Research on local path planning of unmanned vehicles based on improved driving risk field.

With the rapid development of the field of unmanned vehicles, motion planning based on field theory has become a research hotspot. A driving risk field is an effective means to evaluate driving safety in complex environments, and this method is frequently used in autonomous vehicle motion planning. However, existing risk field models are not sufficiently accurate for describing driving risks, often disregarding the size and driving direction restrictions of vehicles, amongst other aspects. Considering the aforementioned problems, this research improves and establishes a new risk field model, including a motor vehicle risk field, a road risk field and a pedestrian risk field. Simultaneously, it proposes a solution to the local minimum point problem caused by different scenarios and verifies the simulation in MATLAB. Finally, the Prescan and MATLAB/Simulink co-simulation platform is used to compare the traditional and improved field theory algorithms. Results show that the trajectory generated by the improved field theory algorithm is smoother, and the fluctuation amplitude and number of parameters, such as heading angle, yaw rate and roll angle during driving, are significantly reduced. These outcomes improve the stability of driving whilst smoothly reaching the target point, demonstrating high application potential for the proposed model.