Search Results

This paper presents the design and evaluation of an emergency driving support (EDS) algorithm. The control objective is to assist driver's collision avoidance maneuver to overcome a hazardous situation. To support driver, electrically controllable chassis components such as motor driven power steering (MDPS) and differential braking and surrounding sensor systems such as radar and camera are used. The EDS algorithm is designed for 3 parts: monitoring, decision, and control. The proposed EDS algorithm recognizes a collision danger using minimum lateral acceleration to avoid collision and time-to-collision (TTC) and driver's intention using sensor systems. The control mode is determined using the indices from monitoring process and the collision avoidance trajectory is derived with trapezoidal acceleration profile (TAP).

This paper describes a driving control algorithm based on a skid steering for a Robotic Vehicle with Articulated Suspension (RVAS). The RVAS is a kind of unmanned ground vehicle based on a skid steering using independent in-wheel drive at each wheel. The driving control algorithm consists of four parts: a speed controller for following a desired speed, a lateral motion controller that computes a yaw moment input to track a desired yaw rate or a desired trajectory according to the control mode, a longitudinal tire force distribution algorithm that determines an optimal desired longitudinal tire force and a wheel torque controller that determines a wheel torque command at each wheel in order to keep the slip ratio at each wheel below a limit value as well as to track the desired tire force. The longitudinal and vertical tire force estimators are required for the optimal tire force distribution and wheel slip control.

This paper presents a closed-loop evaluation of the Vehicle Stability Control (VSC) systems using a vehicle simulator. Human driver-VSC interactions have been investigated under realistic operating conditions in the laboratory. Braking control inputs for vehicle stability enhancement have been directly derived from the sliding control law based on vehicle planar motion equations with differential braking. A driving simulator which consists of a three-dimensional vehicle dynamic model, interface between human driver and vehicle simulator, three-dimensional animation program and a visual display has been validated using actual vehicle driving test data. Real-time human-in-the loop simulation results in realistic driving situations have shown that the proposed controller reduces driving effort and enhances vehicle stability.

An estimation algorithm for vehicle driving load has been proposed in this paper. Driving load is an important factor in a vehicle's longitudinal motion control. An approach using an observer is introduced to estimate driving load based on inexpensive RPM sensors currently being used in production vehicles. Also, the new torque estimation technique using neural network has been incorporated in this estimation algorithm to achieve better performance over variations in the automotive power transmissions process. The effectiveness of the observer-based method is demonstrated through the use of a nonlinear full vehicle simulation model in various scenarios. The proposed method using an observer has good performance, both over modeling error in powertrain system and under the uncertain environment of a running vehicle.

The performance of an integrated Vehicle Stability Control (VSC) system depends on not only control logic itself, but also the performance of state estimator and control threshold. In conventional VSCs, a control threshold is designed by vehicle characteristics and is centered on average drivers. A VSC algorithm with variable control threshold has been investigated in this study. The control threshold can be determined by phase plane analysis of side slip angle and angular velocity. Vehicle side slip angle estimator has been evaluated using test data. Estimated side slip angle has been used in the determination of the control threshold. The performance of the proposed VSC algorithm has been investigated by human-in-the-loop simulation using a vehicle simulator. The simulation results show that the control threshold has to be determined with respect to the driver steering characteristics.

CW/CA (Collision Warning /Collision Avoidance) Systems have been an active research and development area as interests and demands for the advanced vehicle increase. A CW/CA ‘Hardware-in-the-Loop Simulation (HiLS)’ system has been designed and used to test a CW/CA algorithm, radar sensors, and warning displays under realistic operating conditions in the laboratory. A CW/CA algorithm has two parts. One is a distance decision algorithm that determines the critical warning and braking distance and the other is a brake control algorithm for collision avoidance. The CW/CA HiLS system consists of a controller in which a DSP chip is installed, a preceding vehicle simulator, a radar sensor and a warning display. The controller calculates velocities of the preceding and following vehicles, relative distance and relative velocity of the vehicles using vehicle simulation models. The relative distance and velocity are applied to the vehicle simulator that is controlled by a DC motor.

As an effective approach for the design, implementation and test of control systems, hardware-in-the-loop (HIL) test has been used in many research areas. This paper describes a real-time HIL simulation test for an automotive electronic control system. The HIL system proposed in this paper consists of three parts: real-time target hardware, electronic control unit (ECU) of the automotive electronic control systems and a signal-conditioning unit which regulates the voltage levels between real-time target and ECU. The HIL simulation evaluates mechanical and electronic behaviors in real time using off-line simulation models by interfacing real-target with electrical control units via interface box. The model has been developed by MATLAB/Simulink. The model is composed of mechanical part which predicts dynamic behaviors and electronic part to calculate the motor speeds, current and electronic loads under the various conditions.

This paper describes a robust Model Predictive Control (MPC) framework of lane change for automated driving vehicles. In order to develop a safe lane change for automated driving, the driving mode and lane change direction are determined considering environmental information, sensor uncertainties, and collision risks. The safety margin is calculated using predicted trajectories of surround and subject vehicles. The MPC based combined steering and longitudinal acceleration control law has been designed with extended bicycle model over a finite time horizon. A reachable set of vehicle state is calculated on-line to guarantee that MPC state and input constraints are satisfied in the presence of disturbances and uncertainties. The performance of the proposed algorithm has been conducted simulation studies.

This paper presents a longitudinal control algorithm for an adaptive cruise control (ACC) with collision avoidance (CA) in multiple vehicle traffic situations. The proposed algorithm consists of a multi-target tracking filter, a primary target selection algorithm and an integrated ACC/CA system. The multi-target tracking filter is used to smooth the sensor signal, and makes it possible to apply to a control system. The primary target selection algorithm decides an in-lane target and provides the information to an integrated ACC/CA system in order to drive a subject vehicle smoothly and improve safety in complex traffic situations. Finally, the integrated ACC/CA system computes the desired acceleration. The performance and safety benefits of the multi-vehicle ACC/CA system is investigated via simulations using real data on driving. Simulation results show that the response of multi-vehicle ACC/CA system is more smooth and safer at a change of traffic situations.