for lateral control of autonomous highway vehicles
Kraftfahrwesen | 1991
|