Predicted Potential Risk-based Vehicle Motion Control of Automated Vehicles for Integrated Risk Management

Abstract

In recent years, global passenger vehicle sales exceed 60milion units per year. With the increasing number of vehicles on the road, safety has become a focal issue. In order to deal with the safety issue, a number of active safety systems have been developed in passenger vehicles, such as brae assist system (BAS), adaptive cruise control(ACC), lane keeping control(LKS), and collision mitigation(CM). The functionalities of the systems include the assistance in recognizing hazards on roadway e.g. forward vehicles, obstacles, the unexpected lane departure. Beyond the development of each independent safety system, the integrated safety system has been considered nowadays. This dissertation describes design, real-time implementation and test of a fully automated driving algorithm for automated driving in complex urban scenarios and motorways with a satisfactory safety level. The proposed algorithm consists of the following three steps: surround recognition, motion planning, and vehicle control. A full recognition of environment is achieved by data fusion and data interpretation based on the dynamic measurements from the environmental sensors. The recognition of vehicle state including longitudinal, lateral velocity, and position, and driving environment is transformed into a risk potential representation based on probabilistic prediction. The surround recognition system consists of three main modules: object classification, vehicle/non-vehicle tracking and map/lane-based localization. All system modules utilize information from surround sensors close to market such as vision sensors, radars and vehicle sensors. The objective of the motion planning module is to derive an optimal trajectory as a function of time and the surround recognition results. A safety envelope is represented as a complete driving corridor that leads to destination while making sure all objects are either on outside of the left or right corridor bounds. In the case of moving objects such as other traffic participants, their behaviors are anticipated within specific time horizon. The optimal trajectory planning uses the safety envelope as a constraint and computes a trajectory that the vehicle stays in its safe bounds considering drivers pattern and characteristics based on predicted risk potential method. The performance of the proposed algorithm has been verified via computer simulations and vehicle test. From the simulation and vehicle test results, it has been shown that the proposed automated driving control algorithm enhances safety with respect to the potential risk considering driver acceptability.